Main

Despite advances in perinatal and obstetric care, perinatal hypoxic-ischemic brain injury remains a significant health problem and a major cause of mortality and cerebral disabilities in newborn infants(1, 2). Even relatively mild asphyxia may be associated with impaired cerebrovascular autoregulation which may result in major fluctuations in cerebral perfusion. In the past years, efforts have been directed toward understanding the pathophysiology of hypoxic-ischemic injury in the newborn. Using various experimental models, numerous neurochemical processes, including reoxygenation-associated generation of free radicals, have been shown to be involved in the cascade of events that culminate in brain damage(3). During resuscitation of asphyxiated babies, oxygen itself would have both benefits and risks. Clearly, neonates suffering perinatal asphyxia, apnea, bronchopulmonary dyplasia, or other respiratory complications often fluctuate during therapy between low and high Pao2. In clinical emergencies such as neonatal resuscitation, assisted ventilation begins classically with high oxygen concentrations until cyanosis disappears(4), and Pao2 is often difficult to measure under such hectic conditions. The side effects of relative brief periods of hyperoxemia, or at least fluctuating changes in Pao2, are difficult to evaluate and are still unknown; however, it has been shown that they cause a significant decrease in cerebral blood flow(5) and increase the frequency of retinopathy(6).

Although neuronal cells in primary culture serve only as a simplified model, the situation in vivo being much more complex, they appear to be useful in elucidating cellular and molecular mechanisms in various diseases, including cerebral hypoxia. The present study was designed to test the importance of the restoration of oxygen delivery on neuronal cell integrity and functional activity. The influence of two levels of reoxygenation conditions, i.e. normoxia and transient hyperoxia, has been investigated on the consequences of a severe episode of asphyxia in a model of cultured neurons from the embryonic rat brain, by analyzing cell morphology, LDH (a marker of cell injury) release, energy metabolism, and superoxide radical production.

METHODS

Animals. When they were in the proestrus period, as shown by the observation of daily vaginal smears, Sprague-Dawley female rats (R. Janvier, Le Genest-St-Isle, France) were housed together with males for 24 h. They were then maintained in separate cages, under standard laboratory conditions on a 12:12 h light/dark cycle (lights on at 0600 h) with food and water available ad libitum. All animal experimentation was carried out with the highest standards of animal care, according to the NIH Guide for Care and Use of Laboratory Animals.

Neuronal cell cultures. Neuronal cell cultures were obtained from 14-d-old rat embryo forebrain. Pregnant female rats were anesthetized with halothane, and living embryos were excised by cesarian section under sterile conditions. Whole embryos were placed in culture medium previously equilibrated at 37°C, and forebrains were carefully collected. Neuronal cell suspensions were then obtained as previously described(7). Brain tissues were dissected free of meninges and gently dispersed in a mixture of Dulbecco's modified Eagle's medium and Ham's F-12 medium (50/50) supplemented with 5% inactivated FCS. After centrifugation at 700 × g for 10 min, the pellet was redispersed in the same medium and passed through a 46-μm pore size nylon mesh. Density of the cell suspension was measured, and aliquots were transferred into 35-mm Petri dishes(Falcon) precoated with poly-L-lysine to obtain a final density of 106 cells/dish. Cultures were then placed at 37°C in a humidified atmosphere of 95% air/5% CO2. The following day, the culture medium was removed by aspiration and then replaced with a fresh hormonally defined serum-free medium consisting of the Dulbecco's modified Ealge's medium/Ham's F-12 mixture enriched with human transferrin (1 mM), insulin (1 mM), putrescine (0.1 mM), progesterone (10 nM), estradiol (1 pM), and sodium selenite (30 nM). Subsequent medium changes with serum-free medium were performed twice a week. Neuronal cell cultures were kept for 6 d under a standard normoxic gas mixture.

Hypoxia-reoxygenation procedure. After 6 d in vitro(T = 0), a first set of culture dishes was maintained for 6 h under normoxic conditions (95% air/5% CO2), whereas a second set was deprived of oxygen by incubation for 6 h in a gas mixture consisting of 95% N2/5% CO2. Half of each set of culture dishes was then placed for 3 h in normoxic atmosphere, whereas the remaining dishes were incubated under hyperoxia (95% O2/5% CO2) for the same period. All cultures were finally returned to standard normoxic atmosphere for 15 additional hours (T = 24 h) or 3 d (T = 72 h).

Just before the beginning of the experiment (T = 0), then immediately at the end of the hypoxic insult (T = 6 h), and finally as a function of time after reoxygenation, aliquots of the cell incubating medium were rapidly collected from representative dishes using a sterile syringe and analyzed for their O2 and CO2 contents and pH by means of a gas analyzer (Corning, Halstead, UK), as described by Sher(8).

Cell morphology was routinely assessed by phase-contrast microscopic observations. At the end of the experimental period, cultured cells were finally rinsed twice with 2 mL of 0.9% NaCl and scrapped off with 1 mL of 1 M NaOH for the determination of their protein concentrations, according to Bradford(9).

LDH release. The release of LDH from the neuronal cells exposed to hypoxia-reoxygenation was monitored as a function of time in the extracellular medium, using the spectrophotometric method described by Amadoret al.(10). LDH catalyzes the oxidation of lactate to pyruvate with simultaneous reduction of NAD. The formation of reduced NADH results in an increase in absorbance at 340 nm which is directly proportional to LDH activity in the sample.

Glucose transport and incorporation. Specific transport and incorporation of glucose into the cultured neurons was evaluated by the measurement of the specific uptake of 2-[3H]DG(11).

The culture dishes were washed twice in a HEPES-buffered Krebs-Ringer solution (125 mM NaCl, 4.8 mM KCl, 1.2 mM MgSO4, 1.3 mM CaCl2, 1.2 mM KH2PO4, 25 mM HEPES, pH 7.40) and incubated in 2 mL of the Krebs-Ringer saline medium for 15 min, at 37°C. The medium was then discarded, and the assay was then initiated by adding 1.5 mL of the buffer solution containing 1 mM 2-[3H]DG (960 GBq/mmol, DuPont NEN, Boston, MA). After a 30-min incubation, the radioactive medium was quickly removed, and the cultures were rapidly washed three times with physiologic saline. After drying, the cells were solubilized in 1 mL of 1 M NaOH, and samples were taken for scintillation counting. Nonspecific uptake of 2-[3H]DG was assessed by measuring the residual radioactivity in the cultured cells by the same experimental procedure performed in the additional presence of 100 mM D-glucose.

Superoxide radical production. The presence of superoxide radicals was quantified in the cell incubating medium by a second derivative spectrophotometric method using acetyl-cytochrome c as previously described(7).

Due to interference of the original culture medium containing phenol red with the spectrophotometric signal resulting from the reduction of acetyl-cytochrome c, cultured neurons were rinsed twice, immediately before cell exposure experiments, with 2 mL of a sterile HEPES-buffered Krebs-Ringer saline solution. Then, 1.5 mL of the buffer solution containing 60 μM acetyl-cytochrome c were added to each culture dish. Aliquots of culture incubating medium were sampled at various time intervals, diluted in degassed distilled water, and analyzed spectrophotometrically for the amount of acetyl-cytochrome c reduced. The absorption and the second derivative spectra were recorded between 500 and 600 nm against distilled water in a double-beam spectrophotometer (Uvikon 820, Kontron). The solution was then fully reduced by the addition of a few grains of sodium dithionite, and the second order derivative was recorded again. The amplitude variations were measured between the valley at 550 nm and the peak at 556 nm in both cases, and the amount of acetyl-cytochrome c reduced was expressed as a ratio of the amplitudes measured before and after the reduction of the preparation by sodium dithionite. As previously reported, the additional presence of superoxide dismutase (1 330 U/mL) in the Krebs-Ringer solution during cell incubation totally abolished the reduction of acetyl-cytochrome c(7).

Statistical analysis. For each parameter studied, data obtained under the various gassing regimens were compared with those obtained under the respective control conditions. In this respect, normoxia-hyperoxia and hypoxia-normoxia groups were compared with the normoxia-normoxia group, whereas the hypoxia-hyperoxia group was compared with the normoxia-hyperoxia group. In addition, comparisons were made between hypoxia-hyperoxia and hypoxia-normoxia groups to evaluate significant changes related to transient hyperoxia during the reoxygenation period. Statistical analyses were performed by means of the Bonferroni test for multiple comparisons(12). Conservative multiple comparison procedures were chosen to reduce the likelihood of type II errors.

RESULTS

Characteristics of the extracellular fluid. As shown in Table 1, exposure of the cultured neurons to an anaerobic environment for 6 h reduced Po2 by 78%. Three hours after return to normoxic conditions, Po2 increased by 314% in the incubating medium, but reached a lower value than in control cultures. By contrast, Po2 was significantly above control values 24 h later. Whereas Pco2 was not modified by hypoxia, pH of the extracellular medium decreased compared with controls, and was still low 3 and 24 h after the end of exposure to hypoxia. Similar changes in pH values were recorded in the cell culture medium after hypoxia-hyperoxia, whereas extracellular Po2 decreased by 78% after hypoxia, and then increased by 235% above normal values during hyperoxic reoxygenation. Finally, transient hyperoxia alone altered only Po2 values, without significant changes in other parameters, i.e. pH and Pco2.

Table 1 Evolution of pH and Po2 in the incubating medium of cultured neurons
Full size table

Neuronal cultured cells. Six-day-old cultures grown in serum-free medium on a polylysine substratum showed a very high percent (more than 97%) of living cells, as demonstrated by Trypan blue exclusion, with very few (<7%) nonneuronal elements, as previously shown in preliminary immunocytochemical characterization (data not shown). The cultures exhibited polygonal perikarya, with some cellular clumps, interconnected by a dense fiber network.

When not associated with preliminary hypoxia, hyperoxia for 3 h did not induce any changes in cell morphology over the whole experimental period,i.e. 72 h.

No significant morphologic alteration was noticeable after hypoxia-reoxygenation until cell examination at 72 h posthypoxia. Accordingly, protein levels in cultured neurons previously exposed to hypoxia-reoxygenation did not differ significantly from controls at T = 24 h (169.8± 18.4 versus 168.0 ± 10.6 μg/dish). Conversely, protein concentrations in the cells were reduced by 32% (p < 0.05) at 72 h posthypoxia. Cellular proteins exhibited a nonsignificant 18% reduction when transient hyperoxia occurred during reoxygenation. Concomitantly to the reduction of their protein contents, cultured neurons displayed marked morphologic alterations, i.e. neurite fragmentation associated with floating dead cells, reflecting cell degeneration. With the methodology used for the evaluation of cell morphology and viability, it was not possible to show any difference at this end point between hypoxia-normoxia and hypoxia-hyperoxia groups.

Lactate dehydrogenase release. Neurons in primary culture maintained under standard conditions released detectable levels of LDH that accumulated in the incubating medium. Although there was no significant differences between the successive medium samples taken at the various time intervals, extracellular levels of LDH tended to increase as a function of time (Table 2).

Table 2 Effects of hypoxia and reoxygenation on LDH release into the extracellular medium
Full size table

Whereas hyperoxia alone did not enhance the release of LDH by the neurons, hypoxia increased significantly the amount of LDH in the extracellular fluid. It increased by 106% immediately at the end of the hypoxic insult, and by 223% 3 d later. In the hypoxia-hyperoxia group, LDH efflux from the cells also increased, but it appeared to reach lower levels at T = 72 h than in the hypoxia-normoxia group (p < 0.01).

Glucose transport and incorporation. As shown in Table 3, specific uptake of 2-[3H]DG by the cultured neurons was significantly increased after hypoxia, and such changes in the cell energy metabolism lasted for at least 24 h after reoxygenation. By contrast, 2-[3H]DG uptake decreased by 23% below basal values 3 d posthypoxia, when morphologic alterations became apparent. A similar profile of glucose transport was recorded after hypoxia-hyperoxia, although the reduction of 2-[3H]DG uptake at T = 72 h was smaller than after hypoxianormoxia. Finally, a transient exposure of the neuronal cells to hyperoxia enhanced glucose transport and incorporation (22% at T = 24 h).

Table 3 Effects of hypoxia and reoxygenation on specific transport and incorporation of 2-[3H]DG into cultured cells
Full size table

Superoxide radical formation. In control neurons grown in our culture conditions, a basal level of superoxide radicals, as reflected by a specific reduction of acetyl-cytochrome c, was measured in the extracellular medium. It corresponds to 3.8 and 8.6 nmol of superoxide detected/100 μg of protein after 6 and 24 h, respectively(Table 4).

Table 4 Effects of hypoxia and reoxygenation on superoxide radical production in the cell surrounding medium
Full size table

After hypoxia, a significant increase in the radical formation was measured. It reached 53% over basal values at the end of exposure to hypoxia, and 108% the day after reoxygenation. Whereas hyperoxia by itself did not alter the cellular formation of superoxide compared with normoxic conditions, transient hyperoxia after hypoxia did not seem to exacerbate the hypoxia-induced production of superoxide, because the final amounts of superoxide radicals detected at T = 24 h in the hypoxia-hyperoxia culture group were not significantly different from those in the hypoxia-normoxia group. Unfortunately, because of the inability to keep cultured neuronal cells for several days in Krebs-Ringer buffer such as the one necessary for the measurement of acetyl-cytochrome c reduction, it was not possible to analyze superoxide production at T = 72 h.

DISCUSSION

Several studies have focused on the effects of oxygen concentration in postasphyxia resuscitation using various animal models(13–15) and in clinical investigation(16). They were primarily devoted to report survival rates or responses of physiologic parameters after the use of either 21 or 100% oxygen. The data generally support the conclusion that room air is as effective as 100% oxygen during resuscitation at birth, and adequate ventilation appears to be more important than oxygen supplementation. A critical question is whether posthypoxic oxygen supplementation has significant effects on brain damage when tissue is exposed to above normal oxygen pressures, because complex biochemical changes, especially free radical generation, are known to occur in the CNS during hyperoxia(17–19). In this respect, Huang et al.(15) showed that, although increasing the fractional concentration of inspired oxygen (Fio2) to 100% after asphyxia can improve cortical oxygenation, it leads to greater disturbance of brain dopamine metabolism which may contribute to cerebral injury.

To investigate the influence of transient hyperoxia after hypoxia on cell energy metabolism and superoxide production, the use of neuron-enriched cultures is particularly appropriate, because neurons are the most vulnerable cells to hypoxia-ischemia(20), and glial cells have been reported to protect neurons against anoxia-induced injury(21). When used in this study, neurons in culture are fully differentiated, express functional synapses, and have reached a developmental stage theorically comparable to that achieved in the newborn rat brain(22).

Po2 values in the incubating medium of such monolayer cultures was reduced by 75-80% under anaerobic conditions (95% N2/5% CO2) compared with controls. The decrease in Po2 was accompanied by a significant reduction in pH values, which probably reflects the accumulation of lactate and H+, secondarily to the shift from oxidative to anaerobic metabolism. Although some experimental evidence suggests that acidosis can alter the brain cell integrity, including that of newborns(23), it has been shown that anaerobic energy metabolism produces smaller amounts of lactate in immature than in adult brain tissue(24), and that, in neurons in vitro, hypoxia-induced changes in pH, lactate, and bicarbonate are not neurotoxic by themselves(25). One day after standard reoxygenation(95% air/5% CO2), the extracellular pH remained low, whereas Po2 was significantly higher than control values. Similar data have been reported previously by Sher(8) who attributed late Po2 changes to an increased tissue oxygen tension resulting from a posthypoxic decreased cerebral metabolic rate of O2.

The posthypoxic evolution of cell morphology and protein levels confirms the relative resistance of immature brain cells to asphyxia(26). Indeed, no cell alteration could be detected with our method within 24 h after the hypoxic insult when neurons were kept in serum-free culture medium or even in Krebs-Ringer solution. By contrast, LDH efflux appears as a good and early marker in reflecting the adverse effects of the reduction in oxygen supply, because LDH levels in the extracellular medium increased as soon as the end of exposure to hypoxia, and such an increase lasted for the whole experimental period, i.e. 72 h. When neuronal cells were maintained for 24 h in Krebs-Ringer for the determination of superoxide radicals, LDH was shown to increase more sharply (20-22% over values measured in serum-free culture medium), undoubtedly because the incubation medium in these experiments was less optimal than usual. Marked morphologic alterations, including neurite fragmentation and cell lysis, that were concomitant with a significant loss in protein contents, were currently observed only 3 d posthypoxia, suggesting that neurons in primary culture can develop “delayed cell injury,” as already described in various animal models in vivo(27).

Under severe hypoxic conditions, there is a rapid failure in ATP generation, with a concomitant activation of anaerobic glycolysis. It has been demonstrated that glucose is the primary energy substrate for the brain in fetuses and newborns as well as in adults(28, 29). Although other substrates such as ketone bodies may provide alternate fuels in some pathologic conditions, their consumption requires oxygen and thus only glucose is capable of sustaining brain energy metabolism, via anaerobic glycolysis, in hypoxia-ischemia. Therefore, the increase in 2-[3H]DG incorporation into the cultured cells which was observed immediately at the end of the exposure to hypoxia probably reflects the activation of glycolysis, as previously reported in vivo(30). Several hours after cell reoxygenation, 2-[3H]DG-specific transport remained high above control values. Such an observation might be related to a cellular hypermetabolism, maybe consecutively to specific changes in neurotransmitter systems, especially excitatory amino acids that are known to participate to death of hypoxic neurons(31).

The role of the superoxide radical in hypoxic-ischemic brain tissue injury and edema has been well documented(32). Although there is only little information regarding the relationship between free radical formation and perinatal hypoxic-ischemic brain damage(3), superoxide radicals have been detected in the extracellular fluid of the newborn pig brain during postischemic reperfusion(33), and it has been shown that these compounds formed within the cells can reach the extracellular space(34). Whereas control neuronal cultures produced low levels of superoxide, a 6-h exposure to hypoxia increased significantly the free radical amounts measured in the incubation medium. Accordingly, it has been demonstrated that the formation of superoxide radicals does not occur only during tissue reoxygenation, but also under physiologic conditions as well as during hypoxia, when cells are exposed to very low oxygen concentrations(35). Three hours after the onset of reoxygenation and then on the following day, the production of superoxide increased sharply, and this phenomenon may account for final cell injury. The mechanisms that are known to be involved in the formation of free radicals include 1) activation by cytosolic calcium of intracellular oxidases, especially xanthine oxidase which converts hypoxanthine to xanthine(36), 2) release of excitatory amino acids and subsequent stimulation of their receptors(37), and3) impairment of mitochondrial energy conservation(38). In their study, Littauer and de Groot(39) reported that the release of reactive oxygen species, as reflected by the formation of hydrogen peroxide, during reoxygenation of cultured hepatocytes originates from the mitochondrial respiratory chain when the O2 reintroduction increased from 0 to 2%, whereas other sources, i.e. nonenzymatic ones, contribute to a lesser extent to the generation of such compounds when O2 content is further increased. In addition, free radical detoxifying enzymes,e.g. glutathione peroxidase and superoxide dismutases, are inhibited during and after hypoxia(40).

During reoxygenation of the cultured neurons with hyperoxic gas mixture(95% O2/5% CO2), the mean Po2 in the incubating medium increased from 30 to 480 mm Hg. When compared with standard reoxygenation, there was no apparent difference in the morphologic outcome of the neurons. Under both reoxygenation conditions, cellular alteration became apparent only 3 d posthypoxia, and then neurons exhibited vacuolization and neurite fragmentation, concomitantly with decreased energy metabolism and enhanced LDH efflux. Moreover, final protein concentrations, glucose incorporation, and LDH leakage tended to be less altered when neurons were submitted to transient hyperoxia during reoxygenation.

A burst of reactive oxygen-derived radical production can be generated in the CNS as soon as reoxygenation occurs(41), and one may expect that such a process is amplified during posthypoxic hyperoxia. Surprisingly, superoxide amounts detected in the cell surrounding medium were not enhanced by hyperoxia, suggesting that “normal” reoxygenation is as effective as hyperoxia to generate these radicals. It should be noticed, however, that in all culture models, normoxia is classically achieved by incubating cells in 95% air/5% CO2 gas mixture, leading to oxygen tension around 140 mm Hg. It is clear that this value is significantly higher than the oxygen tension in the brain of newborn animals, even when 100% oxygen was given previously(42). This may be important as it has been shown in liver and kidney in vitro that lipid peroxidation increased sharply at oxygen tension between 10 and 50 mm Hg(43). On the other hand, in vivo animal studies after brain ischemia have provided conflicting results, and it was postulated by Agardh et al.(44) that generation of free radicals during reoxygenation is critically dependent on the extent of ischemia, but not on the different oxygen tensions.

The data obtained in the present study indicate that transiently elevated oxygen concentrations during reoxygenation does not appear to be more deleterious for the cell integrity, at least in newborn brain neurons. Such observations are in accordance with the well known higher resistance of neonates to the lethal effects of exposure to high oxygen levels(45), and also correlate with the hypothesis that sudden normoxic reoxygenation results in near maximal oxygen radical production. However, in our work, only the direct effects of hypoxia/reoxygenation on the neurons were tested, and other cell types may also be important for the ultimate brain damage. For instance, microglial cells, forming a network of potential immunoeffector cells, have been shown as a cellular source of oxygen free radicals(46). In addition, the enzyme system xanthine dehydrogenase/oxidase is enriched in brain capillaries, and it was assumed that endothelial cells and the microvasculature are especially prone to oxygen radical damage, although some studies indicated that xanthine oxidase is not the predominant source of radicals in the ischemic brain, and microvasculature is not the primary target of radical-mediated damage(47).

Recently, evidence has been provided that the overall amount of oxygen received by the newborn may be less detrimental than fluctuating changes in oxygen concentrations. Indeed, with respect to the pathogenesis of retinopathy of prematurity, it has been shown that excessive oxygen alone is not the single overriding factor, and that fluctuating changes in oxygen pressure play a critical role(48). Our data are in good agreement with such a hypothesis regarding hypoxia-induced neuronal cell damage. It is conceivable that alternating hypoxia and reoxygenation may cause more severe injury. Also, it may be hypothesized that our observations may be related to an hyperoxia-induced increase in antioxidant enzyme activities in the neonatal brain, as already shown in the lung(49), and further investigations are now required to answer this question.