Richard P. Wennberg

The Blood–Brain Barrier and Bilirubin Encephalopathy

Abstract1. The pathogenesis of bilirubin encephalopathy is multifactorial, involving the transport of bilirubin or albumin/bilirubin across the blood–brain barrier and delivering bilirubin to target neurons.2. The relative importance of the blood–brain barrier, unconjugated bilirubin levels, serum binding, and tissue susceptibility in this process is only partially understood. Even at dangerously high serum levels, bilirubin traverses the intact blood–brain barrier slowly, requiring time for encephalopathy to occur, although deposition of bilirubin can be rapid if a surge in plasma unbound bilirubin is produced by administering a drug which competes with bilirubin for binding to albumin.3. There may be maturational changes in permeability both in the fetus and postnatally which protect the brain from bilirubin.4. Disruption or partial disruption of the blood–brain barrier by disease or hypoxic ischemic injury will facilitate transport of bilirubin/albumin into brain, but the relative affinities of albumin and target neurons will determine whether the tissue bilirubin load is sufficient for toxicity to occur.

Experimental Bilirubin Encephalopathy: Importance of Total Bilirubin, Protein Binding, and Blood-Brain Barrier

ABSTRACT. The cause of bilirubin encephalopathy has been variously ascribed to elevated total serum bilirubin concentration, high free bilirubin levels (or impaired albumin binding), and disruption of the blood-brain barrier. An experimental rat model for acute bilirubin encephalopathy was developed in which these three factors could be varied independently. Osmotic opening of the blood-brain barrier in the right hemisphere was produced by infusing a hypertonic arabinose solution into the right carotid artery. The total bilirubin level and bilirubin binding state were varied by adjusting the amount of bilirubin infused intravenously and/or by infusing human serum albumin. Brain electrical activity (EEG) served as an indicator of developing encephalopathy. Neither staining nor EEG changes occurred if the blood-brain barrier remained intact. Bilirubin staining without EEG evidence of encephalopathy sometimes occurred when the blood-brain barrier was open. Discriminant analysis showed that EEG changes were best predicted by the degree of blood-brain barrier opening (as indicated by brain bilirubin content) and by the quality of serum bilirubin binding. Serum total bilirubin concentration was not an important discriminator of encephalopathy.

Cytotoxicity Is Predicted by Unbound and Not Total Bilirubin Concentration

Although it has been suggested that the unbound, free, (Bf) rather than total (BT) bilirubin level correlates with cell toxicity, direct experimental evidence supporting this conclusion is limited. In addition, previous studies never included a direct measurement of Bf, using newer, accurate methods. To test “the free bilirubin hypothesis”, in vitro cytotoxicity was assessed in four cell lines exposed to different Bf concentrations obtained by varying BT/Albumin ratio, using serum albumins with different binding affinities, and/or displacing unconjugated bilirubin (UCB) from albumin with a sulphonamide. Bf was assessed by the modified, minimally diluted peroxidase method. Cytotoxicity varied among cell lines but was invariably related to Bf and not BT. Light exposure decreased toxicity parallel to a decrease in Bf. In the absence of albumin, no cytotoxicity was found at a Bf of 150 nM whereas in the presence of albumin a similar Bf resulted in a 40% reduction of viability indicating the importance of total cellular uptake of UCB in eliciting toxic effect. In the presence of albumin-bound UCB, bilirubin-induced cytotoxicity in a given cell line is accurately predicted by Bf irrespective of the source and concentration of albumin, or total bilirubin level.

The Importance of Free Bilirubin Acid Salt in Bilirubin Uptake by Erythrocytes and Mitochondria

ABSTRACT: The binding of bilirubin to tissue was studied using adult human erythrocytes and rat liver mitochondria. Tissues were incubated with varying bilirubin-albumin molar ratios, varying albumin concentrations of a given bilirubin-albumin molar ratio, and varying pH. Bilirubin binding by tissue was reversible and stoichiometric with the concentration of the free (nonalbumin bound) bilirubin acid salt (bilirubin monovalent anion). Minimal binding of the bilirubin dianion, the predominant state of bilirubin in plasma, was also suggested. The observations support the “free bilirubin theory” where tissue and albumin compete for binding the bodys bilirubin pool. Binding to tissue, however, is not determined by the free bilirubin concentration, but by the concentration of the pH dependent subfraction, the free bilirubin acid salt. Tissue binding and toxicity of bilirubin may result from the surfactant properties of the monovalent anion.

Interactions of bilirubin with isolated presynaptic nerve terminals: Functional effects on the uptake and release of neurotransmitters

Summary1.The functional effects of bilirubin:albumin solutions (10:1, mol/mol) on several synaptosomal functions were investigated using rat cortical, striatal, and hippocampal synaptosomes prepared by iso-osmotic Percoll/sucrose gradient centrifugation.2.Bilirubin (10–80µM) depolarized synaptosomes in a tetrodotoxin-insensitive manner as assessed by the equilibrium distribution of tetra-[3H]phenylphosphonium. Depolarization induced by bilirubin was of a lesser magnitude than that caused by KCl or veratridine. Steady-state pH gradients across the synaptosomal membrane were determined using the transmembrane distribution of [14C]methylamine. Bilirubin (20–40µM) did not modify the intracellular pH in physiological buffers. The pigment effected a 0.14 ΔpH change when the synaptosomes were suspended in a Ca2+ and Na+ free choline medium containing ouabain.3.Bilirubin (20–80µM) had no effect of its own on [7,8-3H] dopamine release from striatal synaptosomes. In contrast, it inhibited the initial rate of synaptosomal uptake of the catecholamine and its intrasynaptosomal content at 10 min. The pigment (20 and 40µM) reduced the 35 mM KCl-induced release of endogenous acetylcholine from hippocampal synaptosomes by 20 and 36%, respectively.4.The association of bilirubin with synaptic plasma membrane vesicles was characterized by a chloroform:methanol 2:1 (v/v) extraction method. At total concentrations of 10 to 80µM bilirubin, the molar percentage of the pigment in synaptic plasma membrane phospholipids was 1–4%.5.It is proposed that the two main functional consequences of the bilirubin-nerve ending interaction are an impairment of specific membranebound neurotransmitter uptake mechanisms and a reduction of the response to depolarizing stimuli. This may be the basis for rapid alterations in synaptic transmission documented in early reversible bilirubin encephalopathy.

The pathochemistry of kernicterus

The stoichiometry of bilirubin--albumin interaction has been analyzed and quantitated in several recent studies, confirming that albumin binding of bilirubin obeys the law of mass action [4, 5, 14, 16, 26, 36, 43, 46, 61, 65, 73, 92, 111]. These studies provide a basis for interpreting bilirubin transport, cell uptake and toxicity from physicochemical and pharmacologic perspectives [35, 42, 58, 59]. In this report, we propose a model of the pathogenesis of kernicterus which views serum albumin and tissue as competing with each other for binding the miscible bilirubin pool. Evidence is presented to show that bilirubin normally binds reversibly to cellular membranes and certain soluble enzymes just as it does to albumin; the unbound bilirubin concentration is the driving force for both albumin and tissue binding. We propose that albumin binding is determined by the concentration of free bilirubin anion (which is essentially unaffected by physiologic pH changes), and that tissue binding is mainly determined by the concentration of free bilirubin acid (which is greatly influenced by pH). When bilirubin--tissue complexes are formed, essential cell functions may be inhibited, producing cellular acidosis, irreversible intracellular aggregation of bilirubin, and cell death. In developing this argument, we will sequentially discuss relevant features of bilirubin chemistry, the binding of bilirubin to albumin, the formation of bilirubin--tissue complexes, bilirubin toxicity, alternative viewpoints of bilirubin transport, and, finally, the implications of this model to the clinical management of jaundiced infants. It should be emphasized that this paper is an attempt to analyze bilirubin transport and toxicity using basic chemical principles; it is an extension of previously published proposals [17, 77], and will undoubtedly require further modification as additional experimental data becomes available.

Bilirubin-Induced Changes in Brain Energy Metabolism after Osmotic Opening of the Blood-Brain Barrier

ABSTRACT: Acute and residual effects of blood-brain barrier disruption and bilirubin on brain metabolism were studied in a rat model after osmotic opening of the blood-brain barrier under pentobarbital anesthesia. Arabinose (1.5 M) was infused via the right external carotid artery over 30 s, resulting in opening of the barrier within the right hemisphere. Two min later, bilirubin was infused i.v. over 3 min, raising the serum bilirubin concentration to 37–44 mg/dL (633–752 μmol/L). The animals were euthanized at 15 min or 4 h by freezing the brain in situ. Opening the blood-brain barrier produced small changes in cerebral energy metabolism in some animals at 15 min. Compared with saline-infused control animals, two out of nine rats had decreased brain phosphocreatine and three out of nine developed increased brain lactate levels. Infusion of bilirubin in rats with a disrupted blood-brain barrier produced profound decreases in brain energy metabolites, glucose, and glycogen and a markedly increased lactate/pyruvate ratio at 15 min. The markedly increased lactate in the presence of normal or low pyruvate in bilirubin-treated animals indicates accumulation of NADH and probably reflects severe mitochondrial dysfunction. Four h after the arabinose/bilirubin infusions, the barrier would be expected to be repaired and bilirubin levels were negligible, but two out of five arabinose and three out of six bilirubin rats continued to have severely altered brain metabolism indicating residual brain injury in some animals.

The effect of paraben preservatives on albumin binding of bilirubin

The interaction of methylparaben and propylparaben with bilirubin-albumin complexes was studied using difference spectra, Sephadex gel filtration, red blood cell uptake of bilirubin, and the peroxidase assay. Methylparaben was found to be a weak competitor with bilirubin for binding to primary albumin-binding sites but a strong binding competitor (similar to sulfisoxazole) at secondary sites. The displacing effects of methylparaben and sulfisoxazole were additive. Propylparaben bound to albumin but did not displace bilirubin. Drugs and injectable saline and water preparations which contain methylparaben should be avoided in jaundiced newborn infants when the high-affinity albumin-binding sites approach saturation.

Changes in the Auditory Brainstem Response Associated with Intravenous Infusion of Unconjugated Bilirubin into Infant Rhesus Monkeys

ABSTRACT: The auditory brainstem response (ABR) was monitored in nine infant rhesus monkeys during the intravenous infusion of 50-168 mg/kg of unconjugated bilirubin. Sulfisoxazole (200 mg/kg) was sometimes given near the end of or just before the bilirubin infusion if no obvious ABR change had yet occurred. Five of the animals were term gestation, four were preterm, and they ranged from 1 to 40 days of age at the time of study. The three oldest term animals, studied at 20, 35 and 40 days of age, respectively, showed variable changes in the ABR waves during bilirubin infusion and these changes were not altered further by sulfisoxazole administration. The other two term infants, studied at 1 and 6 days of age, respectively, showed sulfisoxazole enhanced ABR wave latency increase and amplitude reduction followed by loss of the ABR. Both of these animals became apneic following ABR loss and eventually died. The ABR reappeared in one animal prior to death. Minimal gross and microscopic changes were present in the brain of the 6-day-old animal at autopsy. The four preterm animals all had a progressive wave amplitude decrease followed by loss of the ABR with bilirubin alone. These preterm animals were sacrificed shortly after the ABR loss with only one showing yellow staining of the basal ganglia at autopsy. The infant rhesus monkey may be a useful paradigm for bilirubin-induced ototoxicity as manifested by potentially reversible ABR changes. The changes are dependent on gestational and chronological age of the animal and appear to occur in the peripheral eighth nerve or cochlea as well as in brainstem pathways.

Brainstem bilirubin toxicity in the newborn primate may be promoted and reversed by modulating PCO2

ABSTRACT: The auditory brainstem response (ABR) was monitored during infusion of bilirubin in six ventilated newborn rhesus monkeys (138–145 d gestation) while acute changes in pH were produced by varying inspired CO2. Prolonged respiratory acidosis without bilirubin infusion produced minimal changes in the ABR (one animal). CO2 exposure, usually initiated when the bilirubin level reached ∼20 mg/dL, decreased arterial pH to values ranging from 6.85 to 7.10. ABR changes, including prolongation of the wave II-IV peak to peak intervals and decreased wave amplitudes, first developed 2–4 h after initial exposure to CO2. Total and unbound bilirubin levels at this time ranged from 376 to 564 μmol/L (22–33 mg/dL) and 38 to 65 nmol/ L (2.5–3.8 μg/dL), respectively. Correction of respiratory acidosis produced partial to complete reversal of ABR changes within 3 to 20 min. Reexposure to CO2 immediately reproduced the ABR abnormality. Production and reversal of the abnormal ABR was obtained through two to three cycles in three animals. Thus, when the brainstem bilirubin level was near the threshold for toxicity, the effect of changes in PCO2 on the ABR were immediate, suggesting that auditory pathway toxicity is initially mediated by a reversible pH-dependent bilirubin-membrane complex. In contrast to humans, in monkeys auditory toxicity appeared to be a late manifestation of bilirubin toxicity, inasmuch as all monkeys were obtunded and apneic 30–70 min before ABR abnormalities appeared. Notwithstanding these limitations, the results support the hypothesis that bilirubin toxicity can be both promoted and reversed by modulating brain pH.