Main

Transport from blood into the brain is mediated by a diffusion-type transport system, i.e. members of the GLUT supergene family (1, 2). This stereospecific set of integral membrane proteins carries out the facilitated diffusion of D-glucose across the plasma membrane. Tissue-specific expression of one or more members determines in part the net rate of glucose entry into the cell. Structure, detailed function, and regulation are well documented and reviewed (36). GLUT1, which is expressed and localized at the endothelial cells of the blood-brain barrier, takes the first step in the transport of glucose from the blood into the tissue layers (7). The next step of glucose transport from extracellular space into neuronal cells is taken by GLUT3, localized at the neuronal cell membrane (4). After it is within the cell, glucose is phosphorylated to glucose-6-phosphate by HX, a key enzyme for glucose utilization in the cell. The second major rate-limiting enzyme in glucose metabolism and glycolysis is phosphofructo-1-kinase (PFK), which catalyzes the ATP-dependent phosphorylation of fructose-6-phosphate.

Glucose transport and metabolism are of utmost importance in IS (ischemia, hypoxia, and asphyxia) and have been studied extensively. Deterioration of glucose transport and glucose metabolism in IS has been shown in vitro (811) and in vivo with conflicting results. The outcome of studies was not comparable, as adult IS differ significantly from perinatal IS (1220). Different time points studied by the individual research groups in various animal models of IS do not allow the comparison of results or the drawing of a general conclusion on glucose transport and metabolism in IS. Data on glucose transport and glucose-metabolizing enzymes in the same system at the same time and in the early phase, which may be crucial for survival and the development of brain damage, have never been reported; this prompted us to investigate these mechanisms in a nonsophisticated animal model of PA.

We induced graded asphyxia in neonatal rat pups up to 20 min postpartum and determined brain glucose, CBF, GLUT1, GLUT3 mRNA and protein, as well as HX and PFK mRNA and activity. These parameters should be useful in answering the question of glucose transport and utilization during the early phase of PA in the brain.

METHODS

Animals and experimental design.

We used a well-documented nonsophisticated rat model of PA resembling the clinical situation (2126). Asphyxia was induced in pups delivered by cesarean section on pregnant Sprague-Dawley rats. Within the last day of gestation, as determined by established protocols, animals were killed by neck dislocation and hysterectomized. The uterus horns, still containing the fetuses, were extirpated and placed in a water bath at 37°C for various periods ranging from 5 to 20 min. Cesarean-delivered control and asphyctic pups were obtained from the same mother, because each rat delivered approximately 10–14 pups. After the asphyctic period, i.e. incubation at 37°C, the uterus horns were rapidly opened and the pups were removed. Pups were cleaned, the umbilical cord was ligated, and the animals were allowed to recover on a heating pad. Only litters with pups weighing >4.5 g at the time of delivery were used in the experiments. The animal studies were conducted according to the guidelines of the American Physiologic Society. Animals used for biochemical studies were killed by decapitation 10 min after delivery, and the brain was kept at −80°C until assays were performed.

Ten pups per group (normoxia and 5, 10, 15, and 20 min of asphyxia) were studied. In the group with normoxia and the groups with asphyxia of 5 and 10 min, 100% survived; in the 15-min group, 89%; and in the group with 20 min of asphyxia, 74%. At 21 min (not used for the experiments) of asphyxia, only 10% of the pups survived. All of the following methods were conducted on 10 brains per group (1 = normoxia, 2 = 5 min, 3 = 10 min, 4 = 15 min, and 5 = 20 min of asphyxia).

Measurement of brain tissue pH.

pH was measured in the brain of pups frozen 10 and 40 min after the asphyctic period by inserting the tip of a tissue pH electrode (2 mm in diameter; TFK 325/HC; pH 320 set, Wissenschaftlich-Technische Werkstaetten, Vienna, Austria) into the whole tissue kept at 4°C. Measurements were performed in triplicate with pepsin-distilled water rinsing cycles (26).

Measurement of protein in homogenate supernatant.

Protein was evaluated in supernatants of homogenized tissue samples (26) by the Bradford method (27).

Measurement of glucose in homogenate supernatant.

Glucose was measured in 4000 ×g supernatants of brain homogenates by a standard glucose oxidase method on a Kodak autoanalyzer.

Measurement of CBF.

A method specifically developed for measurements of CBF in small immature rats was used (25). Briefly, animals were injected subcutaneously with iodo-[14C]antipyrine (5 μCi), and, after a variable period, each rat pup was decapitated. Brains were removed immediately and prepared for isotopic counting.

mRNA isolation from brain/Northern blots and dot blots.

The organs were obtained at autopsy and placed into liquid nitrogen. Frozen brain samples were ground, and mRNA extraction was performed using the Quick Prep Micro mRNA Purification kit (Pharmacia). Subsequently, 6 μg of mRNA each, dissolved in 6 μL of diethylpyrocarbonate (DEPC) water, 11.8 μL of DMSO, 2.4 μL of 0.1 M NaHPO4 buffer, and 3.5 μL deionized 40% glyoxal was incubated for 1 h at 50°C and subsequently cooled down on ice. To this solution, 6.3 μL of a mixture of 50% glyoxal, 10 mM sodium phosphate, and 0.4% bromophenol blue was added. The sample was applied onto 1.2% agarose gel (29) and electrophoresed at 3–4 V/cm for 2.5 h in circulating 0.01 M phosphate buffer, pH 7.0. RNA was then transferred to a positively charged nylon membrane (Hybond N+, Dupont, NEF 986) by capillary blotting (30), fixed with 0.05 N sodium hydroxide for 5 min at room temperature, and finally equilibrated at pH 7.0 with three washes in 2 × sodium salt citrate (SSC).

The cDNA probes purchased were β-actin (ATCC 9800), GLUT1 (ATCC 59631), GLUT3 (ATCC 61614), HK1 (HX, ATCC 61513), and 6-PFK (ATCC 86428); these probes were used for Northern and dot blots. Nonradioactive labeling of probes was performed as described recently (31). After fixation of bound RNA, the nylon membrane was incubated in prehybridization solution [0.25 M phosphate buffer, pH 7.2, containing 5% SDS wt/vol, 1 mM EDTA, and 0.5% blocking reagent (from Dupont NEL 203)] for 12 h at 65°C in a hybridization oven. The blots were hybridized (50 ng/mL of prehybridization buffer) overnight at 65°C with the labeled probes. After hybridization, nonspecifically bound material was removed by posthybridization washes with 0.5X and 0.1X prehybridization buffer, 2 × 10 min each. The 0.5X and 0.1X prehybridization buffer was brought up to 65°C before use, and the second wash was performed at room temperature.

Hybridized blots were blocked with 0.5% blocking reagent in 0.1 M Tris-HCl, pH 7.2, and 0.15 M NaCl for 1 h at room temperature. Membranes were then incubated with antifluorescein horseradish (HRP) antibody (Dupont NEL 203) at a 1:1000 dilution in the solution given above for 1 h under constant shaking. Membranes were washed four times for 5 min each in the solution given above. The nucleic acid chemiluminescence reagent (Dupont NEL 201) was added to the membranes and incubated for 1 min. Excess detection reagent was removed by the use of filter paper, and the membrane was placed in Saran Wrap (Bio-Rad, Vienna, Austria) and exposed to autoradiography reflection films (Dupont NEF 496) for 15 min at room temperature.

Dot blots were performed according to the method of White and Bancroft (32). One microgram of mRNA in 10 μL of DEPC-H2O was added to 30 μL of a mixture of 500 μL of deionized formamide, 162 μL of 37% formaldehyde, and 100 μL of 10X MOPS. This solution was brought to 65°C for 5 min and cooled down on ice for 3 min. A nylon membrane (Hybond N, Amersham) was moistened in 10X MOPS and placed on two layers of moistened Whatman 3MM paper in the dot-blot apparatus. The chamber was closed tightly, and 40 μL of each sample was applied under vacuum. Under vacuum, 0.05 M NaOH was added; the nylon membrane was removed from the dot-blot apparatus and again incubated with 0.05 M NaOH for 5 min. Finally, the nylon membrane was equilibrated in 2 × SSC to a pH of 7.5.

Prehybridization was performed using 0.1 mL of prehybridization solution [5 × SSC, 0.1% (wt/vol) SDS, 0.5% (wt/vol) blocking reagent (Dupont), 5% dextran sulfate; dissolved at 60°C and used at room temperature] per square centimeter of membrane, and the volume of the probe was 20 ng/mL, wrapped in a plastic bag. Subsequent washing steps and detection of the products were performed according to the supplier's instructions (RenaissanceR Random Primer Fluorescein-Labeling kit; Dupont). Densitometry of films was performed using the Hirschmann Elscript 400 densitometer (Germany).

Western blots for the determination of GLUT1 and 3 protein.

Western blots were performed using the method given in a previous publication (33). Antibodies applied were rabbit anti-GLUT-1 and rabbit anti-GLUT-3 (Chemicon International; Temecula, CA). Densitometry was performed as above.

Determination of HX activity.

PFK catalyzes the phosphorylation of fructose-6-phosphate by ATP to fructose-1,6-diphosphate. In the assay used, fructose-1,6-diphosphate is measured by the conversion to dihydroxyacetone phosphate through the aldolase and triose phosphate isomerase reactions. The dihydroxyacetone phosphate is then reduced by the action of α-glycerophosphate dehydrogenase, oxidizing NADH to NAD, which is measured at 340 nm. The protocol of Beutler was followed (34).

Determination of PFK activity.

HX catalyzes the reaction ATP + glucose → glucose-6-phosphate + ADP.

We used the standard assay by Beutler (34) in which glucose-6-phosphate is measured by linking its further oxidation to 6-phosphogluconate to the reduction of NADP through the glucose-6-phosphate dehydrogenase reaction: glucose 6-phosphate + NADP+ → 6-phosphogluconate + H+.

Statistical methods.

ANOVA with subsequent Kruskal-Wallis test and t test was applied, and a level of p < 0.05 was considered significant. Linear regression analysis was used for the determination of correlations.

RESULTS

Results of brain pH, brain tissue glucose levels, CBF (Fig. 1), and HX and PFK activity are shown in Table 1. Brain pH decreased; brain tissue glucose levels and CBF increased with the length of the asphyctic period. HX and PFK activities did not change during the observation period. mRNA steady state levels for GLUT1, GLUT3, HX, PFK, and the housekeeping gene β-actin were comparable between groups (Table 2). Western blots did not show any differences in GLUT1 and 3 protein levels between groups (Table 2).

Figure 1
figure 1

Steadily increasing CBF correlating with brain tissue glucose levels and inversely correlated with brain pH is shown.

Table 1 Results of brain pH, brain glucose levels, and CBF, HX, and PFK activity
Table 2 Results of brain glucose transporters, hexokinase, and phosphofructokinase

There was a significant positive correlation between the parameters brain tissue glucose and CBF (r = 0.87, p < 0.01) and a significant negative correlation between the parameters brain pH and brain tissue glucose (r = −0.82, p < 0.01) and the parameters brain pH and CBF (r = −0.91, p < 0.001).

DISCUSSION

Brain glucose gradually increased with the length of the asphyctic period, a finding compatible with reports in the literature (35) and the probable biologic meaning of supplying fuel for brain energy metabolism (36). Our concomitant finding of increased CBF paralleling brain glucose levels would fit and complement this hypothesis (3740). Increased brain glucose in IS may be damaging or protective (41), but this question is still open, and the basic pathophysiologic knowledge for the final understanding of the contribution of brain glucose levels to brain damage in PA has not been provided yet.

Our data on GLUT1 at the protein and transcriptional level may show that during the asphyctic periods, no changes in glucose transport from the blood into the brain occur. In vitro studies are in agreement with our findings: Takao et al. (42) clearly showed that both GLUT1 mRNA and protein were decreased only after prolonged exposure (days) to high glucose in the medium. Schroeder et al. (43) found no detectable changes of GLUT1 and 3 expression in brain of near-term fetal hyperglycemic rats. McCall et al. (19) studied an adult model of forebrain ischemia; by an immunocytochemical technique, they found increased GLUT1 protein after 24 h. Thus, their observation does not contradict our data. Lee and Bondy (13) used in situ hybridization to show changes of GLUT1-GLUT4 after middle cerebral artery occlusion; they found a significant increase of GLUT1 mRNA as soon as 1 h after the insult. Again, this finding does not contradict our data, as a different age group (mature system), a vascular occlusion model resembling stroke and not PA, was used and no protein levels were determined.

Urabe et al. (12) used an adult model of cerebral artery occlusion and found increased GLUT1 mRNA after hours in vascular cells and neurons, which was, however, not accompanied by increased GLUT1 protein. Also, this model would not challenge our observation of unchanged GLUT1 mRNA and protein, as graded PA in our animal model was studied until 20 min; asphyxia-asphyctic rat pups in our study did not survive longer periods. Devaskar et al. (44) have clearly shown significant differences between quantity and distribution of GLUT1 expression in mature and neonatal animals.

Vannucci et al. (20) applied an immature model of unilateral brain damage to answer the question of GLUT1 and 3 expression using Western blotting. Four hours after the insult, GLUT1 and GLUT3 protein were approximately 20% higher in both hemispheres, accompanied by an increase of brain glucose. Their findings may parallel ours, taking into account that the increase in GLUT was found hours after the insult and that 7-d-old pups were used, animals that are fully adapted to postnatal live and may not be assigned to the (age, maturation) group of PA. Their study also shows that glucose transport into brain cells as represented by GLUT3 expression may not be moderately increased before 4 h after the IS.

This view is supported by the article of Gerhart et al. (14) who demonstrated ischemia-induced GLUT1 and 3 expression after 3 h in young gerbils. Our own data for GLUT3 mRNA and protein showed no changes in the course of the asphyctic periods. The biologic meaning of unchanged GLUT in our model of PA is that glucose transport is neither reinforced nor hampered and may not serve as a limiting factor for glucose availability to the brain. We further learn that in IS facilitated glucose transport is neither affected by hypoxia per se as proposed by Ebert et al. (8) nor by inhibition of oxidative phosphorylation as suggested by Behrooz and Ismail-Beigi (11). We do not agree, however, that GLUT1 and 3 are related to glucose utilization as proposed by Vannucci et al. (20), as they did not test glucose utilization directly. We can contribute to this open question in our model of PA, as two enzymes involved in glucose utilization were assayed. HX and PFK are key enzymes of glucose utilization regulated at the transcriptional as well as at the activity level (4547) and presenting with low perinatal expression in the brain (48).

HX mRNA and activity as well as PFK mRNA and activity did not change during the asphyctic period and were comparable to the normoxic state in our experiments. The main question arising is why high glucose did not induce enzymes of glucose utilization. On the other hand, the decreasing brain pH must be taken into account for the interpretation of glucose-handling enzymes: the shift from the pH optimum of HX and PFK may have been counteracting their activation by increased glucose levels. It is intriguing that the significant decrease of brain pH observed with increasing length of asphyxia (Fig. 1) did not reduce transcription per se as reflected by the housekeeping gene nor specific transcription glucose-handling enzymes or GLUT.

We conclude that brain GLUT1 and 3 were comparable in graded asphyxia and normoxia, suggesting that during the early asphyctic period glucose transport may not be deteriorated despite hypoxia, accumulation of toxic metabolites (49), and acidosis. It must be taken into account, however, that transporters actually present at the plasma membrane were not determined. Brain glucose utilization as reflected by rate-limiting key enzymes HX and PFK in turn was not altered by the asphyctic state, a finding never described before and one that may serve for the interpretation of glucose handling and metabolism in PA.