Main

The long-term neurologic morbidity for preterm and ill term human newborns remains a serious concern. Among these infants, hypercapnia is a common occurrence. Current clinical practice often allows infants ventilated acutely or chronically to remain hypercapnic (undergoing “permissive hypercapnia”) to limit ventilator-induced lung injury. The effects of hypercapnia on the newborn brain are controversial as there is not a known safe range of arterial CO2 pressure (Paco2).

Some recent clinical studies have not shown an immediate harmful effect of hypercapnia in human infants (1–3). However, in most of these studies, long-term neurologic outcomes were not reported. In addition, subtle alterations in brain structure and function may not be detectable in these clinical studies.

Hypercapnia (Paco2 >60 mm Hg) results in increased CBF and decreased cerebral vascular resistance in newborn infants (4–6) and has been associated with intraventricular hemorrhage in some studies (7–9) but not in others (10,11). There are some clinical data suggesting that hypercapnia may be harmful to the newborn brain even if it does not result in intraventricular hemorrhage (12–14). Moderate elevations in Paco2 (end tidal Pco2 of 60 mm Hg) in healthy preterm (33–36 wk) infants significantly impeded brain stem auditory evoked responses (15). In other studies, very-low-birth-weight infants and term infants with impaired CBF reactivity during hypercapnia had poor neurologic outcomes or developed hypoxic ischemic encephalopathy (16,17).

During hypercapnia, neuronal injury may result from alterations in cerebral blood flow, decreased extracellular and intracellular pH, the generation of oxygen free radicals, and/or increased intracellular and intranuclear Ca++ flux. In neurons, increased intracellular Ca++ flux may be due to acidosis-induced activation of the cerebral NMDA receptor (18). Intracellular Ca++ activates proteases, phospholipases, and nitric oxide synthase (19,20). Activation of these enzymes may result in the generation of oxygen free radicals and in the peroxidation of nuclear membrane lipids and an increase in intranuclear Ca++-influx. Intranuclear Ca++ forms a complex with calmodulin that initiates a signal cascade by activating various protein kinases including CaMK IV (21). Nuclear CaMK IV activates transcription factors including CREB by phosphorylating CREB at its active site Ser133, which then allows it to bind to DNA and results in the transcription of apoptotic genes (22), such as Bax. If elevated levels of Bax protein increase the ratio of proapoptotic to antiapoptotic (Bcl-2) proteins, cell death will be promoted. We propose that hypercapnia induces an increase in CaMK IV activity that may result in the subsequent phosphorylation of CREB protein and increased expression of Bax protein in the newborn brain.

To begin to examine the effects of normoxic hypercapnia on these early steps of transcription and apoptosis in the newborn brain, we studied anesthetized, ventilated newborn piglets exposed to two levels of increased arterial Paco2 for 6 h. In the present study, we tested the hypothesis that hypercapnia alters nuclear function in the newborn brain and specifically that hypercapnia results in a decrease in cerebral energy metabolism, activation of CaMK IV, phosphorylation of CREB, and increased expression of the proapoptotic protein Bax in the cerebral cortex of newborn piglets.

METHODS

Studies were performed in three groups of anesthetized, ventilated 1- to 3-d-old piglets, six normocapnic, seven with a Paco2 of 65 mm Hg, and seven with a Paco2 of 80 mm Hg. The experimental protocol was approved by the Institutional Animal Care and Use Committee of the Drexel University College of Medicine. Anesthesia was induced with 4% halothane and maintained with 0.8% halothane. Lidocaine 1% was injected locally for performance of a tracheostomy and insertion of aortic and inferior vena caval catheters. Intravenous fentanyl (10 μg/kg initially and every hour) and tubocurarine (0.3 mg/kg) were given and the animals placed on a pressure ventilator using 75% nitrous oxide and 25% oxygen. Arterial blood pH, Pao2, Paco2, glucose, lactate, heart rate, and blood pressure were recorded at least every 15 min in all animals. Temperature was maintained with a warming blanket at 38–39°C rectal. After a 1-h period of stabilization, the piglets were divided into the following three groups: those ventilated for 6 h with a Paco2 40 ± 2 mm Hg and a Pao2 80–100 mm Hg (normocapnic or CO2 40 group), those ventilated with CO2 added into the respiratory circuit to maintain a Paco2 of approximately 65 ± 2 mm Hg and Pao2 80–100 mm Hg (CO2 65), and those ventilated with added CO2 to maintain a Paco2 of approximately 80 ± 2 mm Hg (CO2 80). Circuit CO2 values were monitored continuously in all animals. Piglets were maintained on D5 0.5 normal saline with calcium gluconate 100 mg/100 mL at 100 mL/kg/d. At the end of the experimental period, the brains were removed and within 4 s were placed either in buffer for the isolation of cell nuclei or in liquid nitrogen and stored at −80°C for biochemical analysis.

Brain concentrations of ATP and PCr were determined by a coupled enzyme reaction as described by Lamprecht (23).

Cerebral cortical neuronal nuclei were isolated by homogenizing cortical tissue in 15 volumes of a medium containing 0.32 M sucrose, 10 mM Tris-HCl (pH 6.8), and 1 mM MgCl2, filtered through nylon cloth (mesh 100) and centrifuged at 850 × g for 10 min. The pellet was resuspended and mixed with a medium containing 2.4 M sucrose, 10 mM Tris-HCl (pH 6.8), and 1 mM MgCl2 to achieve a final concentration of 2.1 M sucrose at which neuronal nuclei settle. The nuclear fraction was purified by centrifugation at 53,000 × g for 60 min, a modification of the Guifrida et al. (24) method. The nuclear pellet was suspended in the medium (0.32 M sucrose, 10 mM Tris-HCl buffer, pH 6.8, and 1 mM MgCl2) and the purity of neuronal nuclei was assessed by a phase contrast microscope (Olympus, Tokyo, Japan). The neuronal nuclei were characterized by the presence of a centrally located nucleolus (one nucleolus/nucleus) compared with the presence of multiple nucleoli in the astrocytic and oligodendrocytic nuclei. The final nuclear preparation was devoid of any microsomal, mitochondrial, or plasma membrane contaminant, with a purity for neuronal nuclei of ≥90% (25).

CaMK IV activity was determined as previously described (26) using 10 μL of a 20 μcg/10 μL preparation of neuronal nuclei for the reaction mixture by 33P incorporation (2 min at 37°C) into syntide-2 in 25 μL of a medium containing 50 mM HEPES (pH 7.5), 2 μM DTT, 40 μM syntide-2, 10 mM Mg acetate, 5 μM PKI 5–24 (protein kinase A inhibitor), 2 μM PKC 19–36 (protein kinase C inhibitor), 1 μM microcystin-LR (protein phosphatase 2A inhibitor), 200 μM sodium orthovanadate (inhibitor of ATPase, alkaline phosphatase, protein tyrosine phosphatase), 0.2 mM ATP, 1 μCi 33P-ATP, and either 1 μM calmodulin and 1 mM CaCl2 (for total activity) or 1 mM EGTA (for Ca++/CaM independent activity). Twenty microliters of phosphorylated peptide medium was placed on phosphocellulose P81 membranes, washed, dried, and 33P gamma labeled phosphorus and radioactivity counted. The difference in the presence and absence of Ca++/calmodulin was calculated and the enzyme activity was expressed as picomoles per milligram protein per minute.

CREB phosphorylation, Bax, and Bcl-2 protein density were determined through Western blotting techniques (27). Neuronal nuclear membranes were prepared as described above in the presence of protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 10 g/mL aprotinin, and 0.2 mM sodium orthovanadate). Protein content was determined by the method of Lowry et al. (28), and the nuclear membrane preparation was diluted to a final concentration of 100 μg/100 μL. Five milliliters of Laemelli buffer were added to each 20 μL of nuclear membrane protein. Equal amounts of each neuronal nuclear protein (35 μL of the 100 μg/100 μL preparation) were separated by 12% SDS-PAGE and transferred electrophoretically to nitrocellulose membranes. Nitrocellulose membranes in duplicate were then blocked with 8% nonfat milk in PBS. The membranes were subsequently incubated with anti-phosphorylated (Ser133) CREB protein antibody (Upstate Biotechnology, Lake Placid, NY) or with antibodies specific to Bax or Bcl-2 (Santa Cruz Biotechnology, Santa Cruz, CA). Recombinant human Bcl-2 BV-lysate (BD PharMingen, San Diego, CA) and human promyelocytic leukemia [ATCC (R) F-13737 240-CCL HL-60 018554, HL-60 promyelocytic leukemia human peripheral blood, Rockland, Gilbertsville, PA] served as positive controls for Bcl-2 and Bax proteins, respectively. Immunoreactivity was detected by incubation with horseradish peroxidase-conjugated secondary antibody (Rockland). Specific complexes were detected by enhanced chemiluminescence method using the ECL detection system (Amersham Pharmacia Biotech, Little Chalfont, Buckinghamshire, UK) and analyzed by imaging densitometry (GS-700 densitometer, Bio-Rad, Hercules, CA). The densitometric scanning data were expressed as autoradiographic values (OD × mm2) per immunoblot protein representing CREB protein phosphorylation at Ser133, or Bax, or Bcl-2 protein density. In another set, CREB protein expression was also determined by anti-CREB antibody using the above method.

All data were reported as mean ± SD. Physiologic data were analyzed using an area transformation of each piglet's data using SigmaPlot (Systat Software, Inc., Point Richmond, CA). After transformation, the data were averaged for the normocapnic and the hypercapnic groups. Statistical analysis between the three groups was performed by one-way ANOVA with subsequent Tukey tests for pair-wise comparisons of the mean responses of the different treatment groups. p Values of < 0.05 were considered statistically significant.

RESULTS

There was a significant decrease in arterial pH in both the CO2 65 (7.30 ± 0.06) and the CO2 80 (7.25 ± 0.01) groups compared with the CO2 40 group (7.42 ± 0.07) as demonstrated in Table 1. Although the pH was lower in the CO2 80 group than the CO2 65 group, there was not a statistically significant difference between the two groups. Arterial Paco2 levels (mm Hg) were significantly higher in the CO2 80 group (81 ± 4) than in the CO2 65 group (68 ± 1), and both were higher than the CO2 40 group (42 ± 3). There was no difference in Pao2, mean blood pressure, or heart rate between the three groups as demonstrated in Table 1.

Table 1 Physiologic data of hypercapnic and normocapnic newborn piglets
Full size table

ATP levels (μmol/g wet brain) were lower in the both the CO2 65 (3.85 ± 0.41) and 80 (4.0 ± 0.14) groups than in the CO2 40 group (4.65 ± 0.27), p < 0.05 for both. Tissue PCr levels (μmol/g brain) were also lower in both the CO2 65 group (3.27 ± 0.41) and the CO2 80 group (3.18 ± 0.17) than in the CO2 40 group, (3.85 ± 0.20), p < 0.05 for both. Tissue levels of both ATP and PCr were not different between the CO2 65 and the CO2 80 groups.

CaMK IV activity (pmol/mg protein/min) as shown in Figure 1 was higher in the CO2 65 (1685 ± 188) and the CO2 80 groups (1529 ± 198) than in the CO2 40 group (1251 ± 100) group. CaMK IV activity was not different between the CO2 65 and the CO2 80 groups.

Figure 1
figure 1

Neuron nuclear CaMK IV activity (pmol/mg protein/min) for the three groups, CO2 40, CO2 65, and CO2 80, in the cerebral cortex of newborn piglets. *p < 0.05 vs Paco2 40 mm Hg.

Full size image

There was an increase in density of phosphorylated CREB protein in both the CO2 65 and the CO2 80 groups compared with the CO2 40 group as demonstrated in Figure 2, a representative gel of phosphorylated CREB protein. The data show that the density (expressed as OD × mm2) of phosphorylated CREB protein increased from 103 ± 14 in the CO2 40 group to 158 ± 12 in the CO2 65 group (p < 0.05) and 150 ± 15 in the CO2 80 group (p < 0.05 versus CO2 40). There was no difference in phosphorylated CREB density between the CO2 65 and the CO2 80 groups. In another set of experiments, CREB protein expression did not change in the CO2 65 or the CO2 80 group compared with the CO2 40 group.

Figure 2
figure 2

Representative Western blots of phosphorylated CREB Ser133 protein expression in CO2 40, CO2 65, and CO2 80 groups in neuronal nuclei of newborn piglets. The bands were visualized at a molecular weight of 43 kD.

Full size image

Representative Western blots of Bax and Bcl-2 protein expression in CO2 40, CO2 65, and CO2 80 piglets are shown in Figure 3, a and b, respectively. Results demonstrate an increase in Bax protein density in the CO2 65 and CO2 80 groups compared with the CO2 40 group but no change in expression of Bcl-2 protein for the three groups. Bax protein expression (OD × mm2) was significantly higher in the CO2 65 group (194 ± 28) and the CO2 80 group (168 ± 18) compared with the CO2 40 group (95 ± 12), p < 0.05 for both versus CO2 40. There was no difference in Bax protein expression between the CO2 65 and CO2 80 groups. Bcl-2 protein (OD × mm2) was similar in all three groups, 114 ± 32 in the CO2 65 group, 100 ± 35 in the CO2 80 group and 114 ± 57 in the CO2 40 group. The ratio of Bax/Bcl-2 was higher in the CO2 65 (1.83 ± 0.66) and CO2 80 (1.86 ± 0.75) groups than in the CO2 40 group (0.91 ± 0.18), p < 0.05 for both. There was no difference in the ratio of Bax/Bcl-2 protein in the CO2 65 and the CO2 80 groups.

Figure 3
figure 3

Representative Western blots of Bax (a) and Bcl-2 (b) protein expression in neuronal nuclei of newborn piglets from the CO2 40, CO2 65, and CO2 80 groups.

Full size image

DISCUSSION

In the present study, we have described the effects of two levels of elevated Paco2 on the newborn brain, a Paco2 of 65 mm Hg and a Paco2 of 80 mm Hg. The data demonstrate a decrease in high-energy phosphates, an increase in CaMK IV activity, an increase in phosphorylation of CREB protein, and increased Bax protein expression in piglets exposed to either a Paco2 of 65 mm Hg or a Paco2 of 80 mm Hg compared with the normocapnic piglets.

The results demonstrate that hypercapnia (either a Paco2 of 65 mm Hg or a Paco2 of 80 mm Hg) resulted in an increase in nuclear CaMK IV activity. The increase in CaMK IV activity seen during hypercapnia may be due to hypercapnia-induced intracellular acidosis. The isoelectric point of CaMK IV is 4.65 (29), indicating that the enzyme will be fully charged in an acidotic environment. Indeed, activity of CaMK IV is increased during acidosis (30). In the carboxy region of the molecule, there is a stretch of amino acids of which 81% are glutamate residues including a string of 11 glutamate residues (31). During hypercapnia-induced acidosis, these glutamate residues may be fully ionized allowing maximal enzyme activity.

In neuronal nuclei, CaMK IV is the predominant form of CaM kinase (32) and results in the phosphorylation of CREB protein and Ca++-dependent gene transcription (21,22). Transcription is activated only if CaMK IV is present (33). CaM kinases are activated by binding to a Ca++/calmodulin (Ca++/CaM) complex, which induces a conformational change exposing the active site of the kinase previously blocked by the autoinhibitory domain (21,34). The hypercapnia-induced increase in CaMK IV activity may result in the subsequent phosphorylation of CREB protein and increased expression of Bax protein in the newborn brain.

During both levels of hypercapnia there was an increase in phosphorylated CREB Ser133 protein density in the cerebral cortex of newborn piglets compared with normocapnic piglets. Activation of the cerebral NMDA receptor results in increased intracellular Ca++, which may potentiate Ca++ flux into the nucleus and increase CaMK IV activity and CREB phosphorylation. Indeed, increased CREB Ser133 phosphorylation is dependent on intranuclear Ca++ concentrations, and intracerebroventricular injection of NMDA potentiates binding of CREB protein to DNA in mice (35). In addition, both the phosphorylation and the activity of CaMK IV is increased by increasing intracellular Ca++ flux through glutamate-induced NMDA receptor activation (36). We speculate that hypercapnia-induced acidosis results in alteration of the NMDA receptor and increases NMDA receptor-mediated intracellular Ca++ influx, leading to an increase in nuclear Ca++ influx and phosphorylation of CREB protein in newborn piglets.

Phosphorylation of CREB protein results in gene transcription (37). CREB is a member of the leucine-zipper class of proteins. These proteins can form homodimers and heterodimers that bind to a regulatory DNA sequence known as the cAMP-responsive element (CRE) (37), resulting in the transcription of genes containing CRE sequences (38). The binding of CREB to CRE does not induce transcription unless phosphorylation of CREB at Ser133 has occurred (39). When CREB protein is phosphorylated, there is an increase in the spherical site and net-positive surface charge of the CREB/DNA complex. Therefore, it appears that phosphorylation of CREB protein alters its binding affinity, its secondary structure, and the charge characteristics of the molecule (40), all of which may be enhanced in an acidotic environment.

The study shows that during hypercapnia there was an increase in expression of Bax protein resulting in an increased ratio of the proapoptotic protein Bax to the anti-apoptotic protein Bcl-2 compared with normocapnic piglets. Both Bax and Bcl-2 proteins are expressed in neurons of the central and peripheral nervous system during brain development (41,42) and are critical regulators of programmed cell death. Bcl-2 protein is antiapoptotic, enhancing cell survival possibly through regulating cytosolic and intranuclear Ca++ concentrations (43). In contrast, the proapoptotic protein Bax has been shown to promote cell death by activating caspases, especially caspase-3 (44,45). The active form of Bcl-2 heterodimerizes with Bax and their ratio determines the cellular susceptibility to apoptotic stimuli (46,47). During hypercapnia, the increase in Bax protein and ratio of Bax/Bcl-2 indicates a propensity of the cells to undergo programmed cell death in the newborn piglet brain.

In addition to Bax, hypercapnia has been shown to induce the expression of other nuclear proteins as well. The nuclear protein FOS is expressed in brainstem chemosensitive neurons (48) during hypercapnia. FOS expression appears to be mediated by activation of MAP kinases, PKCα and PKCβ (49). Therefore, it appears that hypercapnia leads to the activation of kinases and the subsequent expression of neuronal apoptotic proteins.

We have postulated that during hypercapnia the ensuing respiratory acidosis alters brain cell membranes and results in increased CaMK IV activity, phosphorylation of CREB protein, and Bax protein expression in the cerebral cortex of newborn piglets.

During hypercapnia, there is a decrease in extracellular and intracellular pH (50–52). The hypercapnia-induced intracellular acidosis has been shown to retard oxidative phosphorylation. Nuclear magnetic resonance studies during hypercapnia demonstrated that there was a linear correlation between inorganic phosphate (Pi) /nucleotide triphosphates (NTP) and brain pHi (53). During hypercapnia (Paco2 98 ± 3 mm Hg) in newborn lambs, there was a significant increase in the ratio of Pi to NTP and a significant decrease in the ratio of PCr to Pi. We have also demonstrated the impact of hypercapnia on cellular metabolism with the hypercapnia-induced decrease in high-energy phosphates. We anticipate that the decrease in tissue levels of ATP and PCr in our newborn piglets was due to the increased metabolic demand on the neurons to maintain ionic homeostasis during acidosis.

In tissue slices (54) and cultured astrocytes (55), a reduction in pHe leads to brain edema and cell necrosis (56). During hypercapnia, when both pHi and pHe are lowered, restitution of normal pHi may not be feasible resulting in neuronal injury and possibly cell death through activation of nuclear transcription mechanisms. It has been suggested that the hypercapnia-induced increase in intracellular [H+] may compete for binding sites with intracellular Ca++, resulting in an increase in free cytosolic Ca++ (57). We propose that a hypercapnia-induced increase in intracytosolic Ca++ results in an increase in intranuclear Ca++-influx as membrane enzymes are altered by acidosis. The increase in intranuclear Ca++ then may activate CaMK IV, phosphorylation of CREB, and the transcription of apoptotic genes as demonstrated in our current study.

The two groups of piglets, the CO2 65 and CO2 80 groups, had a similar decrease in high-energy phosphates, increased CaMK IV activity, increased phosphorylation of CREB protein, and increased expression of Bax protein. However, even though these groups had statistically different levels of Paco2, the pH range for each group was not statistically different. We anticipate that the effects of hypercapnia on the brain are due to decreases in pH, and not due to differences in Paco2.

CONCLUSION

We conclude that 6 h of either a Paco2 of 65 mm Hg or 80 mm Hg alters nuclear enzyme activity and protein expression in the cerebral cortex of newborn piglets. Specifically, hypercapnia results in a decrease in cerebral energy metabolism, an increase in CaMK IV activity, phosphorylation of CREB protein, and the expression of apoptotic proteins. We propose that the acidosis induced by hypercapnia may be deleterious to the newborn brain and alters nuclear membrane enzymes leading to an increase in intranuclear Ca++ that results in an increase in CaMK IV activity, the subsequent expression of apoptotic genes, and cell death in the cerebral cortex of the newborn piglet.