Main

The role of NO has been subject of numerous investigations during recent years. This applies to its function as a gaseous messenger molecule and to its vasodilative properties. NO is released by hydroxylation of nitrogen at the guanidino group of L-arginine, a reaction which is catalyzed by the enzyme NO synthase (NOS). At least three isoforms of NOS are known at the moment, two constitutive forms (endothelial NOS and neuronal NOS) and the inducible form (iNOS). The latter is induced by bacteria and cytokines. NO concentrations produced by iNOS are much higher than those generated by constitutive NOS (1), and have been shown to exert antimicrobial effects against some bacterial pathogens in vitro(2).

As soon as NO is generated in the target cell (cancer cell, parasite), binding at copper and iron containing proteins takes place (3). As a result copper and iron are released and, in combination with oxygen, toxic radicals are formed. This is referred to as “massive oxidative injury” and requires the expression of iNOS leading to the production of huge amounts of NO, whereas low levels of NO synthesized by the constitutive forms of NOS (eNOS, nNOS) act as a signal molecule (4). Thus, the concentration of NO determines its biologic effect upon neighboring cells and tissues. The reaction products of nitrogen and oxygen include the strong oxidant peroxynitrite, which is the result of the interaction of NO and superoxide anion (4).

The enzyme iNOS is predominantly expressed in macrophages in response to inflammatory stimuli. Induction occurs by bacteria, bacterial products and cytokines (5). At least in lung tissue NO can exert both pro-oxidant and antioxidant effects (6). Again the ultimate response of the lung most likely depends on the dose of NO administered and the presence of other reactive intermediates (7). At low levels of NO production, antioxidant properties of NO predominate. However, with high or sustained generation of NO due to up-regulation of iNOS or administration of exogenous NO, cytotoxic reactive nitrogen intermediates can be formed (8). In the brain, iNOS expression has been characterized in astrocytes, microglial cells, and to a lesser extent in endothelial cells (9–12). In numerous disease states, beneficial as well as deleterious effects have been attributed to the production of NO by glial iNOS.

Hyperoxia in combination with increased production of NO has been shown to induce cell death and DNA fragmentation in pulmonary tissue (13). This latter effect has been attributed to the formation of peroxynitrite, which results from the reaction of the superoxide anion with NO (14, 15). Peroxynitrite, neither NO nor superoxide, can nitrite tyrosine residues of proteins and form the stable product 3-nitrotyrosine (16). This reaction occurs spontaneously but is also catalyzed by superoxide dismutase or carbon dioxide (17–19).

The effects hyperoxia on the immature retina (20) and lungs (21–23) have been extensively investigated. However, data on the effect of hyperoxia on the immature brain, although scarce, indicate that it may lead to damage in the premature brain (24). Having in mind the steep postnatal increase in oxygen tension (intrauterine pO2: 27–35 mm Hg, postnatal pO2: 40–100 mm Hg), we set out to investigate effects of hyperoxia on iNOS expression in immature brain tissue.

We hypothesized that hyperoxia induces iNOS expression in the immature rat brain and sought to determine the cellular source of iNOS. Furthermore we intended to determine whether up-regulation of iNOS leads to increases in 3-nitrotyrosine, which would point toward peroxynitrite-mediated cellular damage.

METHODS

Animals and oxygen exposure.

Seven-day-old Wistar rat pups were housed with their dames. Altogether 138 rat pups derived from 11 litters were investigated, the number of animals in each experiment varied depending on the size of the individual litter (8 to 24). Oxygen exposure was achieved by placing the animals in an oxygen chamber with a minimal FiO2 of 0.8 for a duration of 24 h (≥ 80% oxygen, <1% carbon dioxide, rest nitrogen). Control animals were kept at room air. There were no hyperoxia-associated symptoms in the rat pups, mortality rate was <5% in both groups. Following oxygen exposure the pups were anesthetized and subsequently killed. Brains were perfused transcardially with heparinized PBS followed by paraformaldehyde (4%) as described previously (25). Briefly, following thoracotomy and right atrial incision the perfusion needle was advanced into the ascending aorta. After PBS and paraformaldehyde perfusion (20 mL each) continuous perfusion with cold paraformaldehyde (4%) was performed for at least 10 min. Brains were stored in paraformaldehyde for a total of 3–7 d, thereafter moved to acidified PBS and embedded in paraffin.

Fresh brain tissue for RT-PCR was stored at minus 70°C following decapitation.

All experiments were performed in accordance with the German Animal Welfare Act and were approved by the local Animal Protection Authority.

RT-PCR.

Whole brains were quickly removed following decapitation, cut into six anatomically defined areas (striatum, hippocampus, frontal cortex, cerebellum, thalamus, retrosplenial cortex), and frozen at minus 70°C until analysis. RNA was isolated by the use of RNeasy® (Qiagen, Bothell, Washington, U.S.A.), a concentration of 0.1 μg/μL was aimed at in all specimens. Briefly, approximately 80 mg of rat brain tissue were homogenized for 30 s and added to 1 mL RLT-buffer and 10 μL β-mercaptoethanol. Lysates were centrifuged at 5000 rpm at 4°C for 3 min, for analysis 500 μL of the supernatant were diluted with an equal amount of ethanol 70%. 500 μL of the lysate-ethanol-mixture were put on top of the spin column (RNeasy® mini) and subsequently centrifuged at 10,000 rpm at 4°C for 15 s. The percolation was discarded, then the whole procedure was repeated with the remainder of the lysate-ethanol-mixture. Again the percolation was discarded. RW1 buffer (700 μL) was added to the column, which was spun down at 10,000 rpm for 15 s. A new 2 mL container was connected to the RNeasy® mini spin column, filled with 500 μL RPE buffer, and then centrifuged at 10,000 rpm for 15 s. After discarding the percolation, the whole procedure was repeated with equal amounts of RPE buffer, at this step the spin down was performed for 4 min at 13,000 rpm. For eluation of the RNA from the column another 1.5 mL container was used. Fifty microliters of RNAse-free water was pipetted onto the membrane and thereafter centrifuged at 13,000 rpm for 5 min. This procedure was repeated once. The resulting volume of extracted RNA was approximately 100 μL. The concentration of RNA was measured photometrically (E260/E280), if required the concentration was modified to achieve 0.1 μg/μL.

Reverse transcription of 1 μg RNA into cDNA was performed by use of random hexamers and AMV reverse transcriptase and 1× PCR buffer (20 mM Tris · HCl, pH 8.3, 50 mM KCl, 2 mM MgCl, and 100 μg/mL BSA). The cDNA was amplified by PCR in a total volume of 30 μL with the use of 5 U/μL Taq DNA polymerase (GIBCO, Eggenstein, Germany). Furthermore 4 μL 1mM dNTP, 0.75 μL MgCl (20 mM), 5 μL 10xPCR-Puffer, 15.6 μL Aqua and 1 μL of each primer were added.

Amplification was achieved in a microprocessor-driven thermocycler (Multicycler PTC-200 MJ Research, Biozym, Hessisch Oldendorf, Germany).

Primers of the housekeeping gene (β-actin) had the following sequences (PCR-product: 897 bp): 5′-CCC TAA ggC CAA CCg TgA AAA gAT g-3′ (sense; 1663–1687nt) and gAA CCg CTC ATT gCC gAT AgT gAT g-3′ (antisense; 2535–2559nt). Amplification of β-actin was terminated after 30 cycles (30″ at 94°C, 60″ at 60°C and 60″ at 72°C). Sequences of iNOS-primers were (PCR-product: 388 bp; MWG-Biotech, Ebersberg, Germany): 5′-AgC ATC ACC CCT gTg TTC CAC CC-3′ (sense; 1592–1613nt) and 5′-Tgg ggC AgT CTC CAT TgC CA-3′ (antisense; 1979–1960nt). Again 30 cycles were run for the iNOS-PCR (30″ at 94°C, 60″ at 60°C and 60″ at 72°C). PCR-products were separated by gel electrophoresis (agarose 2%), stained with ethidium bromide and subsequently exposed to UV light.

iNOS-Western blot.

For Western blotting analysis of brain tissue, animals were killed, and brains were removed, micro-dissected, and immediately snap frozen in liquid nitrogen.

Tissue was then homogenized at 4°C in a Tris-HCL buffer (50 mM, pH 7.6) containing: 150 mM KCL, 5 mM ethylene diamine tetraacetic acid (EDTA), 1 mM phenylmethane sulfonyl fluoride (PMSF), 1% Triton X-100, 0.5 μg/mL leupeptin and 1 μg/mL pepstatin. Homogenate was centrifuged at 15,000 g for 20 min and the supernatant used as the cytosolic fraction.

Total cellular proteins (25 μg/lane cytosolic fraction) were separated on a 10% SDS-polyacrylamide gel and electrotransferred onto nitrocellulose membranes (Hybond ECL, Amersham International, Bucks, U.K.). Equal loading and transfer of proteins was confirmed by temporally staining the membranes with Ponceau S solution. The nonspecific binding of proteins was prevented by treating the membranes with 5% nonfat dry milk in Tris-buffered saline/0.1% Tween 20 for 2 h at room temperature. The membranes were incubated overnight at 4°C with anti-iNOS antibody (1:10,000; Transduction Laboratories). After incubation with secondary antibody conjugated to horseradish peroxidase (anti-mouse, 1:2000 dilution), immunoreactive proteins were detected by the enhanced chemiluminescence system (ECL, Amersham International) and serial exposures were made to radiographic film (Hyperfilm ECL, Amersham International). Densitometric analysis of the blots was performed with the image analysis program TINA 2.09 g.

Immunocytochemistry.

Paraffin-embedded serial coronal sections were cut at a thickness of 10 μm (Leica microtome VT 1000S, Leica, Nussloch, Germany) and mounted on silane-coated slides. The anti-iNOS-antibody was purchased from Santa Cruz (Cat.-No. Sc-7271, Santa Cruz, CA, U.S.A.), the anti-nitrotyrosine antibody was acquired from Upstate Biotechnology (Cat.-No. 06–284, Upstate, Lake Placid, NY, U.S.A.). Antibodies were used in concentrations of 1:100 (iNOS) and 1:200 (nitrotyrosine), respectively. In brief, sections were initially deparaffinized by washing in xylene and decreasing ethanol concentrations, then boiled in citrate buffer to unmask antigen sites. Following cooling with tap water, slides were washed in PBS. Endogenous peroxidase activity was blocked by exposing slides to 0.6% hydrogen peroxide in PBS for 30 min. After washing in PBS nonspecific binding was avoided by incubation with 5% blocking solution (5% normal goat serum in PBS) for 20 min. Sections were incubated overnight with the respective primary antibody. Following removal of the antibody and repetitive rinsing with PBS, slides were incubated with a biotinylated goat anti-rabbit IgG antibody (1:200, Vector Laboratories, Burlingame, CA, U.S.A.). Immunocytochemical identification of positive cells was performed by the use of an avidin-biotinylated peroxidase complex (ABC-kit Vectastain, Vector Laboratories, Burlingame, CA, U.S.A.) and DAB substrate (Vector Laboratories, Burlingame, CA, U.S.A.). After intensive washing in PBS and Tris buffer slides were counterstained with methyl green 10% or hematoxylin, respectively. Sections were washed with butanol, treated with xylene and subsequently mounted. All specimens were examined by light microscopy (CH2, Olympus, Shinjuku-ku, Tokyo, Japan), photographs were taken with a digital camera (C-3030, Olympus, Shinjuku-ku, Tokyo, Japan).

Light microscopy.

Light microscopy was performed with a conventional microscope (CH2 or BX60, Olympus, Shinjuku-ku, Tokyo, Japan). Brain regions examined were:

In anterior brain areas primary motor cortex (MOP), secondary motor cortex (MOS), and medial septal nucleus (MSN) [Fig. 1, left]

Figure 1
figure 1

Anterior and posterior rat brain slices of the areas investigated (MOP: primary motor cortex; MOS: secondary motor cortex; RSC: retrosplenial cortex; HIP: hippocampus with areas CA1 and CA3; THA: thalamus; MSN: medial septal nucleus).

Full size image

In posterior brain areas retrosplenial cortex (RSC), hippocampus (HIP) including regions CA1 and CA3, and thalamus (THA) [Fig. 1, right].

Statistical analysis.

Data are shown as mean and SEM. Statistical evaluation was performed using the Wilcoxon test, significant differences between control and hyperoxia animals were assumed at p < 0.05.

RESULTS

iNOS-immunohistochemistry.

Whereas there was only occasional staining in perivascular areas in controls, iNOS-positive staining was present in all hyperoxia-exposed animals. Staining was confined to microglial cells as evident from the typical microscopic shape and appearance. Positive cells were predominantly present in the retrosplenial cortex (Fig. 2), hippocampus (Fig. 3), and frontal cortex (Fig. 4), less so in the thalamus (Fig. 5).

Figure 2
figure 2

iNOS immunohistochemistry in the hippocampus (1000×) of the 7-day-old control rat without evidence of specific staining.

Full size image
Figure 3
figure 3

iNOS immunohistochemistry in the hippocampus (1000×) of the 7-day-old hyperoxia-exposed rat with brown staining of microglial cells (black arrows).

Full size image
Figure 4
figure 4

Frontal cortex (iNOS staining; 1000×) in the 7-day-old hyperoxia-exposed rat; note reticular pattern of staining confined to microglial cells (black arrows).

Full size image
Figure 5
figure 5

Thalamus (iNOS staining; 1000×) in the 7-day-old hyperoxia-exposed rat; microglial cells stain positive as visible by the brown, net-like appearance (black arrows).

Full size image

Nitrotyrosine-immunohistochemistry.

Staining was strongly positive in some neurons in the frontal cortex, other neurons remained unstained. Particularly apical dendrites showed marked staining (Fig. 6), this was confined to neurons within the frontal cortex.

Figure 6
figure 6

Nitrotyrosine staining (frontal cortex, 1000×) in the hyperoxia treated 7-day-old rat pup; staining is most pronounced in neurons and their apical dendrites (black arrows).

Full size image

iNOS-RT-PCR.

Expression of iNOS was only slightly increased 2 h after the onset of hyperoxia, significantly so at 6 h and 12 h (Fig. 7). iNOS-mRNA had a similar time course in frontal cortex and thalamus, after 24 h of oxygen exposure iNOS expression was almost back to initial levels.

Figure 7
figure 7

Time course of iNOS-mRNA expression (RT-PCR) during the first 24 h of hyperoxia exposure in frontal cortex (square and original blot) and thalamus (triangle) (n = 6).

Full size image

iNOS-Western blot.

Expression of iNOS protein was significantly increased in retrosplenial cortex and hippocampus after 24 h of hyperoxia (t test), iNOS expression in control animals was set as 100% (Fig. 8).

Figure 8
figure 8

iNOS protein expression (Western blot) of controls (n = 7) and 24 h hyperoxia exposure (n = 7) for the brain regions retrosplenial cortex (RSC) and hippocampus (HIP).

Full size image

DISCUSSION

Here we demonstrate up-regulation of iNOS and subsequent peroxynitrite formation in different areas and cell types in the immature rat brain previously exposed to hyperoxia. Whereas iNOS protein staining was confined to microglial and to a lesser extent to endothelial cells, peroxynitrite formation took place in apical dendrites of neurons. This differential observation is supported by in vitro co-cultures of activated microglia and embryonic cortical neurons, in which microglia was shown to produce nitric oxide and adjacent neurons exhibited neurodegenerative damage (26).

Exposure to high oxygen concentrations has long been known to cause pulmonary morbidity (21, 27, 28). It has been shown that hyperoxia serves as a potent trigger of up-regulation of iNOS mRNA expression in pulmonary cells within several hours of oxygen exposure in mice (29) as well as in rats (30). When wild-type animals were compared with iNOS-deficient mice, lung injury was more severe in iNOS-deficient animals indicating a protective effect of iNOS due to reduction of pulmonary leukocyte accumulation. Since nitrotyrosine was detectable in both wild type and iNOS deficient mice following hyperoxia, existence of an iNOS-independent pathway for protein nitration was concluded (29).

Once the effects of hyperoxia on distant organs are investigated (for example the brain), matters become more complicated. Although there is evidence that hyperoxemia leads to reduced cerebral blood flow in several animal models, information about resulting tissue oxygen tensions remains scarce (31). In a rat model of 2-day-old pups the combined effect of hyperoxia and inhibition of glutathione synthesis led to an increased amount of histologically detectable apoptosis (32). Other distant organs have been investigated with respect to the integrity of their anti-oxidant capacities. In the isolated perfused rat liver hyperoxia not only led to up-regulation of iNOS protein and mRNA levels, but also to impaired anti-oxidant defense mechanisms as indicated by reduced intracellular glutathione levels (33).

Focal ischemia and reperfusion after 1 h in a newborn rat model increased immunohistochemical iNOS and nitrotyrosine staining in neutrophils and perivascular endothelial cells at the age of 72 h (34). This shows that nonresident cells (neutrophils) as well as resident brain cells (perivascular endothelial cells) contribute to the intracerebral formation of peroxynitrite. Unfortunately from the experimental set-up of the model, no differentiation between the underlying triggers for nitration of tyrosine residues can be made (ischemia versus reperfusion i.e. relative hyperoxia).

In a different model of hypoxia-ischemia the authors found increased amounts of iNOS and nitrotyrosine in the infarct area. They were also able to reduce iNOS-mediated damage by the use of S-methyl-isothiourea (a selective inhibitor of iNOS) (35). Although the causative effect of iNOS up-regulation on the resulting brain damage could be demonstrated by specific inhibition, caution is warranted as to what caused the increase in iNOS-mRNA expression in the first place (hypoxia-ischemia or subsequent reoxygenation, in this case with room air).

Both of these aforementioned rat models investigated effects of hypoxia-ischemia with subsequent reoxygenation in the immature rat brain. The maturity of cerebral structures of the 7-day-old rat corresponds approximately to that of the human fetus during the last trimester of pregnancy. This approximation is derived from the temporal assignment of the brain growth spurt in man and rat, which equates the period of the growth spurt with the maximal vulnerability of cerebral integrity (36). In our experimental model we used animals of the same postnatal age, but modified our intervention to a single procedure i.e. hyperoxia. To our knowledge this is the first investigation of the hyperoxia effect on iNOS expression in the immature rat brain. Earlier investigators looked at peroxynitrite formation in cortical glial cultures following iNOS induction (37), iNOS expression following intrastriatal injection of lipopolysaccharide (38, 39), or iNOS expression following traumatic brain injury in immature (40) and adult rats (41–43).

Authors of a porcine model of inflammation concluded that their finding of only minimal induction of iNOS but intense nitrotyrosine immunostaining indicates the existence of alternative pathways of nitrotyrosine production (44).

Since iNOS is not the sole enzyme catalyzing the synthesis of NO, particularly up-regulation of eNOS could lead to the formation of peroxynitrite without concomitant increases in iNOS protein. Activation of this pathway in the context of inflammation has been previously shown (45). Alternatively, nitrotyrosine could be produced by the reaction of tyrosine residues with other reactive molecules, such as perchloric acid, which is produced by infiltrating neutrophils and hydroxyl radicals from myeloperoxidase (46).

In vitro experiments have provided evidence of sufficient protection against nitration of tyrosine residues in the presence of physiologic levels of glutathione and carbon dioxide (47). Whether this is also true during in vivo conditions – particularly in immature animals – needs to be determined.

We conclude that hyperoxia induces up-regulation of iNOS-mRNA and synthesis of iNOS-protein in microglial cells in several areas of the immature rat brain. This increased protein synthesis leads to the formation of peroxynitrite, which indicates potential damage to structures of the neonatal brain. Further research should address underlying mechanisms to identify endogenous and exogenous systems capable to protect the neonatal brain.