Hammerman, C.*; Goldschmidt, D.*; Caplan, M. S.†; Kaplan, M.‡§; Bromiker, R.‡; Eidelman, A. I.*; Gartner, L. M.¶; Hochman, A.**
Neonatal hyperbilirubinemia remains a teleologic enigma. Although it is clear that bilirubin can cross the blood–brain barrier and can cause irreversible brain damage and even death, it is also true that bilirubin possesses antioxidant properties that can be protective in the face of oxidative stress.
Although increased serum bilirubin concentrations have been correlated with total antioxidant potential and bilirubin antioxidant properties have been demonstrated to be protective in a rat model exposed to hyperoxia (1), evidence of bilirubin-associated in vivo protective effect against oxidative stress remains nonspecific. Human clinical studies must, by definition, be clinical surveys showing at best an association between hyperbilirubinemia and clinical outcome of reactive oxygen species–mediated pathology. The results of such studies have been understandably contradictory. First, clinical disease is multifactorial, and the role of bilirubin may be concealed behind multiple other risk factors. With this in mind, we sought to investigate whether jaundice is protective, in an in vivo animal model, against ischemic reperfusion injury to the intestine. Second, although it might be expected that hyperbilirubinemia would be associated with improved antioxidant potential and thus improved outcome, bilirubin may also be consumed during the process, and thus patients with better outcomes may be associated with lower rather than higher bilirubin concentrations.
This model was chosen because the intestine is extremely sensitive to oxidative stress and thus would appear to be a good test of bilirubin's antioxidant properties. Ischemia predisposes the gut to subsequent injury. With reperfusion and the reintroduction of oxygen, reactive oxygen species are generated. It is these potentially toxic reactive oxygen species that, in turn, promote cell destruction and bowel necrosis via peroxidation of membrane lipids (2–4) and that potentially could be inhibited by bilirubin.
Young adult rats were anesthetized with 60 mg/kg intraperitoneal pentobarbital. The intestinal vasculature was exposed via a midline surgical incision, and the superior mesenteric artery was isolated. The abdominal wall was then reapproximated and secured with clamps to prevent evaporative losses, while maintaining easy access for observation and further intervention. Thirty minutes after surgical preparation, baseline blood samples were obtained. Blood samples were obtained from the inferior vena cava by micropuncture using 25-gauge needles. The rats were then randomly assigned to one of the following three groups: 1) group 1, ischemia–reperfusion (IR)/control: after preparation as described above, flow in the superior mesenteric artery (which had been surgically isolated earlier) was occluded with a mosquito clamp; after 45 minutes of occlusion, perfusion was reestablished by opening the clamp; 2) group 2, IR/hyperbilirubinemia: same as IR protocol described above, with a continuous infusion of bilirubin; 3) group 3, hyperbilirubinemia/control (no IR): sham surgical preparation as described, but these animals were not subjected to IR; bilirubin was infused as described above.
Blood samples (1 mL) were taken from all animals in all groups at baseline (30 minutes after completion of surgery), at the end of 45 minutes of ischemia for IR groups (75 minutes after surgery), and after 15 and 30 minutes of reperfusion for the IR groups (90 and 120 minutes after surgery). Samples were immediately frozen in liquid nitrogen and stored at −80°C. These samples were subsequently assayed for thiobarbituric acid reducing substances (TBARS), a measure of lipid peroxidation and for serum bilirubin concentrations. At the conclusion of the experiment, intestinal tissue samples were taken from the distal ileum, immediately frozen in liquid nitrogen, and stored at −80°C for measurement of tissue xanthine oxidase (XO)/xanthine dehydrogenase (XD) and reduced glutathione. Other intestinal samples were obtained and fixated in formalin for histologic examination with hematoxylin and eosin staining. This study was approved by the institutional review board of the Shaare Zedek Medical Center for the care of animals using local guidelines.
Bilirubin was dissolved in 0.1 N NaOH, stabilized with 5% bovine serum albumin, and then diluted with 0.055 mol/L sodium phosphate buffer to a final pH of approximately 8.0 (final concentration of approximately 7 mg/4.5 mL). Each rat selected for hyperbilirubinemia was given a loading dose of 15 mg/kg bilirubin intravenously over 10 minutes, followed by 30 mg/kg intravenously over the ensuing 75 minutes.
Lesions were graded by a blinded observer based on hematoxylin and eosin–stained histologic slides and were assigned a necrotizing enterocolitis (NEC) score on a scale of 0 to 4 as follows: 0 = normal, intact villous epithelium with normal histology; 1 = mild villous edema, with epithelial sloughing confined to the tips of the villi; 2 = mild midvillous necrosis; 3 = moderate midvillous necrosis, with crypts still readily detectable; and 4 = severe necrosis of entire villi with complete absence of epithelial structures.
Thiobarbituric Acid Reacting Substances (TBARS)
Blood (300 μL) was collected in EDTA anticoagulant, snap frozen in liquid nitrogen, and then maintained frozen at −80°C pending analysis of TBARS (5).
Xanthine Oxidase (XO)/Xanthine Dehydrogenase (XD)
Intestinal tissue (500 mg) was homogenized on ice in 0.03 mol/L potassium phosphate buffer (pH 7.4) containing 1 mmol/L dithiothrietol to prevent oxidative conversion of XD to XO in solution, 1 mmol/L EDTA, and the protease inhibitors, 1 mmol/L phenylmethylsulfonyl fluoride and 1 mg/ml leupeptin. The homogenate was then centrifuged at 110,000 x g for 20 minutes at 4°C. The supernatant collected and centrifuged a second time for 10 minutes, and activity was assayed as soon as possible to avoid further changes in the enzymes. XO and XD activities were measured spectrophotometrically and recorded as described (6), and the ratio of XO to XD activity was calculated.
Glutathione/glutathione disulfide ratios (GSSG/GSH)
Glutathione/glutathione disulfide ratios were assayed with a Calbiochem kit (Calbiochem; Novabiochem, La Jolla, CA, U.S.A.). Protein was determined by the Bradford method (7), a rapid and sensitive method for the quantification of microgram quantities of protein using the principle of protein dye binding with bovine serum albumin as standard.
Analysis of Data
Results are expressed as mean ± SD. Two-way analysis of variance with repeated measures was used to determine the effects of time, group, and group–time interaction for serial TBARS. One-way analysis of variance was used to compare values of single variables among groups. Significant differences were accepted at P < 0.05. Post hoc comparisons were conducted with Tukey pairwise comparisons. Paired t tests were used to compare values with baseline values within each group. Statistical analysis was conducted with Sigmastat for Windows.
Thirty-two rats were studied: 12 subjected to IR (IR controls), 10 subjected to IR with hyperbilirubinemia (IR/hyperbilirubinemia), and 10 not subjected to IR but with hyperbilirubinemia (hyperbilirubinemia controls; subjected to a sham surgical procedure similar to that of the IR group). Mean weights were 246 g ± 53 g, 213 g ± 37 g, and 205 g ± 18 g for the IR control, IR/hyperbilirubinemia, and hyperbilirubinemia control groups, respectively. Although there was a trend toward lower birth weight in the hyperbilirubinemia control group, these differences were not significant (P = 0.05). By design, mean serum bilirubin concentrations at the conclusion of the procedure were significantly different in the IR control group than in the two hyperbilirubinemic groups, which were similar to each other (< 1 mg/dL ± 0.4 mg/dL, 17 mg/dL ± 4 mg/dL, 15 mg/dL ± 5 mg/dL, respectively;P < 0.001; 1 vs. 2; 1 vs. 3). In both hyperbilirubinemic groups, serum bilirubin steady states (≥ 13 mg/dL) were achieved from 15 minutes after baseline and continued throughout the experiment.
Histopathology: NEC Scores
All of the IR control animals had evidence of severe injury, with a mean score of 4.0 ± 0.0. Animals subjected to IR while hyperbilirubinemic had less severe injury, with a mean score of 2.8 ± 0.4 (Table 1). Hyperbilirubinemic animals without IR had predominantly normal intestinal architecture, with a histopathologic score of 0.7 ± 0.8. These differences were significant at P < 0.0001, with subsequent Tukey pairwise comparisons showing significant differences among the groups. Representative histopathologic specimens taken from the animals are presented in Figure 1.
Thiobarbituric Acid Reducing Substances
Circulating thiobarbituric acid reducing substances
Figure 2 depicts serial values for TBARS (presented as % baseline) for each of the groups at baseline and after 45 minutes of ischemia and 15, 30, and 45 minutes of reperfusion. We found significantly increased TBARS, indicative of increased lipid membrane peroxidative damage, only in the IR control animals. This finding is consistent with the previously described histopathologic observations.
Tissue thiobarbituric acid reducing substances
Tissue TBARS, like circulating TBARS, were significantly and extremely elevated in the IR control group compared with either of the hyperbilirubinemic groups (Table 1). TBARS in the hyperbilirubinemic rats were similar with and without IR.
XO/XD ratios were not significantly different across the groups, although there was a trend toward higher ratios in the IR group compared with the control group (Table 1). Glutathione/glutathione disulfide ratios were reduced in the IR control animals compared with both other groups.
Reactive oxygen species are generated continuously during physiological conditions. These potentially toxic reactive oxygen species are highly reactive substances that promote cell damage via lipid membrane peroxidation. Once lipid peroxidation is initiated, it self-propagates until terminated by some antioxidant. The mammalian body has an extensive network of protective antioxidant defenses; however, overproduction of free radicals or inadequate antioxidant defenses can disrupt the natural balance, leading to oxidative damage.
Bilirubin is one of the body's natural antioxidants, contributing up to 10% to 30% of the total antioxidant status of premature infants (8,9). Furthermore, during the first postnatal days when general antioxidant defenses are reduced, serum bilirubin is physiologically increased, conceivably implying some beneficial role for this physiological hyperbilirubinemia (9).
It has been proposed that hyperbilirubinemia might protect against oxidative injury. To date, human studies of bilirubinemia and oxygen radical–associated diseases have been retrospective and observational. Because clinical diseases are multifactorial and potential risk factors cannot be controlled for in retrospective studies, results remain contradictory and inconclusive. Some investigators (10,11) have noted lower bilirubin concentrations in neonates with oxygen radical–mediated diseases as compared with controls, implying that bilirubin may be consumed in response to oxygen free radicals. Conversely, a retrospective study of early bilirubin concentrations in relation to the subsequent development of retinopathy of prematurity (12) did not show any such correlation. By definition, such studies are hampered by the fact that any relation between clinical disease and bilirubin is likely to be more complex than a direct correlation. Oxidative stress may consume available antioxidant resources, or, conversely, it may induce antioxidant release (13) or some combination thereof.
Animal studies offer the opportunity to focus investigations on specific factors and to limit confounding variables. In an interesting study by Clark et al. (14), increased endogenous bilirubin production was shown to provide cardioprotection against postischemic–reperfusion injury. Although consistent with our findings, there are several important differences between our study and that of Clark et al. (14). First, heme oxygenase induction also leads to release of carbon monoxide, a vasoactive compound that may contribute to observed vascular effects. Second, this study dealt with an isolated heart preparation, not with an intact animal model and not with intestinal lesions. In another study, Dennery et al. (1) demonstrated that Gunn rats exposed to hyperoxia had less oxidative damage, as evidenced by reduced concentrations of lipid peroxides, conjugated dienes, and carbonyl proteins, than did nonjaundiced rats. However, in homozygous Gunn rats, serum bilirubin concentrations reach approximately 8 mg/dL, considered physiologic in term human neonates. We therefore felt it was important to extend these studies of the protective bilirubin effects in an intact animal model to higher bilirubin concentrations.
The intestinal mucosa is known to be exquisitely sensitive to IR injury. Its baseline high rate of oxygen use renders the intestine relatively incapable of increasing oxygen transport in the face of hypoxic stress and thus more susceptible to ischemic injury. Intestinal ischemia predisposes the gut to subsequent necrosis. Evidence suggests, however, that while sustained blood flow reduction damages the intestine mucosa, restoration of flow and reintroduction of oxygen after deprivation accelerates tissue injury (2).
As predicted, there was significantly less histopathologic and biochemical evidence of damage in hyperbilirubinemic animals than in controls when exposed to similar degrees of IR. Microscopically, there were diffuse necrotic lesions with severe villous necrosis throughout the intestines of the IR control animals, as compared with significantly milder damage in the IR hyperbilirubinemia group. TBARS were significantly also significantly elevated in IR control animals, further suggesting that hyperbilirubinemia protected against lipid membrane peroxidation (15,16).
Xanthine oxidoreductase is an enzyme that plays a major role in the generation of toxic superoxide radicals. It has two forms—the dehydrogenase and the oxidase—which coexist in vivo. During normal physiologic conditions, the dehydrogenase predominates. During ischemia, XDe is converted into XO, which then becomes a major source for toxic superoxide radical production once the tissues are reperfused and oxygen is reintroduced (17). In most species, the intestine is one of the richest sources in the body of XO, an enzyme that exhibits a gradient of increasing activity from the intestinal villous bases to the tips. This gradient might explain the fact that the villous tips of the intestine are more susceptible to postischemic damage than are the bases (18), as seen in both our model and in clinical necrotizing enterocolitis. Because ischemia increases the relative predominance of the oxidase form of the enzyme, an elevated XO/XD ratio indicates ischemic insult. In our animals, there was a trend toward an increase in this ratio in the IR controls; however the differences were not significant.
Because it is recyclable and abundant in many cell types, glutathione is an important nonenzymatic antioxidant. Peroxidation of endogenous lipids leads to the conversion of reduced glutathione to glutathione disulfide (19,20). Glutathione-depleted tissues are more susceptible to subsequent oxidative stress. This is evident in our IR control animals as compared with the other two groups.
In summary, we have confirmed that intestinal IR generates severe histopathologic and biochemical injury associated with increased lipid peroxidation. We have further shown in vivo, in our animal model, that hyperbilirubinemia acts as an antioxidant in ameliorating the extent of this injury, both histopathologically and biochemically. This study adds support to the concept that bilirubin may possess some beneficial as well as toxic properties. Any potential clinical relevance of these data to the premature neonate must, however, take into consideration the fact that serum bilirubin concentrations such as those generated in the current study can be neurotoxic in the human premature. Clearly, no clinical conclusions can be drawn from these data, pending further study.
1. Dennery PA, McDonagh AF, Spitz DR, et al. Hyperbilirubinemia results in reduced oxidative injury in neonatal Gunn rats exposed to hyperoxia. Free Radic Biol Med 1995; 19:395–404.
2. Schoenberg MH, Berger HG. Reperfusion injury after intestinal ischemia. Crit Care Med 1993; 21:1376–86.
3. Saugstad OD. Mechanisms of tissue injury by oxygen radicals: Implications for neonatal disease. Acta Paediatr 1996; 85:1–4.
4. Parks DA, Bulkley GB, Granger DN, et al. Ischemic injury in the cat small intestine: Role of superoxide radicals. Gastroenterology 1982; 82:9–15.
5. Ernster L, Nordenbrand K. Microsomal lipid peroxidation. Methods Enzymol 1967; 10:574–80.
6. Kooij A, Schiller HJ, Schijns M, et al. Conversion of Xanthine dehydrogenase into xanthine oxidase in rat liver and plasma at the onset of reperfusion after ischemia. Hepatology 1994; 19:1488–95.
7. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976; 72:248–54.
8. Frei B, Stocker R, Ames BN. Antioxidant defenses and lipid peroxidation in human blood plasma. Proc Natl Acad Sci U S A 1988; 85:9748–52.
9. Hammerman C, Goldstein R, Kaplan M, et al. Bilirubin in the premature: Toxic waste or natural defense? Clin Chem 1998; 44:2551–3.
10. Benaron DA, Bowen FW. Variation of initial serum bilirubin rise in newborn infants with type of illness. Lancet 1991; 338:78–81.
11. Hegyi T, Goldie E, Hiatt M. The protective role of bilirubin in oxygen radical diseases of the preterm infant. J Perinatol 1994; 14:296–300.
12. Gaton DD, Gold J, Axer Siegel R, et al. Evaluation of bilirubin as a possible protective factor in the prevention of retinopathy of prematurity. Br J Ophthalmol 1991; 75:532–4.
13. Romeo MG, Tina LG, Scuderi A, et al. Variations of blood bilirubin levels in the newborn with and without retinopathy of prematurity (ROP). Pediatr Med Chir 1994; 16:59–62.
14. Clark JE, Foresti R, Sarathchandra P, et al. Heme oxygenase-1 derived bilirubin ameliorates postischemic myocardial dysfunction. Am J Physiol Heart Circ Physiol 2000; 278:H643–51.
15. Esterbauer H, Cheeseman KH. Determination of aldehydic lipid peroxidation products: Malonaldehyde and 4-hydroxynonenal. Methods Enzymol 1990; 186:407–21.
16. Wojciech W, Neve J, Peretz A. Optimized steps in fluorimetric determination of thiobarbituric acid reactive substances in serum: Importance of extraction pH and influence of sample preservation and storage. Clin Chem 1993; 39:2522–6.
17. Parks DA, Granger DN. Xanthine oxidase: Biochemistry, distribution and physiology. Acta Physiol Scand 1986; 548(suppl):87–99.
18. Granger DN, Hollwarth MA, Parks DA. Ischemic reperfusion injury: Role of oxygen derived free radicals. Acta Physiol Scand 1986; 548:47–63.
19. Hansen TN, Smith CV, Gest AL, et al. Biochemical manifestations of oxygen toxicity in the newborn lamb. Pediatr Res 1990; 28:613–7.
20. Adams Jr, JD Lauterburg BJ, Mitchell JR. Plasma glutathione and glutathione disulfide in the rat: Regulation and response to oxidative stress. J Pharmacol Exp Ther 1983; 227:749–54.
© 2002 Lippincott Williams & Wilkins, Inc.