Oxidative stress of the newborn in the pre- and postnatal period and the clinical utility of melatonin
Address reprint requests to Russel J. Reiter, Department of Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, Mail Code 7762, 7703 Floyd Curl Drive, San Antonio, TX 78229-3900, USA.
Abstract: Newborns, and especially those delivered preterm, are probably more prone to oxidative stress than individuals later in life. Also during pregnancy, increased oxygen demand augments the rate of production of reactive oxygen species (ROS) and women, even with normal pregnancies, experience elevated oxidative stress and lipid peroxidation compared with nonpregnant women. Also, there appears to be an increase in ROS generation in the placenta of pre-eclamptic women. In comparison with healthy adults, newborn infants have lower levels of plasma antioxidants such as vitamin E, β-carotene, and sulphydryl groups, lower levels of plasma metal binding proteins including ceruloplasmin and transferrin, and reduced activity of erythrocyte superoxide dismutase. This review summarizes conditions of newborns where there is elevated oxidative stress. Included in this group of conditions is asphyxia, respiratory distress syndrome and sepsis and the review also summarizes the literature related to clinical trials of antioxidant therapies and of melatonin, a highly effective antioxidant and free radical scavenger. The authors document there is general agreement that short-term melatonin therapy may be highly effective and that it has a remarkably benign safety profile, even when neonates are treated with pharmacological doses. Significant complications with long-term melatonin therapy in children and adults also have not been reported. None of the animal studies of maternal melatonin treatment or in postnatal life have shown any treatment-related side effects. The authors conclude that treatment with melatonin might result in a wide range of health benefits, improved quality of life and reduced healthcare costs and may help reduce complications in the neonatal period.
Oxygen-derived metabolites, collectively termed reactive oxygen species (ROS), are normally produced in aerobic organisms. When produced in excess, ROS mutilate molecules and are important mediators of cell and tissue damage [1–3]. The inevitable damage is referred to as oxidative stress. Free radicals are highly unstable and several enzymes and small-molecular-weight molecules with antioxidant capabilities protect against them ; collectively, this is referred to as the antioxidative defense system. Within all organisms there is a critical balance between free radical generation and antioxidant defenses. Free radical reactions cause oxidation of lipids, proteins, polysaccharides and also damage DNA (fragmentation, apoptosis, base modifications and strand breaks) and thus they have a wide range of toxic biological effects [5, 6].
Newborns and especially the preterm infants are especially prone to oxidative stress. The reasons for this are several. Infants very often (a) are exposed to high oxygen concentrations, (b) have infections or inflammation, (c) have reduced antioxidant defense, and (d) have high levels of free iron which enhance the Fenton reaction leading to production of highly toxic radicals . Oxidative stress likely contributes to the severity of several newborn conditions to the extent that Saugstad [8, 9] coined the phrase ‘oxygen radical disease of neonatology’. The idea suggests that oxidative stress affects a variety of organs, often simultaneously, and gives rise to different signs according to the organ most damaged. He includes bronchopulmonary dysplasia/chronic lung disease (CLD), retinopathy of prematurity and necrotizing enterocolitis in this category. Subsequently, it became clear that free radicals are involved in periventricular leukomalacia (PVL)  as well as in influencing the ductus arteriosus and pulmonary circulation [11–13]. If the concept of ‘oxygen radical disease of neonatology’ is valid, then the conditions mentioned are not different diseases but belong to the same category yielding different symptoms according to which organs are most involved.
Oxidative stress and antioxidant status in pregnancy and in pre-eclampsia
Pregnancy is a physiological state accompanied by a high metabolic demand and elevated requirements for tissue oxygen. This increased oxygen demand augments the rate of production of ROS and even women with normal pregnancies experience increased oxidative stress and lipid peroxidation relative to age-matched, nonpregnant women. Several studies have shown that the antioxidative defense system is altered during pregnancy. Circulating levels of lipid peroxides increase significantly in the maternal circulation when a woman becomes pregnant. Various antioxidants, however, including vitamin E, ceruloplasmin, erythrocyte thiols and iron-binding capacity also increase. Several of these elevate progressively with advancing gestation, while serum iron concentrations progressively decrease. While there is a gradual favoring of antioxidant activity over oxidation during normal pregnancy, there is an insufficient increase in antioxidants to offset the rise in free radical generation.
In contrast to the low-molecular weight antioxidants, the activity of an important family of antioxidative enzymes, the superoxide dismutases (SOD), are depressed in the blood of pregnant women . In addition, Walsh and Wang  reported a deficiency in another antioxidative enzyme, glutathione peroxidase (GPx), during pregnancy. GPx is an important antioxidant enzyme present in virtually all tissues. The enzyme limits the accumulation of lipid peroxides and utilizes reduced glutathione (GSH) as its cofactor to convert lipid peroxides into relatively harmless hydroxylated fatty acids, water and glutathione disulfide. Given these actions, it might be expected that a deficiency in this enzyme may lead to elevated oxidative stress during pregnancy. The placenta is a major source of oxidative stress during pregnancy. It is rich in polyunsaturated fatty acids, and the placenta is an abundant source of lipid peroxides which are secreted into the maternal circulation. In normal pregnancy, placental lipid production is believed to be kept under control by placental antioxidant enzymes . Major antioxidant enzymes such as SOD, catalase (CAT), GPx, glutathione reductase, glutathione S-transferase and glucose-6-phosphate dehydrogenase are all present in the placenta. In the normal placenta, the activities of SOD and CAT increase as gestation progresses, while the activity of GPx is diminished. On the other hand, placental production of lipid peroxides progressively drop as normal gestation advances, most likely because of the elevated activities of SOD and CAT. Thus, in normal pregnancy, placental antioxidant defenses are considered sufficient to control lipid peroxidation.
Pre-eclampsia is a multisystem disorder unique to human pregnancy. It is a complication in 5–10% of pregnancies and remains a leading cause of maternal and neonatal mortality and morbidity. This human disorder is a leading cause of premature delivery and intrauterine fetal growth retardation (IUGR). Pre-eclampsia is usually diagnosed in late pregnancy because of increased blood pressure and proteinuria with the symptoms of pre-eclampsia typically disappearing shortly after delivery of the placenta. A significant rise in lipid peroxidation levels in the placenta of pre-eclampsia has been suggested [17–21]. Several lines of evidence support this assumption, including increased lipid peroxidation products, elevated nitrotyrosine immunostaining and reduced antioxidant enzyme activities in pre-eclamptic placentas. In a case–control study, Vanderlelie et al.  measured tissue levels of SOD, GPx and lipid peroxidation in placental samples from women with normal pregnancies (18 women) and with pre-eclampsia (20 women). Placental tissue homogenates from pre-eclamptic patients contained significantly higher levels of lipid peroxides [malondialdehyde (MDA) and 4-hydroxy-2 (E)-nonenal; 20.68 versus 5.33 mm/mg protein], whereas there were significantly lower levels of SOD (2.02 versus 2.48 U/mg protein) and of GPx (11.50 versus 17.33 mmol/min/mg) than in control placentas. These finding are consistent with a limited enzymatic antioxidant capacity and elevated breakdown of lipids in placental tissue of women suffering from pre-eclampsia. Increased levels of thromboxane and lipid peroxides associated with a loss of GPx activity was also reported in placentas from pre-eclamptic patients compared with those from normal pregnancies . In parallel, the in vitro production of lipid peroxides and thromboxane is augmented in both trophoblast cells and villous tissues from women with pre-eclampsia . Furthermore, production of 8-iso-PGF2α and MDA (a lipid peroxide metabolite), as measured by levels in the medium, is higher for pre-eclamptic placental tissue explants than for normal placental explants . Collectively, the data provide convincing evidence that oxidative stress and especially lipid peroxidation are abnormally increased in the placentas of pre-eclamptic women.
Many et al.  found particularly intensive immunoreactivity for nitrotyrosine in invasive cytotrophoblasts in placental biopsies and vascular endothelium in the floating villi obtained from women with pre-eclampsia. The presence of nitrotyrosine is suggestive of damage caused by peroxynitrite, a potent nitrosative agent . Overall, the findings of nitrotyrosine residues in the cellular components of pre-eclamptic placentas may reflect increased production of the superoxide anion radical, as it couples with nitric oxide to generate peroxynitrite.
Placental generation of ROS and reactive nitrogen species (RNS) in pre-eclampsia might be facilitated by a reduction in local antioxidant defense, although it is not clear whether this reduced antioxidant defense is part of the problem or secondary to free radical damage. The activities of placental SOD and glucose 6-phosphate-dehydrogenase are reduced in pre-eclampsia compared to placentas from normal pregnancy . Moreover, the activities and mRNA expression of Cu/ZnSOD and GPx, and tissue levels of vitamin E are significantly lower in placental tissues from pre-eclampsia than from normal pregnancy .
In summary, there appears to be an increment in ROS generation in the placenta of pre-eclamptic women. There is also evidence for increased nitrotyrosine residue formation in the pre-eclamptic placenta suggestive of peroxynitrite formation, perhaps arising from local NO production coupled with increased generation of superoxide anion radical and either regionally decreased or inadequate SOD.
Oxidative stress at parturition
The transition from fetal to neonatal life at birth includes acute and complex physiologic changes. The delivery of the fetus from an intrauterine relatively hypoxic environment with a PO2 of 20–25 mmHg to an extrauterine normoxic environment with a PO2 of 100 mmHg increases oxidative stress. This four to fivefold rise in oxygen tension is believed to induce a greater production of ROS . In addition, labor and childbirth may be associated with periods of both hypoxia and oxidative stress for the newborn, while neonatal plasma is relatively deficient in antioxidants.
Several investigators studied the relationship between the oxidative state of the mother and the newborn at the moment of birth. Arguelles et al.  measured oxidative stress markers [carbonyl groups, lipid peroxides and total antioxidant capacity (TAC)] and found a good correlation between the oxidative status of the mother and of the neonate, with higher oxidative stress correlating with an even higher oxidative stress of the newborn based on measurements in umbilical cord blood. They also report that smoking mothers and their newborns had a higher concentration of the carbonyl groups, lipid peroxides and a lower TAC.
Term labor is typically associated with oxidative stress for the neonate, but there is no difference between the degree of fetal oxidative stress in vaginal delivery and cesarean section [30, 31]. It is unclear, however, whether oxidative stress is related to the delivery itself or whether it reflects a pre-existing fetal oxidative status. Lauries et al.  demonstrated that distressed fetuses delivered by emergency cesarean exhibited increased MDA concentrations, a parameter indicative of oxidative damage, and an enhanced GPx activity in amniotic fluid and umbilical cord blood compared to nondistressed fetuses delivered by elective cesarean section. This is probably an indication of higher fetal oxidative stress.
Oxidative stress in perinatal asphyxia
Perinatal asphyxia is an insult to the fetus or newborn resulting from a lack of oxygen (hypoxia) or a reduced perfusion (ischemia) in various organs. While virtually every organ of the body is affected by asphyxia leading to multiorgan failure, the most severe insult occurs in the central nervous system, which also lacks many repair processes . The mechanism of cellular injury after hypoxia or ischemia is poorly understood, but is probably mediated by an excess release of neurotransmitters, generation of ROS/RNS and the initiation of lipid peroxidation which, in turn, leads to a cascade of damaging events .
At the cellular level, cerebral hypoxia–ischemia sets in motion a series of biochemical events commencing in a shift from oxidative to anaerobic metabolism; this leads to an accumulation of NADH, FADH, lactic acid and H+ ions . If the asphyxic insult persists, the fetus is unable to maintain circulatory centralization, cardiac output and cerebral perfusion falls. Owing to the acute reduction in its oxygen supply, oxidative phosphorylation and ATP production in the brain are diminished [36, 37]. As a result, the Na+/K+ pump in cell membranes are deprived of the required energy to maintain ionic gradients. With a reduced membrane potential, increased numbers of calcium ions flow through voltage-dependent ion channels, down an extreme extra-intracellular concentration gradient, into the cell. Intracellular accumulation of Na+ and Cl− ions leads to swelling of the cells as water enters by osmosis (cytotoxic cell edema) . This damage is thought to be caused by the postischemic production of oxygen radicals, synthesis of NO, inflammatory reactions and an imbalance between excitatory and inhibitory neurotransmitter systems. Part of the secondary neuronal cell damage may be caused by induction of a well-known cellular suicide program referred to as apoptosis . Production of reactive species in the early reperfusion phase plays a substantial role in the resulting brain cell damage. Among the toxicants generated are the superoxide anion radical (O2•−) and hydrogen peroxide (H2O2). The latter agent can be converted to the highly reactive hydroxyl radical by transition metals, in particular free iron, ultimately leading to lipid peroxidation of the brain cell membranes as well as damage to other macromolecules .
Recent studies reported increased intra-erythrocyte free iron levels in infants with asphyxia . Iron may be released from hemoglobin in erythrocytes as result of oxidative stress . As the erythrocyte is a target of extracellular free radicals, free iron release may be followed by extracellular oxidative stress caused by O2•− generation because of phagocyte activation . Intra-erythrocyte free iron concentrations appear to be a reliable marker of cell oxidative stress and an indicator of the risk of oxidative injury in other tissues.
Increased production of free radicals including the O2•− and NO induces oxidative stress in the placenta by formation of the pro-oxidant ONOO−; this reactant is formed when O2•− couples with NO• [44, 45]. ONOO− is cytotoxic due to a number of independent mechanisms including the initiation of lipid peroxidation, the inactivation of a variety of enzymes (most notably, mitochondrial respiratory enzymes) and membrane pumps  and depletion of GSH . Moreover, ONOO− causes DNA damage [48, 49] resulting in the activation of the nuclear enzyme poly (ADP-ribose) synthetase, depletion of NAD and ATP, and ultimately leading to cell death . The higher levels of free radical production associated with repetitive ischemia likely reflect differences in oxygen availability during these events. During repetitive ischemia, the availability of oxygen to the fetus for intermittent periods of reperfusion facilitates free radical production. With the increased availability of oxygen, oxidative reactions rather than reductive reactions are favored.
Respiratory distress syndrome (RDS) and oxidative stress
Hyperoxic exposure itself, although essential for promoting survival of infants with RDS, induces excessive production of ROS in the respiratory system. There exist, however, several potential causes of intracellular and extracellular oxidant stress in the preterm newborns with RDS. The high inspiratory concentrations of oxygen required to achieve adequate arterial oxygenation, pro-oxidant drugs and infections or extrapulmonary inflammation can all promote ROS accumulation and the utilization and depletion of antioxidative agents .
In experimental models of respiratory distress, the specific targets of a hyperoxic insult to the lung are the vascular endothelial cells and the epithelial cells of the alveoli. ROS induce ultrastructural changes in the cytoplasm of pulmonary capillary endothelial cells and cause focal hypertrophy and altered metabolic activity. Thus, increased oxidative stress accompanied by reduced endogenous antioxidant defenses may play a role in the pathogenesis of a number of inflammatory pulmonary diseases including respiratory distress in the newborn [52, 53]. A deficit in the precise balance between exposure to oxidants and endogenous antioxidants obviously leads to elevated oxidative damage. The molecular damage caused by free radicals and related reactants appear to be involved in the pathogenesis of a growing number of diseases, including RDS of the newborn [54, 55].
When phagocytes such as neutrophils are stimulated by microorganisms or other means, they are activated and increase their oxidative metabolism; as a result, toxic oxygen derivatives, i.e. ROS, are formed. If these oxygen-based products are not inactivated, their high chemical reactivity leads to damage to a variety of cellular macromolecules including proteins, carbohydrates, lipids and nucleic acid. This results in cell injury and may induce respiratory cell death . Under these conditions, a surfactant deficiency may be aggravated by the inactivation of the small amount of endogenous surfactant that is produced . Furthermore, if exogenous surfactant is given, it may also be destroyed [53, 55].
Reactive oxygen species also have been implicated in the molecular damage seen in the bronchoalveolar lavage (BAL) fluid of patients with RDS [58, 59]. This assertion is supported by several findings; H2O2 is detected in the expired air of RDS patients, and myeloperoxidase (MPO) and oxidized-1-antitrypsin have been found in BAL fluid. Moreover, increased plasma lipid peroxidation products have been measured in critically ill patients and in patients with sepsis and at risk of developing RDS. Also, evidence of augmented levels of oxidized lipids and proteins have been found in the plasma of patients with RDS. Elevated levels of ROS also have been implicated in the molecular damage seen in the BAL fluid of patients with RDS. BAL fluid normally contains a large amount of the antioxidant GSH; however, in patients with RDS this is mostly in the oxidized form .
Consistent with this, oxidative inactivation of 1-antiprotease also has been observed in RDS. Elevated concentrations of xanthine and hypoxanthine are present in the plasma and BAL fluid of patients with RDS and are a potential source of ROS in the presence of exogenously added xanthine oxidase. Also, elevated concentrations of orthotyrosine and metatyrosine in BAL fluid protein imply the formation of the damaging •OH in the lungs of these patients, as orthotyrosine and metatyrosine are isomers of tyrosine thought to be formed exclusively by aromatic hydroxylation of phenylalanine by •OH . Chlorotyrosine and nitrotyrosine also have been found in BAL fluid from patients with RDS. Increased concentrations of chlorotyrosine residues in BAL fluid proteins from patients with RDS indicate hypochlorous acid (HClO) production by the activated inflammatory cells in the lungs of these patients. Chlorotyrosine is formed by HClO-dependent chlorination of paratyrosine. HClO is a damaging oxidant formed from H2O2 and chloride ions by the enzyme MPO, present in activated inflammatory cells. HClO has been implicated as the major damaging species produced by activated neutrophils. While HClO itself is a destructive oxidant, it also may interact with low molecular mass iron or O2•− to produce the •OH.
Nitrotyrosine concentrations also are significantly elevated in the BAL fluid protein of patients with RDS . Nitration of tyrosine residues is an in vivo marker of the formation of ONOO−. Earlier studies reported increased nitrotyrosine concentrations in the lungs of patients with RDS. Additionally, under acidic conditions ONOO− decomposes to form a powerful oxidant with properties similar to •OH. There is, however, another possible explanation for the formation of nitrotyrosine. Recent work shows that nitrotyrosine can arise from the reaction of tyrosine with nitroxyl chloride, an intermediate formed by the interaction of nitrite (the auto-oxidation product of nitric oxide) with HClO. Interestingly, nitrotyrosine concentrations in BAL fluid protein from patients with RDS treated with NO were elevated compared with those found in lung-injured patients not receiving this therapy .
Increased nitrotyrosine concentrations may reflect augmented ONOO− formation in patients receiving NO. As the patients receiving inhaled NO are no sicker, in terms of Acute Physiology and Chronic Health Evaluation II score or FIO2 requirements, than those patients not receiving this therapy, it implies that inhaled NO may react with O2•− in these circumstances to form the nitrating agent . Finally, MPO concentrations are significantly elevated in the BAL fluid from patients with RDS, suggesting lung neutrophil recruitment and activation . Collectively, the data are compelling that RDS is associated with elevated ROS/RNS generation and the consequential increased oxidative damage to the respiratory tree.
Oxidative stress and neonatal sepsis
Sepsis represents a serious problem in newborns with an incidence of one to 10 cases per 1000 live births, with even higher rates in low-birth-weight neonates. Hospital acquired infections in neonatal intensive care units may also occur as frequently as 30 infections per 100 patients. Mortality rates resulting from sepsis in newborns are 30–50% . Sepsis is characterized by alterations in body temperature, hypotension, hypoperfusion with cellular damage which culminates in multiple organ failure .
The initiating event in sepsis is the result of release of endotoxins [i.e. bacterial cell wall lipopolysaccharides (LPS)] from gram-negative and gram-positive pathogenic bacteria . LPS triggers activation of inflammatory cells, including polymorphonuclear leukocytes (neutrophils; PMN), monocytes/macrophages and lymphocytes. LPS also initiates cellular and humoral aspects of the inflammatory immune response. The inflammatory response that occurs as the result of infection is the predominant determinant of outcome in sepsis [69, 70]. A major feature of sepsis is tissue infiltration by phagocytic cells [71–75]. When this occurs, PMN and monocytes/macrophages respond to septic stimulation by producing ROS and RNS . In addition, PMN release enzymes (e.g. elastase, cathepsin, etc.) and the MPO-derived oxidant, HOCl. These reactants contribute to PMN/macrophage-mediated killing of bacteria. However, if produced in excess during sepsis, the ROS/RNS and proteolytic enzymes cause microvascular dysfunction followed by organ shutdown.
An inflammatory response to septic stimuli is crucial for host defense, because it up-regulates anti-inflammatory mediators (e.g. IL-1 receptor antagonist, IL-4, IL-10) and antioxidant enzymes (e.g. CAT, GPx and SOD). However, the excessive production of pro-inflammatory mediators in sepsis overwhelms the anti-inflammatory signaling processes leading to a suppression of innate immune functions (especially those of PMN) and causing immunoparalysis and subsequently an increased susceptibility to infection . It is important to note that, besides immune cells, microvascular endothelial cells also become activated in sepsis which contributes to amplification of the inflammatory response [68, 73, 78]. It is known that septic stimuli (e.g. LPS, TNF-α) initiate activation of transcription factors including NFκB and AP-1, resulting in transcriptional activation of multiple genes. This leads to the release of pro-inflammatory cytokines (e.g. TNF-α, IL-1β, etc), and elevates the expression of adhesion molecules (e.g. E-selectin, ICAM-1, VCAM-1) and chemokines by endothelial cells [73, 79–81]. The central role of ROS/RNS in modulation of the endothelial cell proinflammatory phenotype is well documented . Despite a large amount of research, little progress has been made in improving the outcome of septic newborns . Efforts to block one or more aspects of the sepsis-associated inflammatory pathways have had little impact on patient survival. Of many drugs tested, few have demonstrated efficacy [83–86].
Clinical reports are consistent with the involvement of ROS/RNS in neonatal sepsis and its complications. Batra et al.  documented increased production of oxygen-derived reactants in septic neonates. Also, Seema et al.  found newborns with sepsis have significantly higher levels of TNF-α and increased activity of antioxidative enzymes, SOD and GPx. Finally, Kurt et al.  demonstrated that serum IL-1β, IL-6, IL-8, and TNF-α are mediators of inflammation and can be used at the diagnosis and at the evaluation of the therapeutic efficiency of drugs used to treat neonatal sepsis.
Antioxidant status of newborns
Compounds that prevent cellular damage caused by ROS/RNS may act in numerous ways, ranging from prevention of their formation (such as chelators of metal ions and anti-inflammatory agents) to their interception once they are formed. Most antioxidants used clinically fall into the latter category. Antioxidants may be broadly classified as enzymatic (SOD, CAT and GPx) or nonenzymatic (vitamin C, E, β-carotene, allopurinol, melatonin, etc).
As antioxidative enzymes are normally present in cells, it seemed logical to supplement them, especially in the premature neonate who may be deficient. Several studies indicate that supplementation with a single antioxidant enzyme, e.g. SOD, is not protective; it must be used in combination with another enzyme such as CAT [90, 91]. Furthermore, these enzymes, when administered systemically, do not readily enter cells unless conjugated to polyethylene glycol or encapsulated in liposomes. Benefits of exogenously administered natural surfactant may at least partly relate to its antioxidant properties . Suresh et al.  tested whether exogenously administered SOD is efficacious in the prevention of CLD in preterm infants who are mechanically ventilated. Also, SOD was examined to determine if it reduced damage in the following situations: bronchopulmonary dysplasia, intraventricular hemorrhage, PVL, retinopathy of prematurity, necrotizing enterocolitis, patent ductus arteriosus and mortality. To determine the frequency and nature of any potential adverse effects of SOD, randomized controlled trials involving preterm infants who had developed or were at risk of developing RDS requiring assisted ventilation were treated with the enzyme [94–99]. The data from two randomized controlled trials reported no differences in the pooled data in terms of death prior to discharge, oxygen dependency at 36 wk corrected age, oxygen dependency at 28 days of life or other endpoints [100, 101]. In another study , survivors who had been treated with SOD had a shorter duration of continuous positive airway pressure (4.9 versus 9.7 days), a lower frequency of respiratory problems after discharge and a lower frequency of chest radiograph abnormalities compared to newborns who received placebo. Based on currently available published trials, there is insufficient evidence to draw firm conclusions about the efficacy of SOD in preventing CLD of prematurity. Data from a small number of treated infants suggest that it is well tolerated and has no serious adverse effects.
In a randomized, placebo-controlled trial, prophylactic, intratracheal recombinant human SOD (rhSOD) given at birth to premature infants (birth weight 600–1200 g) at high risk for developing BPD was associated with many fewer episodes of respiratory illness (wheezing, asthma, pulmonary infections) severe enough to require treatment with bronchodilators or corticosteroids at 1 yr corrected age . This suggests that rhSOD may prevent long-term pulmonary injury from ROS in high risk premature infants.
The mechanisms by which ROS influence cellular proliferation and gene expression generally involve changes in the redox state, i.e. the ratio of oxidizing to reducing equivalents. Redox state also varies in a predictable pattern during the cell cycle. Frosali et al.  measured higher GSH, and oxidized glutathione (GSSG), glucose-6-phoshate dehydrogenase (G-6-PDH) and hexokinase levels/activities and lower GSH/GSSG ratios and higher GSH-recycling rates in RBCs of term and preterm babies than those of adults. In preterm babies, significant correlations were found between G-6-PDH and CAT, GSH, GSH/GSSG ratio and GSSG. In term newborns, statistically significant correlations were observed between G-6-PDH and CAT, SOD and GSH. The findings suggest a central role of G-6-PDH activity in antioxidant defense. Frosali et al.  speculate that preterm babies have more rapid involvement of antioxidant defenses than term babies.
In a prospective, randomized, blinded trial, Vento et al.  studied the effects of resuscitation upon oxygenation in a group of asphyxiated newly-born infants receiving room air or 100% oxygen as the gas source. During the acute phase of asphyxia and until the resuscitation procedure concluded, they measured serial blood gases as well as GSH and GSSG, the enzymes involved in the glutathione redox cycle, and antioxidant enzyme activities. They found a significant correlation between hyperoxemia and the intra-erythrocyte GSSG concentration. They hypothesize that hyperoxemia may be one of the triggering factors responsible for an increased oxidation of GSH. Catalase has not been studied in newborn .
The dietary antioxidant, vitamin E (α-tocopherol), was among the first to be used in the hope of preventing neonatal disease. The rationale for the use of vitamin E was based on evidence that neonates are deficient in this vitamin as well as on the actions of this compound in preventing membrane lipid peroxidation which is likely to contribute to lung disease. However, it was soon recognized that measurement of plasma vitamin E as a measure of sufficient antioxidant capacity is seriously flawed . Moreover, prolonged pharmacological use of vitamin E has been shown to interfere with bacterial killing by neutrophils and mononuclear cells. A recent clinical study showed that the use of vitamin E increased the incidence of sepsis and necrotizing enterocolitis in neonates . Exogenously administered vitamin C and β-carotene, in particular, may act as pro-oxidants and actually increase mortality in some groups of patients [108, 109]. Pharmacological preparations of antioxidants may also cause harm, as was thought to be the case with E-Ferol, an intravenous preparation of vitamin E (α-tocopherol acetate) that led to a number of deaths of neonates. It was also assumed that vitamin E could potentially limit the processes that lead to CLD but a systematic review of the vitamin E trials indicates that this therapy does not reduce the incidence of CLD .
Vitamin A is involved in the regulation and promotion of growth and differentiation of many cells. It also maintains the integrity of respiratory tract epithelial cells. Very preterm infants are relatively deficient in vitamin A, which is associated with CLD. A large randomized clinical trial of 807 infants with a birth weight of less than 1000 g has shown that the use of large doses of intramuscular vitamin A reduced the risk of CLD (OR 0.89) . Other smaller trials support a similar conclusion. However, this preventive therapy requires several injections . A trial of daily oral vitamin A therapy for 4 wk in a similar population of infants failed to detect any benefit .
Low selenium status has been documented in preterm infants and may be a risk factor for CLD. A randomized controlled trial suggested that postnatal selenium supplementation in very low birth weight infants did not improve neonatal outcome . Primary outcome measures were oxygen dependency at 28 days and total days of oxygen dependency. No significant differences were seen between the groups with respect to primary or secondary outcome measures, with the exception that fewer supplemented infants who had an episode of sepsis after the first week of life. Lower maternal and infant prerandomization selenium concentrations were associated with increased respiratory morbidity.
One study  determined whether protein carbonyls and the MDA are elevated in plasma of very low birth weight (<1500 g) infants and whether they are affected by selenium supplementation. Moreover, this study determined whether the damaged lipid and protein are associated with poor respiratory outcome or retinopathy. In this study, Winterbourn et al.  found that protein carbonyl concentrations in very low birth weight infants were significantly higher than for adults but lower than in cord blood from term infants. MDA concentrations in very low birth weight infants overlapped the ranges for healthy adults and cord blood from term infants. Selenium supplementation almost doubled plasma selenium concentrations, but carbonyls and MDA did not change as a result of the treatment. There were no significant differences in oxidant marker levels in infants who did or did not develop CLD. Darlow and Austin  demonstrated that supplementing very preterm infants with selenium is associated with benefit in terms of a reduction in one or more episodes of sepsis. Supplementation was, however, not associated with improved survival, a reduction in neonatal CLD or retinopathy of prematurity.
Marro et al.  evaluated the efficacy of allopurinol in inhibiting purine metabolism via the xanthine oxidase pathway in neonates with severe, progressive hypoxemia during rescue and reperfusion with extracorporeal membrane oxygenation (ECMO). Hypoxanthine concentrations were significantly higher in isolated ECMO circuits and increased over time during bypass. This demonstrates that allopurinol reduces the production of ROS during reoxygenation and reperfusion of hypoxic neonates recovered on bypass. Russell et al.  demonstrated that during admission, plasma hypoxanthine concentrations were significantly higher in infants who subsequently developed PVL, BPD, or ROP, but there was no difference in the primary endpoint (PVL) between the allopurinol-treated and control groups. The failure of allopurinol prophylaxis in this group of infants is probably related to the complex nature of the pathological processes and to the timing of treatment. If oxidant processes are an important mechanism of cellular injury in these preterm infants, an alternative biochemical modulator may be required to reduce the damage or a combination of agents may be effective.
Evidence exists that allopurinol reduces delayed cell death in animal models of perinatal asphyxia and in human patients with other forms of organ reperfusion injury. The Cochrane database of 2008  showed that the available data are not sufficient to determine whether allopurinol has clinically important benefits for newborn infants with hypoxic-ischemic encephalopathy.
Melatonin as an antioxidant
Melatonin, an endogenously produced indoleamine, is a highly effective antioxidant and free radical scavenger. That melatonin may be a free radical scavenger was first suggested by Ianas et al. . Tan et al.  were the first to document that melatonin detoxifies the highly reactive •OH. Since these reports, there have been numerous confirmatory studies using a wide variety of methodologies [122, 123]. Pure chemical, cell-free systems, as well as cultured cells and animal studies have shown melatonin to scavenge the •OH [123–125]. The average calculated rate constant for the scavenging of the •OH by melatonin is similar to that of other known efficient •OH scavengers . According to Zang et al. , melatonin interacts with H2O2 as indicated by the dose–response reduction in its concentration in a mixture to which increasing amounts of melatonin were added. Details of the interaction mechanisms of melatonin with H2O2 were, however, not provided. According to Tan et al. , melatonin scavengers H2O2 with the formation of N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK); this molecule and others also possess significant scavenging activity [129–132]. Pieri et al. [133, 134] concluded that melatonin may be a more efficient scavenger of the peroxyl radical than is trolox (water soluble vitamin E). Melatonin and its metabolites efficiently interact with various ROS/RNS as well as with organic radicals, upregulates antioxidant enzymes (including GPx and glutathione reductase) and downregulates pro-oxidant enzymes (NO synthases, lipoxygenases) [135–137].
After birth, the full-term neonate does not produce melatonin for 2–4 months, leading to transient melatonin deficiency [138–143]. Previous studies  have reported differences between plasma levels of melatonin in the umbilical artery and vein in humans, with higher values in the vein and day/night differences in melatonin concentrations, and evidence of the absence of a circadian rhythm of melatonin during the early neonatal period . Prematurity itself does not hasten the maturation of the neurological network controlling melatonin secretion. The onset of pineal melatonin secretion is even more delayed when there is exposure to neurological insults [139, 142]. Thus, in premature neonates the melatonin deficiency is more prolonged [138, 139, 141, 142]. An infant born 3–4 months prematurely may lack significant melatonin levels for 7–8 months or longer. Melatonin in the nocturnal breast milk does not correct this deficiency [141, 146, 147] and does not affect neonatal morbidity [148–151]. Abnormal brain and ocular development and IUGR frequently occur in very premature infants. Studies in animal models suggest that these may be related to a melatonin deficiency and can be corrected with melatonin supplementation. In summary, the premature infant, and especially its developing brain, is uniquely vulnerable to hypoxic-ischemic injury, with a complex evolution of the injury that affords opportunities for intervention [152, 153]. Melatonin, in reducing free radical generation, may be an effective neuroprotective treatment for the fetus  as it is in adults [154, 155].
Intrauterine fetal growth retardation and fetal distress in human infants have been associated with a pronounced reduction in melatonin secretion during the first 3 months of life. Additionally, urinary 6-sulphatoxylmelatonin (a melatonin metabolite) (6SaMT) excretion is impaired in adults who were growth restricted prenatally or were born after 40 wk of gestation . The urinary excretion of 6SaMT, which is measurable in children but not in newborns [156, 157], suggests there is a relation between melatonin production and body size at birth. The mean daily excretion of 6SaMT was significantly lower in response to an apparent life-threatening event than in normal infants and siblings of sudden infant death syndrome victims.
Neonatal stress (acute fetal distress) increases nocturnal melatonin production compared with normal term and preterm neonates . Muñoz-Hoyos et al.  postulated that the high melatonin concentrations they observed, which may remain so for a prolonged period, protect against ROS. Melatonin is a potent small-molecular-weight antioxidant that often attains especially high levels during physiological ischemia/reperfusion episodes affecting newborns during delivery. According to Tan et al.  melatonin plays an important role in physiological ischemia/reperfusion and protects against the potential oxidative injury induced by transitory physiological ischemia/reperfusion.
Several clinical studies on melatonin showed that it reduces oxidative stress in newborns with sepsis, distress or other conditions where there is excessive ROS production. In one of these studies, a product of lipid peroxidation, MDA and the nitrite + nitrate levels were measured in the serum of asphyxiated newborns before and after treatment with melatonin given within the first 6 hr of life . The results documented that melatonin may be beneficial in the treatment of newborn infants with asphyxia. Serum levels of lipid peroxidation products (MDA and nitrite + nitrate concentrations) in newborns with asphyxia are significantly higher than those in normal infants; moreover, they were significantly reduced by melatonin. The protective actions of melatonin in this study likely related to the antioxidant properties of the indole and its metabolites as well as to the ability of melatonin to increase the efficiency of mitochondrial electron transport [161, 162].
Gitto et al.  examined the changes in the clinical status and serum levels of lipid peroxidation products [MDA and 4-hydroxylalkenals (4-HDA)] in 10 septic newborns treated with melatonin given within the first 12 hr after diagnosis. Ten other septic newborns in a comparable state were used as ‘septic’ controls, while 10 healthy newborns served as normal controls. Serum MDA + 4-HDA concentrations in newborns with sepsis were significantly higher than those in healthy infants without sepsis; in contrast, in septic newborns treated with melatonin there was a significant reduction of MDA + 4-HDA levels to the values measured in the normal controls at both 1 and 4 hr. Melatonin also improved the clinical outcome of the septic newborns as judged by measurement of sepsis-related serum parameters after 24 and 48 hr.
We also tested whether melatonin treatment would lower IL-6, IL-8, TNF-α and nitrite/nitrate levels in 24 newborns with RDS of III or IV grade (radiographically confirmed) diagnosed within the first 6 hr of life . Compared with the melatonin-treated RDS newborns, in the untreated infants the concentrations of IL-6, IL-8, TNF-α were significantly higher at 24, 72 hr and at 7 days after onset of the study. In addition, nitrite /nitrate levels at all time points were higher in the untreated RDS newborns than in the melatonin-treated babies. Following melatonin administration, nitrite/nitrate levels decreased significantly, whereas they remained high and increased further in the RDS infants not given melatonin.
We also measured proinflammatory cytokines (IL6, IL-8 and TNF-α) and the clinical status of 110 preterm newborns with RDS ventilated with different modalities [conventional ventilation, pressure-support ventilation (PSV) and with guaranteed volumes (GV) and high-frequency oscillatory ventilation] before and after treatment with the antioxidant melatonin . Compared with the melatonin-treated RDS newborns, the concentrations of inflammatory cytokines were significantly higher in the newborns given only the diluent. When comparing serum levels of IL-6, IL-8 and TNF-α for two groups, melatonin treatment clearly had anti-inflammatory effects. In particular, it was noted that newborns mechanically ventilated in PSV mode with GV presented a greater reduction of serum levels of inflammatory cytokines than did newborns ventilated in conventional mode or in oscillatory ventilation. The measured inflammatory cytokines were most markedly elevated in infants mechanically ventilated but not given melatonin. Newborns not treated with melatonin developed CLD have much higher levels of proinflammatory cytokines than infants without CLD . In conclusion, we showed the melatonin lowers interleukin IL-6, IL-8, TNF-α and nitrite/nitrate levels and modifies serum inflammatory parameters in surgical neonates improving their clinical course . The anti-inflammatory actions of melatonin in human newborns is consistent with similar actions of this indoleamine in other animals [167, 168].
Oxidative stress is defined as an imbalance between pro-oxidant and antioxidant forces resulting in an overall pro-oxidant insult [1–3]. This augmented oxygen requirement increases the rate of production of ROS which damages the newborn [5, 6]. Further information is required to better understand the interaction between oxidative stress and disease processes in the perinatal period; however, these should be evaluated by both biochemical and clinical investigations. As discussed herein and elsewhere [120–122], free radicals and the associated molecular damage are critical components of several diseases of the newborn.
In comparison with healthy adults, lower levels of plasma antioxidants such as vitamin E, β-carotene, melatonin, sulfhydryl groups, as well as lower levels of plasma metal binding proteins such as ceruloplasmin and transferrin, and reduced activity of erythrocyte SOD are typical of newborn infants. It is possible that antioxidant therapy might be useful in the management of neonates with oxidative stress-related problems but further biochemical investigations are required to define the most effective antioxidant therapy.
There is general agreement that short-term melatonin therapy has a remarkably benign safety profile, even when neonates are treated with pharmacological doses [151, 160–166]. Moreover, significant complications with long-term melatonin therapy in children and adults have not been reported. Additionally, none of the animal studies on maternal melatonin treatment and in postnatal life have shown treatment-related side effects [169–173]. We believe that melatonin treatment might result in a wide range of health benefits, improved quality of life and reduce healthcare costs and may help reduce complications in the neonatal period.