Experimental hemolysis model to study bilirubin encephalopathy in rat brain

Journal of Neuroscience Methods 168 (2008) 35–41

Experimental hemolysis model to study bilirubin encephalopathy in rat brain Gerardo Barrag´an Mejia a , Cecilia Ridaura Sanz b , Marco Mart´ınez Avila c , Armando Valenzuela Peraza c , David Calder´on Guzm´an a , Hugo Ju´arez Olgu´ın d,∗ , Aline Morales Ram´ırez a , Edna Garc´ıa Cruz a a

Laboratorio de Neuroqu´ımica, Instituto Nacional de Pediatr´ıa (INP), Mexico b Departamento de Patolog´ıa, INP, Mexico c Laboratorio de Neuromorfometr´ıa, INP, Mexico d Laboratorio de Farmacolog´ıa, Instituto Nacional de Pediatr´ıa (INP), Avenida Im´ an # 1, 3rd Piso, Colonia Cuicuilco, UNAM, CP 04530 M´exico, D.F., Mexico Received 8 March 2007; received in revised form 7 September 2007; accepted 7 September 2007

Abstract None of experimental models used to study the toxic effect of unconjugated bilirubin brain accumulation, reproduce the conditions in which the hyperbilirubinemia is a consequence of a hemolytic process, i.e. when important amounts of bilirubin and iron are released. The aim was to develop an animal model to determine the role of bilirubin and iron, in the encephalopathy secondary to a hemolytic disease. Male Wistar rats 7 days old (n = 30) were treated with phenylhydrazine as hemolytic at 75 mg/kg body weight intraperitoneally for 2 days and euthanized 24 h after the last dose. Hemoglobin, hematocrit, serum and brain bilirubin, serum iron and lipoperoxidation products, as well as neuronal damage and iron positive staining were evaluated and compared among treated and untreated (n = 10) animals. The animals with induced hemolysis showed significant reduction in hemoglobin and hematocrit, increased concentration of total and conjugated bilirubin, as well as of serum iron and lipid peroxidation products. The neuronal damage in treated animals included the presence of altered neurons spread out among normal cells, as well as of iron-staining positive cells. With the use of appropriated pharmacological procedures, the characteristics of the model can be useful to dissect the participation of both bilirubin and iron, on the bilirubin encephalopathy secondary to hemolysis. © 2007 Elsevier B.V. All rights reserved. Keywords: Encephalopathy; Experimental model; Hemolysis; Bilirubin; Iron; Phenylhydrazine; Rats

1. Introduction The neuronal damage observed in newborns with hemolytic disease has been attributed to the toxicity that produces the unconjugated forms of bilirubin accumulated in specific areas of the brain. Secondary clinical manifestations of this damage compose the “bilirubin encephalopathy” and the established treatment is focused in diminishing the bilirubin blood levels. Although the toxic effect of bilirubin is unquestionable, other metabolites that have not properly been evaluated may participate in the pathogenesis of this encephalopathy. One of the

∗

Corresponding author. Tel.: +52 55 1084 3883; fax: +52 55 1084 3883. E-mail address: [email protected] (H.J. Olgu´ın).

0165-0270/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2007.09.003

consequences of hemolysis, besides bilirubin production, is free iron (Fe2+ ) release. In infants with rhesus hemolytic disease (RHD), high ferritin levels, low transferrin values, diminished latent iron-binding capacity and increased lipid-peroxidation products have been observed; these findings suggest the possibility that iron toxicity occurs in infants with severe RHD (Aygun et al., 2004; Berger et al., 1990). Bilirubin synthesis occurs through the heme catabolism, mainly. The heme group is degraded to bilirubin by an enzymatic process, in which the enzymes heme-oxigenase (HO) nicotinamide-adenine dinucleotide phosphate (NADPH)cytochrome P450 reductase, and biliverdin reductase produce biliverdin, with iron and carbon monoxide (CO) release (Kikuchi et al., 2005; Rodgers and Stevenson, 1990; Shibahara et al., 2002).

36

G.B. Mejia et al. / Journal of Neuroscience Methods 168 (2008) 35–41

To exert its toxic effects on the central nervous system, bilirubin must cross the brain–blood barrier (BBB). In this regard, it has been demonstrated that several conditions such as a decrease of the bilirubin–albumin binding capacity, or alterations in BBB functioning, favor the increment of bilirubin within the brain (Hanko et al., 2003; Hansen, 2002; Odell, 1959; Wennberg, 2000). Studies related with the role of the BBB in experimental models in rats, have shown that the permeability of the brain to bilirubin is significantly superior in animals 10 days old, that in those of 21 days. This finding suggests a higher susceptibility of immature organisms to hyperbilirubinemia (Roger et al., 1996). Additionally, it is well known that the transport of one of the products of the heme catabolism (i.e. iron) from the plasma to brain is highly regulated, and that an alteration of the BBB could favor its traffic into the central nervous system (CNS) (Burdo et al., 2003; Moos and Morgan, 2000; Wu et al., 2004; Yokel, 2006). Iron is an essential component of almost all type of cells; since it exists in two oxidation states (Fe2+ and Fe3+ ), it is reliable to participate in fundamental redox reactions within the organism. The brain also requires iron for its metabolism and may suffer alterations if there is a deficiency or an excess of this element. Neurobehavioral disruptions, as well as reduced activity, altered reflexes and retardation of rejection conditioned response are known to occur after an iron overload in rats (Sobotka et al., 1996). Administered iron to newborn mice induces lowered learning capacity and decreased motor activity at 3 month-age, due to an iron accumulation in basal ganglia, rather than in cerebral cortex (Fredriksson et al., 1999, 2000). Other studies have shown that the iron released by the activity of the heme-oxigenase is involved in lipid peroxidation in both microsomal and cellular systems (Lamb et al., 1999). Previous observations suggest that iron can play an important role as neurotoxic agent during unconjugated hyperbilirubinemia associated to hemolytic diseases (Abzhandadze et al., 2006; Aygun et al., 2004; Berger et al., 1990). In order to study the physiophatology of the bilirubin encephalopathy (BE), experimental models have been used to demonstrate the toxic effect of bilirubin, in one approach, the strategy consists in the simultaneous administration of exogenous bilirubin and drugs that displace it from albumin (i.e. sulfizoxasole, sulfadimethoxine) so that bilirubin concentration in the brain is increased (Hanko et al., 2003; Park et al., 2001; Roger et al., 1996). The other approach is to use the congenital animal model (Gunn rats) that reproduces the Crigler–Najjar syndrome, in which there is an increase in the free bilirubin concentration that produces jaundice in the animals, due to the lack of activity of the enzyme glucuronyl transferase (Ahlfors and Shapiro, 2001; Dennery et al., 1995; Shapiro, 2002). Those models have been widely used to demonstrate the toxic effect of bilirubin accumulation in nervous tissue. However, none of these models reproduce the conditions in which the hyperbilirubinemia is a consequence of an acute hemolytic process, and in which, due to the activity of the heme-oxigenase; important amounts of bilirubin and iron are released. A model with these characteristics would be of great utility to determine the role of bilirubin and iron, in bilirubin encephalopathy, secondary

to hemolytic disease. Therefore the objective of this work was focused on the creation of an experimental model that reproduces these conditions. 2. Material and methods For the creation of the experimental model, 40 male Wistar rats of 7-day old were used. According to Vannucci et al. (1999), the development of the brain of a rat at this age is histologically similar to that of a 32–34-week gestation human newborn infant. Hemolysis was induced by intraperitoneal administration of 75 mg/kg/day phenylhydrazine hydrochloride (Merck), for two consecutive days based on previous data (Hansen and Allen, 1996). Phenylhydrazine reacts readily with the carbonyl group (–C O) of different biologically important molecules. It interacts with hemoglobin and cytochrome p450 through an oxidation reaction, leading to the generation of destructive free radicals, which are responsible for subsequent hemolysis (World Health Organization Geneva, 2000). Control animals received the same volume of 0.9% saline solution. Twenty-four hours after the last administration of the drug the animals were humanely sacrificed. Half of the animals of both groups were used for histological studies and the other half for biochemical analyses. For the measurement of biochemical indicators the animals were sacrificed by decapitation, blood samples were collected and the brain was extracted and frozen at −70 ◦ C until use. For the histological studies, animals were sacrificed under general anesthesia (sodium Phenobarbital 25 mg/kg/ip), and transcardially perfused with 10% formalin in 0.1 M phosphate buffer, pH 7.2. Their brains were extracted and fixed in 10% formalin until processed. In order to verify the BBB permeability in 7 days old animals, Evans blue (2%) (Saija et al., 1997) 4 ml/kg/iv was administered under general anesthesia to a group of newborn animals (n = 3) and adults (6 months old) (n = 3). Twenty minutes after administration, the animals were perfused with 0.9% saline and their brains extracted. Slices 5 ␮m thick, obtained in fresh with a cryomicrotome at −15 ◦ C, were evaluated under the fluorescence microscope to determine the presence and distribution of the Evans blue in the cerebral tissue. Hemolysis severity indicators were the hemoglobin (Hb) level and the hematocrit (Ht), which were determined by automated methods. Hemoglobin was quantified spectrophotometrically as cyanomethemoglobin (Beckman-Coulter LH-750) and the hematocrit was determined indirectly from a previous red blood cell count. The measurement of total and direct bilirubin was carried out using the Jendrassik-Grof method. A concentration of serum total bilirubin higher than 3 mg/dl was considered as an index of hyperbilirubinemia (Dominguez et al., 1997). Brain bilirubin levels were measured following the method reported by Sawasaki et al. (1976). Briefly, brain tissue was homogenized in 3 volumes of a solution of 0.25 M sucrose and bilirubin extracted using a mixture of 4 ml distilled water, 5 ml chloroform and 12 ml methanol. The mixture was then centrifuged at 10,000 × g for 30 min, and the bilirubin suspended in the chloroform phase was determined spectrophotometrically

G.B. Mejia et al. / Journal of Neuroscience Methods 168 (2008) 35–41

by difference of readings at 452 and 490 nm, using commercial bilirubin (Sigma) as standard. Blood iron determination directly after sacrificing the animals was carried out with a Randox SI250 kit, based on the acidinduced ferric iron dissociation from blood transferrin, followed by iron reduction to its ferrous state, which forms a complex with a chromogen absorbing at 595 nm. As indicator of oxidative stress, lipid peroxidation products were measured in the brain of the treated animals and compared with those of the control group. The technique for the determination of thiobarbituric acid reactive substances (TBARS) consisted, in 30 min boiling of 1 ml brain homogenate added to 2 ml of a mixture of 0.375 g thiobarbituric acid, 15 g trichloroacetic acid plus 2.5 ml of concentrated hydrochloric acid, diluted in 100 ml of water. After cooling, the sample was centrifuged at 3000 × g for 15 min, and the supernatant was read in the spectrophotometer at 532 nm. The results are expressed as micromoles of malondialdehyde per milligram of protein (Gutteridge and Halliwell, 1990). The brains of the animals were fixed in 10% formalin/0.1 M phosphate buffer, pH 7.2, and were included in paraffin. Two expert pathologists, blinded to the slides observed, evaluated the stained sections to observe the presence or absence of neuronal damage (angular retraction, cell fragmentation, neuronofagia) as well as acidophilia and reactive glyosis in different sections of the brain: cerebral cortex, hippocampus, basal nuclei, ependimal cells and choroid plexus. Both control and experimental individuals were examined. The samples were stained using the method of Perl’s modified by Smith et al. (1997) to visualize the presence of iron: slices 6 ␮m thick were incubated during 15 min in 7% potassium ferricyanide diluted in 3% hydrochloric acid, and subsequently in a solution of 3,3 diaminobenzidine (0.75 mg/ml) and 0.015% hydrogen peroxide, for 5–10 min, in order to intensify the reaction. Treatment-blinded personnel evaluated the slices by means of light microscopy. The representation of the data was made using graphics with absolute and mean ± standard deviation values. For inference analysis the strategy consisted in comparing the hemoglobin, hematocrit, serum and brain bilirubin levels, iron concentration and thiobarbituric reactive substances between groups, by the Student-t test for independent samples, or by means of the Mann–Whitney U-test. Probability values <0.05 were considered statistically significant. 3. Results Fluorescence microscopy evaluation of Evans Blue extravasations in the brain of mature and immature animals showed that the brain–blood barrier remains permeable in the 7 days old rats (Fig. 1). As a consequence of the induction of hemolysis, a significant reduction in both hemoglobin and hematocrit was observed for the experimental group. A significant increment in the concentration of total and direct serum bilirubin is observed in the experimental group as compared with the control group. And even when the difference was not statistically significant, there was an increment of indirect bilirubin in the

37

Fig. 1. Evans blue administration shows that in mature animals (a) BBB is not permeable. When the encephalic slices were analyzed by fluorescence microscopy the staining was confined only to the blood vessels (black arrow). In the 7 days old animals (b) Evans blue extravasations was observed staining the intercellular space (white arrow) and the neuronal nuclei (black arrow). This shows that at this age the BBB is permeable.

experimental group. The concentration of bilirubin in the brain was slightly higher in experimental group than in the control group, but the difference was not statistically significant. The animals treated with drug showed a significant increase in the serum iron, and significantly higher level of lipid peroxidation was observed in the experimental animals as compared to the control group (Table 1). It was possible to visualize deposits of iron in the brain tissue of the animals of the experimental group. In the choroid plexus, abundant iron staining appeared, with preponderance in the basal membrane. The stain was observed in the cytoplasm of the ependymal cells; its distribution was irregular and, in general, the neighboring cells did not present alterations (Fig. 2). In the external granular layer of the cortex, neurons with positive nuclear iron staining were observed. These cells had lost its normal appearance; the neurons were angular in shape and positive Fe neurons appeared alternating with cells of the same type that did not present Fe staining (Fig. 3). The glial cells were negative to any mark. The neuropile was observed disorganized. At the hippocampus very few neurons presented the alterations described for cerebral cortex; no Fe stain appeared in the ependymal cells. In the brain of the control animals the neurons presented normal morphology, i.e. with well-defined borders and nuclear

38

G.B. Mejia et al. / Journal of Neuroscience Methods 168 (2008) 35–41

Table 1 Values of some biomarkers to determine the effect of phenylhydrazine in Wistar rats (mean ± standard deviation) Control mean ± s.d. Hemoglobin (g/dl) Hematocrit (%) Total serum bilirubin (mg/dl) Direct serum bilirubin (mg/dl) Indirect serum bilirubin (mg/dl) Brain bilirubin (␮g/mg protein) Serum iron (mg/dl) Lipid peroxidation (␮mol malondialdehyde) a

11.07 34 1.44 0.53 1.1 0.108 0.012 0.025

± ± ± ± ± ± ± ±

1.34 2.71 0.5 0.13 0.36 0.03 0.009 0.002

PHZ treated mean ± s.d. 2.84 9.42 5.04 3.23 1.8 124 0.44 0.050

± ± ± ± ± ± ± ±

0.75 2.4 1.86 1.97 1.21 0.07 0.08 0.009

Significancea P < 0.00001 P < 0.00001 P < 0.0001 P < 0.0001 P = 0.068 P = 0.65 P < 0.0001 P < 0.0001

After Student-t test or Mann–Whitney U-test.

membrane, fine chromatin, evident nucleolus, homogeneous neuropile, and unaltered glial cells (Fig. 4). The cerebral cortex and hippocampus cellular layers, the basal nuclei, the ependymal cells, and the choroid plexus were well preserved. In contrast, in the brain of the animals with experimentally induced hemolysis, damaged neurons were observed, which presented a process of gradual acidophilia, as evidenced by darkening of the nucleus and cytoplasm, followed by contraction of the pericaryon, and finally loss of the nuclear limits and cytoplasmatic membrane with total cellular disorganization. The neuropile appeared vacuolated in a degree generally related to tissue disorganization. Neurons without apparent alterations appeared alternating with those damaged (Fig. 5). In Table 2 a semi-quantitative summary of the main neuronal damage findings in animals treated with phenylhidrazine is showed. Injury gradation was done using morphologic semi-quantitative criteria by the presence or absence of neuronal damage as described in material and methods in 20 serial slices by subject. Presence of Iron staining positive cells in damaged areas was semi-quantified. Almost imperceptible damage or iron stained cells were observed in control subjects.

that include exogenous administration of bilirubin, pharmacological displacement from albumin and artificial induction of the blood–brain barrier opening (Hanko et al., 2003; Park et al., 2001; Roger et al., 1996), or the model of Gunn rat, which reproduces the Crigler–Najjar syndrome (Ahlfors and Shapiro, 2001; Dennery et al., 1995; Shapiro, 2002). As we pointed out previously, heme catabolism produces bilirubin and other metabolites (i.e. iron and carbon monoxide) whose participation in the bilirubin encephalopathy have not properly been evaluated. As it is well documented, the increased levels of serum bilirubin and its presence in brain specific areas, may be the cause of the

4. Discussion Most of the studies designed to demonstrate the toxic effect of the unconjugated form of bilirubin when accumulated in specific areas of the brain, use experimental models with procedures

Fig. 2. Iron-positive choroid plexus cells (black arrow), unstained cells are observed (white arrow), morphological changes are not observed.

Fig. 3. (a) Iron-positive cerebral cortex cells (black arrow); among unstained cells (white arrow), stained cells show morphological alterations. (b) Pyramidal neurons with positive iron stain (black arrow) alternating with not stained cells (white arrow) in the brain of phenylhydrazine treated rats. The cells show cytoplasmatic condensation, nuclear and cytoplasmatic disorganization, and perycarion deformation with irregular borders. Also, vacuolated neuropile is present. Modified Perl’s–thionin, 800×.

G.B. Mejia et al. / Journal of Neuroscience Methods 168 (2008) 35–41

Fig. 4. Control animal cerebral cortex cells are well defined (white arrow); neuropile (black arrow) with no alterations. Significant morphological alterations are not observed.

Fig. 5. Pycnotic pyramidal neurons (black arrow) are found in the brain of phenylhydrazine treated rats. These cells show nuclei and cytoplasm disorganization. The neuropile appears vacuolated (white arrow). Neurons without apparent alteration (diamond arrow) appear alternating with damaged neurons. Acid fuchsin–thionin, 800×.

neuronal damage observed in hyperbilirubinemic individuals. Studies in individuals with Rh factor incompatibility-induced hemolytic disease (Aygun et al., 2004; Berger et al., 1990) lead us to suggest that the heme group metabolism-derived iron atom

39

may play a critical role in the neuronal damage induction. It is necessary to have an adequate experimental model in order to put this hypothesis to the test. However, none of the models that have been used may achieve this purpose, since both are based only in bilirubin concentration increase and in the process by which it reaches the brain. The desirable characteristics of the model include blood–brain barrier permeability, increased concentration of the metabolites produced by the catabolic activity of the hemeoxigenase, evidence of oxidative stress attributable to free iron release and neuronal damage compatible with bilirubin encephalopathy. The establishment of an animal model in which the BBB is naturally permeable avoids the need for this additional variable and thus, brings the model closer to the circumstances in which the phenomenon of bilirubin encephalopathy, secondary to hemolysis, occurs. A significant increment in the concentration of total and direct bilirubin in the animals of the experimental group was observed in the present work. Remarkable is the fact that the concentration of direct bilirubin was higher than that of indirect. This situation was contrary to what is reported in the literature. It is well known that an acute hemolysis leads to a significant bilirubin increase, and the impairment in bilirubin production/conjugation leads to an increase in free or unconjugated bilirubin (Porter and Dennis, 2002). In search of a possible explanation, we evaluated samples of hepatic tissue to determine if the presence of an obstructive process was responsible for such a situation. The results indicated that it is not, since there were no cholestasis signs. Erythrophagocytosis, as a consequence of the hemolytic process was present, as well as steatosis, but these alterations are not indicative of severe hepatic damage (Fig. 6). The activity of the enzyme glucuronyl transferase in the rat follows a pattern of activation that depends on its development; its activity is low in the neonatal period and it is increased until adult values (Cantarino et al., 2002), a very similar pattern to that in humans. Thus, an explanation based on differences between the rat and the human hepatic function could not be the appropriate, but the existence of an extra hepatic

Table 2 Semi-quantitative summary of the main neuronal damage findings in animals treated with phenylhidrazine Brain areas

Degree of injury

Iron staining positive cells

Amygdala Olfactory cortex Geniculate nucleous Motor fronto parietal cortex Sensorial fronto parietal cortex Hippocampus Thalamus Hipothalamus Caudate putamen Choroid plexus

L L L M S M L M L L

* * * ** *** ** * ** * ***

Scores are L = low, M = moderate, S = severe. Presence of iron staining positive cells in damaged areas was semi-quantified (*) scanty, (**) medium, (***) abundant.

Fig. 6. Liver slide showing parenchymal normal structure, no signs of cholestasis, hemolysis-derived erythrophagocytosis (solid black arrow) and diffused steatosis (dot black arrow).

40

G.B. Mejia et al. / Journal of Neuroscience Methods 168 (2008) 35–41

conjugation in the rat could contribute to this phenomenon. In this regard, Mottino et al. (1988) reports that, the glucuronidation of the bilirubin occurred in the renal and intestinal tissues of the rat is as important as that of the hepatic tissue in vitro. Additionally Li et al. (2004) pointed out that the unconjugated bilirubin can be transformed to a conjugated form in extra hepatic tissues in the rat. When the glucuronidation capacity of the liver is saturated due to administration of a high dose of acetaminophen (150 mg/kg), a dramatic increment in direct bilirubin, attributable to extra hepatic conjugation, is observed. This would support the hypothesis that the concentration of direct bilirubin observed in the experimental animals can be a result of indirect bilirubin conjugation by other tissues. The most important parameter to determine risk of suffering bilirubin encephalopathy derived from a hemolytic process is indirect bilirubin elevation. Based on the results obtained, the conclusion is that this requirement is not fulfilled in the model. However, it is important to make some considerations: The albumin binding of indirect bilirubin is a factor of great importance for the development of hyperbilirubinemia. According to Ahlfors (2000) the concentration of unconjugated bilirubin, not bound to albumin associated to the probability of kernicterus occurrence, oscillates between 0.86 and 1.19 ␮g/dl. The use of drugs that displace the bilirubin from albumin increases the possibility that bilirubin gets into the brain. Taking into consideration these findings it is important to note that the model induced enough concentration of indirect bilirubin, to induce hyperbilirubinemia by pharmacological means. It is important to emphasize that this is in fact one of the procedures used to dissect the participation of the bilirubin during encephalopathy. In addition, the fact that the encephalic concentration of bilirubin between animals with hemolysis induction and controls is not different is a favorable condition for the introduction of pharmacological management directed to increase its levels, so that it is possible to evaluate the neuronal damage that takes place when the levels of unconjugated bilirubin are increased. The increment that is observed in the bilirubin levels, together with the increase in the concentration of iron, indicates a high catabolic activity of the heme-oxigenase in the animals with induced experimental hemolysis. In previous paragraphs we have described the process of bilirubin synthesis, the role of the heme-oxigenase in this process, as well as the participation of the iron released in the generation of oxidative stress. These circumstances are of importance in the proposed model, since the hypothesis that we seek to test arises from the premise that the metabolites derived from heme catabolism, could play an important role as inductors of neuronal damage under conditions in which bilirubin encephalopathy is the result of a hemolytic process. The neuronal damage that we observe in the animals subjected to experimental hemolysis is similar to that reported by Ahdab-Barmada (1984) in autopsies of infants who died of kernicterus: the neuropile appears spongy, the cells show higher affinity for the colorant, the chromatin appears diffuse, there is disorganization of the cell, the limits of the nucleus are lost and apparently normal neurons spared among damaged cells are observed.

With the use of appropriated pharmacological procedures, the characteristics of the model can be useful to dissect the participation of both bilirubin and iron, on the bilirubin encephalopathy secondary to hemolysis. References Abzhandadze TI, Kvezereli-Kopadze AN, Dzhaparidze ES, Tkhilava NG, Sanikidze TV. The role of oxidative metabolism in pathogenesis of hyperbilirubinemia in infants. Georgian Med News 2006;136:77–80. Ahdab-Barmada M, Neonatal kernicterus: neuropathologic diagnoses. In: Hyperbilirubinemia in the newborn. Report of the Eighty-Fifth Ross Conference on Pediatric Research. Columbus, Ohio, USA: Ross Laboratories, 1984, p. 2. Ahlfors CE. Unbound bilirubin associated with kernicterus: a historical approach. J Pediatr 2000;137:540–4. Ahlfors CE, Shapiro SM. Auditory brainstem response and unbound bilirubin in jaundiced (jj) Gunn rat pups. Biol Neonate 2001;80:158–62. Aygun C, Tekinalp G, Gurgey A. Increased fetal iron load in rhesus hemolytic disease. Pediatr Hematol Oncol 2004;4:329–33. Berger HM, Lindeman JHN, van Zoeren-Grobben D, Houdkamp E, Schrijver J, Kanhai HH. Iron overload, free radical damage, and rhesus haemolytic disease. Lancet 1990;335:933–6. Burdo JR, Antonetti DA, Wolpert EB, Connor JR. Mechanisms and regulation of transferrin and iron transport in a model blood–brain barrier system. Neuroscience 2003;121:883–90. Cantarino M, Arahuetes R, Arza E, Cubero F, Garc´ıa-Barrutia M, Ortiz A. Evoluci´on de la actividad de la Bilirrubina-UDP Glucuronosil-Transferasa (bUDP-GT) durante el desarrollo en rata. Ann Biol 2002;24:155–60. Concise International Chemical Assessment Document 19 Phenylhydrazine. Geneva: World Health Organization; 2000. Dennery PA, McDonagh AF, Spitz DR, Rodgers PA. Hyperbilirubinemia results in reduced oxidative injury in neonatal Gunn rats exposed to hyperoxia. Free Radic Biol Med 1995;19:395–404. Dominguez F, Gonz´alez G, Cedars M, Calvo J. Apertura reversible de la barrera hemato-encef´alica inducida por hipercapnia en hiperbilirrubinemia experimental. Ann Espa˜noles Pediatr 1997;46:374–7. Fredriksson A, Schroder N, Eriksson P, Izquierdo I, Archer T. Neonatal iron exposure induces neurobehavioural dysfunctions in adult mice. Toxicol Appl Pharmacol 1999;159:25–30. Fredriksson A, Schroder N, Eriksson P, Izquierdo I, Archer T. Maze learning and motor activity deficits in adult mice induced by iron exposure during a critical postnatal period. Brain Res Dev Brain Res 2000;119:65–74. Gutteridge J, Halliwell B. The measurement and mechanism of lipid peroxidation in biological systems. Trends Biochem Sci 1990;15:129–35. Hanko E, Tommarello S, Watchko JF, Hansen TW. Administration of drugs known to inhibit P-glycoprotein increases brain bilirubin and alters the regional distribution of bilirubin in rat brain. Pediatr Res 2003;54:439–40. Hansen TW. Mechanisms of bilirubin toxicity: clinical implications. Clin Perinatol 2002;29:765–78. Hansen TWR, Allen JW. Hemolytic anemia does not increase entry into, nor alter rate of clearance of bilirubin from rat brain. Biol Neonate 1996;69:268–74. Kikuchi G, Yoshida T, Noguchi M. Heme oxygenase and heme degradation. Biochem Biophys Res Commun 2005;338:558–67. Lamb NJ, Quinlan GJ, Mumby S, Evans TW, Gutteridge JM. Haem oxygenase shows pro-oxidant activity in microsomal and cellular systems: implications for the release of low-molecular-mass iron. J Biochem 1999;344:153–8. Li XD, Xia SQ, Lv Y, He P, Han J, Wu MC. Conjugation metabolism of acetaminophen and bilirubin in extrahepatic tissues of rats. Life Sci 2004;74:1307–15. Moos T, Morgan EH. Transferrin and transferrin receptor function in brain barrier systems. Cell Mol Neurobiol 2000;20:77–95. Mottino AD, Pellegrino JM, Guibert EE, Roma MG, Rodriguez Garay EA. Comparison of hepatic, renal and intestinal bilirubin UDP-glucuronyl transferase activities in rat microsomes. Int J Biochem 1988;20:1113–6. Odell GB. The dissociation of bilirubin from albumin and its clinical implications. J Pediatr 1959;55:268–79.

G.B. Mejia et al. / Journal of Neuroscience Methods 168 (2008) 35–41 Park WS, Chang YS, Chung SH, Seo DW, Hong SH, Lee M. Effect of hypothermia on bilirubin-induced alterations in brain cell membrane function and energy metabolism in newborn piglets. Brain Res 2001;922:276–81. Porter ML, Dennis BL. Hyperbilirubinemia in the term newborn. Am Fam Physician 2002;65:599–606. Roger C, Koziel V, Vert P, Nehlig A. Autoradiographic mapping of local cerebral permeability to bilirubin in immature rats: effects of hyperbilirubinemia. Pediatr Res 1996;39:64–71. Rodgers PA, Stevenson DK. Developmental biology of heme oxigenase. Clin Perinatol 1990;17:275–91. Saija A, Princi P, Trombetta D, Lanza M, De Pasquale A. Changes in the permeability of the blood–brain barrier following sodium dodecyl sulphate administration in the rat. Exp Brain Res 1997;115:546–51. Shapiro SM. Somatosensory and brainstem auditory evoked potentials in the Gunn rat model of acute bilirubin neurotoxicity. Pediatr Res 2002;52:844–9. Sawasaki Y, Yamada N, Nakajima H. Developmental features of cerebellar hypoplasia and brain bilirubin levels in a mutant (Gunn) rat with hereditary hyperbilirubinaemia. J Neurochem 1976;22:577–83.

41

Shibahara S, Kitamuro T, Takahashi K. Heme degradation and human disease: diversity is the soul of life. Antioxid Redox Signal 2002;4:593–602. Smith M, Harris PL, Sayre L, Perry G. Iron accumulation in Alzheimer disease is a source of redox-generated free radicals. Proc Natl Acad Sci USA 1997;94:9866–8. Sobotka TJ, Whittaker P, Sobotka JM, Brodie RE, Quander DY, Robl M, et al. Neurobehavioral dysfunctions associated with dietary iron overload. Physiol Behav 1996;59:213–9. Vannucci R, Connor J, Mauger D, Palmer C, Smith M, Towfighi J. Rat model of perinatal hypoxic-ischemic brain damage. J Neurosci Res 1999;55:158–63. Wennberg RP. The blood–brain barrier and bilirubin encephalopathy. Cell Mol Neurobiol 2000;20:97–109. Wu LJ, Leenders AG, Cooperman S, Meyron-Holtz E, Smith S, Land W, et al. Expression of the iron transporter ferroportin in synaptic vesicles and the blood–brain barrier. Brain Res 2004;1001:108–17. Yokel RA. Blood–brain barrier flux of aluminum, manganese, iron and other metals suspected to contribute to metal-induced neurodegeneration. J Alzheimers Dis 2006;10:223–53.