Effects of Bilirubin on Neutrophil Responses in Newborn Infants

Abstract

Background

Newborns are susceptible to inflammatory diseases due to defects in clearing activated immune cells from tissues. Therefore, mechanisms have likely evolved to protect neonates from leukocyte-mediated cytotoxicity. Bilirubin has antioxidant activity, and it is possible that it also exerts effects on cellular immune responses in jaundiced infants.

Objectives

We hypothesize that bilirubin increases expression of antioxidant genes and decreases production of inflammatory proteins in neonatal neutrophils.

Methods

Neutrophils were isolated from umbilical cord blood, and from adults for comparison, and treated with bilirubin (10–300 μmol/l, equivalent to unbound bilirubin 3–40 nmol/l), in the presence or absence of lipopolysaccharide (LPS). Expression of genes for antioxidant enzymes [superoxide dismutase (SOD), heme-oxygenase-1 (HO-1)] and heme-dependent enzymes involved in inflammation [NADPH oxidase-1 (NOX-1), cyclooxygenase-2 (COX-2)] was measured by PCR. Inflammatory cytokines were measured by bead array analysis using flow cytometry.

Results

We found that LPS induced production of interleukin (IL)-8, IL-1β, and macrophage inhibitory protein-1β (MIP-1β). Bilirubin increased basal production of IL-8 and IL-1β, but downregulated LPS-induced generation of IL-8 and MIP-1β. It also upregulated SOD and HO-1 gene expression. We observed an unexpected bilirubin-induced increase in gene expression of NOX-1 in LPS-activated cells, and of COX-2 in both resting and activated cells.

Conclusions

These findings suggest that bilirubin suppresses inflammation and increases antioxidant enzyme generation in activated neonatal neutrophils. The unexpected increases in NOX-1 and COX-2 expression may represent an early response, with physiologic effects mitigated by increased antioxidant activity. Further studies will be needed to define levels of bilirubin that optimize its protective effects, while minimizing potential inflammatory toxicity.

Introduction

Neonatal jaundice caused by hyperbilirubinemia occurs in 60–70% of newborns. Though some degree of hyperbilirubinemia is considered physiologic and possibly neuroprotective because of bilirubin’s antioxidant properties [1], the modulating effects of bilirubin on inflammation in neonates have not been critically examined. The aim of these investigations was to define the dose-dependent effects of bilirubin on neutrophil function, which is a key determinant of health and disease in the neonatal period. Neutrophils are the main effector cells in early inflammation, producing reactive oxygen intermediates via enzymes such as NADPH oxidase-1 (NOX-1), and eicosanoids via cyclooxygenase-2 (COX-2). In vitro studies have demonstrated that purified human neutrophils are metabolically active, expressing mRNA for proteins mediating inflammation, both constitutively and after activation with bacterial lipopolysaccharide (LPS) or other stimulants [2, 3]. Neonatal neutrophils exhibit delayed apoptosis and prolonged survival under inflammatory conditions. It has recently been shown that neutrophils surviving >24 h generate increased quantities of inflammatory cytokines, including interleukin (IL)-8 and macrophage inhibitory protein-1β (MIP-1β), possibly contributing to tissue injury [4].

Bilirubin, one of the principal metabolic end-products of heme catabolism, is generated by the sequential action of heme oxygenase and biliverdin reductase [1]. It has been shown to exert anti-inflammatory effects, such as decreasing NOX-1 activation by regulating p47 phox-dependent activity [5]. Bilirubin prevents the oxidant-mediated vasoconstrictive actions of tumor necrosis factor (TNF) and angiotensin II [1]. In addition, low concentrations of bilirubin scavenge reactive oxygen intermediates (ROS), potentially reducing oxidant-induced cellular injury and attenuating oxidant stress in vivo [6, 7]. Bilirubin also prevents endothelial cell damage and sloughing in conditions such as diabetes and hypertension by up-regulating extracellular superoxide dismutase (EC-SOD) and plasma catalase activity, and decreasing superoxide production [8]. Consistent with this, we have previously demonstrated that higher bilirubin levels in preterm infants are associated with a lower incidence of inflammatory diseases, including bronchopulmonary dysplasia and necrotizing enterocolitis [9]. In the current studies, we hypothesized that bilirubin: (1) decreases inflammatory cytokine production, (2) increases expression of antioxidant genes [heme-oxygenase-1 (HO-1), SOD], and (3) decreases expression of heme-dependent enzymes (NOX-1, COX-2) associated with oxidant activity in neonatal neutrophils.

Materials and Methods

Reagents

Dulbecco’s modified Eagle’s medium (DMEM), PBS, dextran, RNAase A, and bacterial-derived LPS were purchased from Sigma Chemical Co. (St. Louis, Mo., USA). Ficoll-Paque was from GE Healthcare (Piscataway, N.J., USA). RNA purification kits were purchased from Qiagen (Chatsworth, Calif., USA). Primers for real-time PCR were obtained from Integrated DNA Technologies (Coralville, Iowa, USA). Nucleotides and reagents for PCR were from Applied Biosystems (Foster City, Calif., USA). Cytometric bead array flex sets were from BD Biosciences (San Jose, Calif., USA). Bilirubin was obtained from Calbiochem (San Diego, Calif., USA).

Bilirubin Preparation

A 5-mmol/l stock solution of bilirubin was prepared immediately before use in 0.1 mol/l NaOH. Prior to treatment, pH was restored to 7.4 by addition of equal amounts of 0.1 mol/l HCl [10, 11]. Treatments (10, 30, 100, 300 μmol/l) were added in the dark, incubated and protected from light for 4–24 h. The range of 0–300 μmol/l (0–18 mg/dl), equivalent to unbound bilirubin of between 3 and 40 nmol/l in the neutrophil culture media, was chosen for these experiments because it corresponds to the range of levels that are commonly observed in neonates under conditions of both physiologic and pathologic hyperbilirubinemia.

Subjects and Neutrophil Isolation

Studies were approved by the Institutional Review Board of UMDNJ-Robert Wood Johnson Medical School, and informed consent was obtained from participants. Umbilical cord blood was obtained from healthy term infants (≥37 weeks’ gestation) delivered by elective cesarean section prior to labor between January 2010 and January 2011. Subjects were excluded with clinical evidence of chorioamnionitis or other perinatal bacterial or viral infections, e.g. maternal fever, uterine tenderness, or foul-smelling amniotic fluid. Neutrophils were isolated by dextran sedimentation, followed by Ficoll gradient centrifugation and hypotonic lysis of erythrocytes, as previously described [12]. Isolated neutrophils were suspended in DMEM with 10% fetal bovine serum (FBS), in the presence of LPS (1 μg/ml) or medium control. Bilirubin was added in concentrations of 10, 30 100 and 300 μmol/l.

RNA Expression

Neutrophils were incubated with control, LPS (1 μg/ml), and/ or bilirubin (10, 30, 100, 300 μmol/l) for 4 h. RNA was isolated using an RNeasy Mini Kit (Qiagen). cDNA was synthesized using a high-capacity cDNA reverse transcription kit (Applied Biosystems). Real-time PCR was performed using SYBR Green PCR Master Mix (Applied Biosystems) and amplified on an ABI Prism 7900 sequence detection system, using GAPDH as standard. Full-length coding sequences were obtained from GenBank™. Primers were designed using Primer Express software (Applied Biosystems). Forward and reverse primers used were: β-actin, aaagacctgtacgccaacac and gtcatactcctgcttgctgat; SOD, gtcgtagtctcctgcagcgtc and ctggttccgaggactgcaa; HO-1, gctcaaaaagattgcccaga and gcggtagagctgcttgaact; NOX-1, ccttgcaccggtcattcttt and cggtaaaaccggaggatcct, and COX-2, gcctgatgattgcccgact and gctggccctcgcttatgatct.

Inflammatory Proteins

Culture supernatants were incubated with premixed flex set beads coated with antibodies to IL-1β, IL-8, MIP-1β, IL-6, and VEGF for 1 h in 96-well filtration plates. Premixed phycoery-thrin-labeled detection reagent was then added to the wells. The plates were incubated at room temperature in the dark for 2 h. Bead/protein complexes were then washed and analyzed for fluorescence intensity using a BD FACS Array Bioanalyzer. Data were analyzed using BD FCAP software (version 2.0) with 5-parameter curve fitting.

Measurement of Unbound Bilirubin

Neutrophils from 3 representative subjects were incubated for 4 h in medium containing 10% FBS and bilirubin concentrations of 10, 30, 100, and 300 μmol/l. Culture supernatants were analyzed for unbound bilirubin using the bilirubin probe BL22P1B11-Rh, a combination of mutated fatty acid binding protein labeled with the fluorescent molecule acrylodan and Rhodamine B as we have previously described [13]. BL22P1B11-Rh responds to binding bilirubin by reducing its ratio of 525–575 nm fluorescence (excitation = 375 nm) with a K_d for bilirubin at 22°C of 16 nmol/l. The bilirubin probe was added to the culture media and unbound bilirubin concentrations quantified.

Statistical Analysis

Experiments were repeated 6–14 times. Data were analyzed using Statistica 5.5 (StatSoft, Inc., Tulsa, Okla., USA). The effects of treatments were compared pairwise using t tests.

Results

Levels of unbound (free) bilirubin in the neutrophil culture media were correlated with total bilirubin levels, with 300 μmol/l total corresponding to about 40 nmol/l unbound (fig. 1). First, we examined the effects of bilirubin on mRNA expression of antioxidant genes SOD and HO-1, which are known to play a role in protecting against cytotoxicity and tissue damage. Bilirubin increased SOD and HO-1 gene expression in LPS-activated neutrophils. These effects were dose-dependent, with statistically significant increases at bilirubin concentrations of 300 μmol/l (fig. 2). There was also a trend towards increased expression of SOD and HO-1 genes in response to increasing concentrations of bilirubin in unstimulated neutrophils. We next compared the effects of bilirubin on mRNA expression of NOX-1 and COX-2 genes. These enzymes catalyze the generation of oxidative and pro-inflammatory mediators, including prostaglandins and superoxide anion. They are heme-dependent and inducible, with rapid turnover [14]. Bilirubin was found to upregulate expression of NOX-1 in LPS-activated neutrophils in a dose-dependent manner, with statistically significant increases at bilirubin concentration of 300 μmol/l. Bilirubin also upregulated COX-2 gene expression in both resting and LPS-activated cells, with significant effects at 100 and 300 μmol/l of bilirubin in unstimulated neutrophils, and 300 μmol/l in LPS-activated cells (fig. 3).

Fig. 1.

Association of unbound bilirubin levels with total bilirubin. Neonatal neutrophils were incubated in the presence of bilirubin (10, 30, 100 and 300 μmol/l) for 4 h. Culture supernatants were analyzed for unbound bilirubin using a mutated fatty acid binding protein labeled with the fluorescent molecule acrylodan. Each point represents mean ± SE (n = 3).

Fig. 2.

Effects of bilirubin on expression of antioxidant genes (SOD, HO-1). Neonatal neutrophils were incubated in the presence (grey bars) or absence (black bars) of LPS (1 μg/ml) and bilirubin [0 (medium control), 10, 30, 100 and 300 μmol/l] for 4 h. mRNA expression of the SOD (a) and HO-1 (b) genes was quantified by real-time PCR. Results were normalized to β-actin expression. Each bar represents the mean ± SE (n = 8–9). ** Significantly different (p < 0.05) from control + LPS.

Fig. 3.

Effects of bilirubin on expression of heme-dependent enzymes (NOX-1, COX-2). Neonatal neutrophils were incubated in the presence (grey bars) or absence (black bars) of LPS (1 μg/ml) and bilirubin [0 (medium control), 10, 30, 100 and 300 μmol/l] for 4 h. mRNA expression of the NOX-1 (a) and COX-2 (b) genes was quantified by real-time PCR. Results were normalized to β-actin expression. Each bar represents the mean ± SE (n = 8–12). * Significantly different (p < 0.05) from control. ** Significantly different (p < 0.05) from control + LPS.

We then analyzed the effects of bilirubin on neutrophil production of inflammatory proteins and chemotactic cytokines. Production of IL-1β, IL-8 and MIP-1β was significantly increased in LPS-activated cells compared to resting neutrophils. Bilirubin increased the production of IL-1β and IL-8 at concentrations between 10 and 300 μmol/l in unstimulated cells. In contrast, bilirubin (300 μmol/l) downregulated the production of MIP-1β and IL-8 in LPS-activated cells (fig. 4). No significant effects of bilirubin were noted on the production of IL-6 and VEGF (not shown).

Fig. 4.

Effects of bilirubin on inflammatory cytokine production. Neonatal neutrophils were incubated in the presence (grey bars) or absence (black bars) of LPS (1 μg/ml) and bilirubin [0 (medium control), 10, 30, 100 and 300 μmol/l] for 24 h. The inflammatory mediators IL-8 (a), IL-1β(b), and MIP-1β (c) were measured in culture supernatants using cytometric bead array analysis. Each bar represents the mean ± SE (n = 8–14). * Significantly different (p < 0.05) from control. ** Significantly different (p < 0.05) from control + LPS.

Discussion

Bilirubin formation results from the catabolism of heme-containing proteins such as hemoglobin, myoglobin, cytochromes, catalases, and tryptophan pyrrolase. The microsomal enzyme heme oxygenase catalyzes the first step of this process, resulting in the generation of biliverdin and the release of carbon monoxide. Neonates have a high heme load, as well as developmental immaturity in the metabolism and transport of bilirubin. Consequently, they exhibit transient hyperbilirubinemia, or ‘physiologic jaundice’. This condition is characterized, in the term infant, by a progressive rise in serum bilirubin concentrations to a peak of approximately 100 μmol/l by the third day of life, followed by a drop to 35 μmol/l by the end of the first week of life [15]. At physiologic levels, bilirubin has been shown to have antioxidant and cyto-protective effects. For example, it suppresses postischemic myocardial dysfunction [16], protects against LPS-induced liver injury [17], and ameliorates skin inflammation [18]. In these studies, we also investigated the effects of pathologic levels of exposure (up to 300 μmol/l, equivalent to 40 nmol/l unbound bilirubin in the neutrophil culture media), to which many infants are exposed, often as a result of hemolysis or dehydration.

Neutrophils accumulate in tissues in response to chemotactic factors generated at sites of infection or injury. Inflammatory cytokines and bacterial-derived products also trigger the generation of reactive oxygen intermediates, which are key in bacterial killing but may also play a role in tissue injury. In association with its antioxidant properties, bilirubin may suppress neutrophil phagocytosis and bacterial killing [19]. In these studies, we found that bilirubin upregulated constitutive expression of SOD and HO-1 in neonatal neutrophils. These antioxidant enzymes are key to protecting tissues from reactive oxygen species generated in the fetal and maternal circulation. HO-1 is associated with decreased endothelial cell injury and sloughing in conditions such as diabetes and hypertension [20]. It is induced by physiologic and pathologic stimuli including oxidative stress signals, cytokines, bacterial compounds and growth factors [21]. A growing number of pharmacologic compounds, such as statins [22] and 5-ASA [23], have been shown to provide anti-inflammatory protection via the induction of HO-1. Similarly, increased SOD may exert protective anti-inflammatory effects by decreasing superoxide anion. This reduces cytotoxicity and also allows levels of nitric oxide in tissues to remain stable, promoting endothelial cell progenitor function [8, 16, 20]. Our finding that bilirubin increases expression of HO-1 and SOD in neonatal neutrophils suggests that these protective mechanisms may occur in the neonatal period, which is characterized by the oxidative stress of birth and postnatal adaptation. Neonatal jaundice may have an important anti-inflammatory effect via induction of these enzymes.

In contrast to these antioxidant effects, we found that bilirubin increased the expression of NOX-1 in LPS-activated neutrophils, and COX-2 in both resting and activated cells. NOX-1 is the major enzymatic mediator of superoxide anion generation in neutrophils [24]. Previous studies have shown that neonatal neutrophils produce robust quantities of superoxide anion and inflammatory eicosanoids after stimulation with inflammatory stimuli such as LPS [25]. Our findings suggest that bilirubin induces these oxidative and inflammatory responses in neonatal neutrophils at early time points. Consistent with this, bilirubin has been shown to trigger the induction of COX-2 and inflammatory cytokines in mononuclear cells [26]. The upregulation of COX-2 in neonatal neutrophils by bilirubin observed in our studies could be a time-dependent response, with physiologic effects mitigated by the subsequent activity of HO-1 [14]. Of note, significant effects of bilirubin on expression of both pro- and antioxidant enzymes were observed only at bilirubin levels >100 μmol/l (unbound bilirubin >17 nmol/l), which are considered pathologic rather than physiologic.

Previous studies have yielded conflicting results regarding the effects of bilirubin on the production of inflammatory cytokines. For example, bilirubin induces IL-1β and TNF-α production, but suppresses IL-6, in astrocytes [27]. Bilirubin has also been shown to induce release of TNF-α, IL-1β and IL-6 in microglia [28]. However, previous studies have not examined the effects of bilirubin on resting and activated neonatal neutrophils. We found that bilirubin increased constitutive production of IL-1β and IL-8 in resting neonatal neutrophils, but suppressed IL-8 and MIP-1β in LPS-activated cells. Downregulation of these mediators under inflammatory conditions, in the presence of LPS, suggests that bilirubin may play a protective role under conditions of infection or oxidative stress in newborns.

The portion of circulating bilirubin that is not bound to serum proteins (unbound or free) is thought to be responsible for its biologic actions [29]. We found that free bilirubin levels in the culture media, which contained 10% FBS, correlated directly with total bilirubin levels, with 300 μmol/l total corresponding to about 40 nmol/l free. This is comparable to previous measurements of free bilirubin in neonatal serum, including a large cohort of newborns >2,500 g birth weight who were shown to have a mean level of 21.5 with a range of 0.9–130 nmol/l [30]. These findings suggest that total bilirubin measurements are an acceptable surrogate for free bilirubin in these experiments.

In conclusion, bilirubin may exert protective effects by upregulating antioxidant production and downregulating cytokine production in activated neonatal neutrophils. Bilirubin may also activate resting neutrophils, and the significance of this activity is not known. Hyperbilirubinemia may represent a protective adaptation, with beneficial effects given the oxidative and inflammatory stress that can occur in the perinatal period. Further studies will be required to define levels that optimize these effects while minimizing potential cytotoxicity resulting from bilirubin-induced upregulation of NOX-1 and COX-2 in neutrophils.