Abstract
Abstract. Interaction of reactive oxygen species with nitric oxide promotes nitric oxide inactivation and generation of cytotoxic reactive nitrogen species that attack DNA, lipids, and proteins. Nitration of free tyrosine and tyrosine residues of proteins results in production of nitrotyrosine, which can lead to excitotoxicity and frequently is found in the brain of patients and animals with various degenerative, ischemic, toxic, and other neurologic disorders. According to earlier studies, reactive oxygen species activity is increased and neuronal NO synthase expression in the brain is elevated in animals with chronic renal failure (CRF). It was hypothesized, therefore, that tyrosine nitration must be increased in the uremic brain. This hypothesis was tested, through determination of nitrotyrosine abundance (by Western blot analysis), as well as distribution (by immunohistology), in the cerebrum of rats with CRF 6 wk after 5/6 nephrectomy. The results were compared with those of sham-operated controls and antioxidant (lazaroid)-treated and captopril-treated rats with CRF. Western blot analysis revealed a significant increase in nitrotyrosine abundance in the cerebral cortex of rats with CRF. This was accompanied by an intense nitrotyrosine staining of the neuronal processes, including proximal segments of dendrites, axons, and axon terminals of the cortical neurons. Both antioxidant therapy and captopril administration alleviated oxidative stress (as evidenced by normalization of plasma lipid peroxidation product malondialdehyde) and significantly reduced nitrotyrosine abundance in the cerebral cortex of the treated CRF group. In conclusion, CRF resulted in oxidative stress and increased tyrosine nitration in the cerebral cortex. Antioxidant therapy and angiotensin-converting enzyme inhibition alleviated the CRF-induced oxidative stress and mitigated tyrosine nitration in the rats with CRF.
Advanced renal failure frequently is associated with the brain dysfunction generally referred to as uremic encephalopathy. Uremic encephalopathy presents with a wide range of electrophysiologic and clinical manifestations that include slowing of the α rhythm on electroencephalography, impaired somatosensory evoked potential, abnormal cognitive event-related potential, fatigue, apathy, emotional instability, sleep disorder, and neuroendocrine dysregulation. Other manifestations of uremic encephalopathy include forgetfulness; perceptual error; agitation; impaired cognition; and, in severe cases, delirium, delusions, hallucinations, coma, and convulsions, which point to the malfunction of the advanced brain areas. Furthermore, specific frontal release phenomena such as snout, grasp, and palmomental reflexes are found in some cases (1,2,3,4,5,6).
Nitric oxide (NO), originally identified as the endothelium-derived relaxing factor, now is known to be a critical intra- and intercellular signal molecule that plays a fundamental role in regulation of a wide variety of biologic functions (7). In addition to its important physiologic functions, NO is involved in various pathologic processes that lead to cytotoxicity (8,9). In this regard, interaction of NO with reactive oxygen species (ROS) leads to generation of highly reactive and cytotoxic byproducts, such as peroxynitrite, which can attack DNA, lipids, and proteins (10,11). For instance, peroxynitrite reacts with free tyrosine and tyrosine residues in protein molecules to produce nitrotyrosine. Alternatively, ROS can activate tyrosine to form tyrosyl, a radical that, in turn, oxidizes NO to produce nitrotyrosine (11,12).
Oxidative stress and heightened NO-ROS interaction have been implicated in the pathogenesis of a variety of degenerative, ischemic, inflammatory, and toxic disorders of the central nervous system (13). The reactive byproducts of NO-ROS interaction can promote direct cytotoxicity by attacking cellular constituents. In addition, nitrotyrosine, the end product of ROS-NO-tyrosine interaction, has been shown to cause excitotoxicity via activation of N-methyl-D-aspartate receptors (14). Accumulation of nitrotyrosine is reported in the brain of transgenic mouse model and in the nonhuman primate model of Huntington's disease (15,16,17). Similarly, nitrotyrosine accumulation occurs in the aged brain (18) and in the brain of patients with Alzheimer's disease (19,20), multiple sclerosis (21), amyotrophic lateral sclerosis (22), carbon monoxide toxicity (23), stroke (24,25), and head trauma (26). Administration of neuronal NO synthase (NOS) inhibitors has been shown to attenuate striatal lesions (17,24,27) and mitigate the rise in hydroxyl radical production and nitrotyrosine generation in the nonhuman primate model of Huntington's disease (16,23,28). Similarly, antioxidant therapy has been shown to retard the progression of Alzheimer's disease (29). Thus, attenuation of ROS activity and NO production may favorably influence the course of certain central nervous system disorders.
The underlying mechanisms that are responsible for brain dysfunction in uremia have not been elucidated fully. The accumulation of uremic toxins, excess parathyroid hormone, impaired cellular Na-K-ATPase activity, increased cellular calcium content, erythropoietin-deficiency anemia, and impaired melatonin metabolism have been implicated in this process (1,3,6,30,31,32). Earlier studies by Vaziri et al. (33) demonstrated strong evidence that increased ROS activity leads to enhanced NO inactivation in the uremic rats. They showed further that enhanced NO inactivation by ROS plays a role in the pathogenesis of uremic hypertension (33). Furthermore, neuronal NOS (nNOS) expression is elevated in the brain of uremic rats (34). In view of the compelling evidence for increased ROS activity and disturbed brain NO metabolism in uremia, we hypothesized that the concomitant increase in ROS activity with elevated brain tissue nNOS expression in uremia may favor generation and accumulation of nitrotyrosine in the uremic brain. We further considered that if this hypothesis is true, then antioxidant therapy should ameliorate this process. The present study was undertaken to test this hypothesis and to discern the localization of nitrotyrosine in the brain, particularly the cerebral cortex, in a rat model of chronic renal failure (CRF).
The antioxidant chosen for this study was des-methyl-tirilazad, which is a powerful scavenger of ROS and a potent inhibitor of lipid peroxidation (35). This compound and its closely related derivatives have been used widely to study the role of oxidative stress in a wide range of neurologic disorders (35). In addition, we used this agent in our earlier studies of the role of oxidative stress in the pathogenesis of uremic and lead-induced hypertension (33,36).
Materials and Methods
Animals
Male Sprague-Dawley rats with an average body weight of 250 g were housed in a climate-controlled, light-regulated space with 12-h light and dark cycles. They were fed a basic diet that contained 24% protein (Purina Mills, Brentwood, MO) and water ad libitum. The rats were assigned randomly to three groups of six rats each. The first group was subjected to 5/6 nephrectomy to produce CRF. With the rats under general anesthesia (pentobarbital, 50 mg/kg intraperitoneally), the upper and lower thirds of the left kidney were excised by use of a dorsal incision. This was followed by the removal of the right kidney 4 d later. After a 4-wk recovery period, the rats with CRF were divided into placebo-treated (CRF) and lazaroid-treated (CRF-LZ) groups. The CRF-LZ group received the potent antioxidant and lipid peroxidation inhibitor des-methyl-tirilazad (U-74389G; UpJohn, Inc., Kalamazoo, MI) 10 mg/kg twice daily by gastric gavage, for 2 wk. The CRF and the sham-operated control groups received the vehicle. In an attempt to discern the possible contribution of CRF-associated hypertension, we treated a separate group of CRF rats with the angiotensin-converting enzyme (ACE) inhibitor captopril (50 mg/kg per d in the drinking water [CRF-cap group]). The choice of an ACE inhibitor was based on the widely accepted indication of this class of antihypertensive agent in CRF.
At the end of the study period, rats were placed in metabolic cages for a timed urine collection. Arterial BP was monitored by tail plethysmography, as described in our earlier study (37). The rats then were anesthetized with pentobarbital (50 mg/kg intraperitoneally), blood was obtained by cardiac puncture, and the brain was removed immediately after decapitation and frozen by immersion in isopentane over dry ice, then stored at -70°C until processed for Western blotting and immunohistochemical analysis. Plasma and urine creatinine concentrations were measured by use of standard laboratory methods.
Measurement of Malondialdehyde
Plasma malondialdehyde (MDA) was determined as an indicator of ROS-mediated lipid peroxidation by means of HPLC. The procedure was carried out in a manner identical to that previously described by Vaziri et al. (33).
Western Blot Analysis
The frozen brains were defrosted, and samples of the neocortex were isolated and processed as described earlier (19). Briefly, 100 mg of protein from the cortex was separated by electrophoresis in 4 to 12% Tris-Glycine gel, and nitrotyrosine was detected after a 24-h incubation with monoclonal mouse antiserum directed against 3-nitrotyrosine (Upstate Biotechnology, Inc., Lake Placid, NY) at a final concentration of 2 μg/ml, followed by a 2-h incubation with goat anti-mouse IgG conjugated with peroxidase, diluted at 1:3000. The bands were visualized by use of enhanced chemiluminescence detection reagents (Amersham Life Science, Inc., Arlington Heights, IL) and quantified by a laser densitometer.
Immunohistochemistry
The brains used for immunocytochemistry were fixed in 4% paraformaldehyde in 0.1 M Sorensen buffer (pH 7.3) for 48 h and stored in 0.1 M phosphate-buffered saline that contained 20% sucrose. Tissue blocks that contained the entire cerebrum were cut at the coronal plane at 50 μm and collected in Tris-HCI-buffer (TB, 0.1 M [pH 7.4]) in cell culture wells and stained for nitrotyrosine by use of a modification of the avidin-biotin-peroxidase complex method, as described in detail elsewhere (19,38). Briefly, free-floating cerebral sections from the study groups were pretreated with 50% formic acid for 5 min and incubated in TB with the primary antibody (diluted to 2 μg/ml) overnight at room temperature. After several washes with TB, the sections were reacted with biotin-conjugated goat anti-mouse IgG (1:200) at room temperature for 2 h. The bound antibodies were detected by use of the ABC kit, and the immunoreactive product was visualized with a diaminobenzidine substrate kit (Vector Laboratories, Burlingame, CA). Control experiments for the immunostaining in Western blot and immunohistochemistry included the above-described incubation procedures with the primary antibody omitted or preabsorbed with the 3-nitrotyrosine antigen. No specific staining was found in the control preparations. To minimize experimental bias, we processed sections from different groups under identical conditions. Thus, sections from the study groups were incubated with the same prepared antibody solutions. In other cases, sections from different groups first were marked by removal of a part of an unrelated portion (for later orientation) and then reacted in the same container throughout the entire staining procedure.
Statistical Analyses
ANOVA and multiple range test were used in evaluation of the data from blood and urine measurements and Western blot analyses, which are presented as mean ± SEM. P < 0.05 was considered to be significant.
Results
Physiologic Data
Data are given in Table 1. Plasma creatinine concentration in the CRF group was comparable with that of the CRF-LZ and CRF-cap groups and was significantly higher than that found in the control group. Creatinine clearance in the CRF, CRF-LZ, and CRF-cap groups was significantly lower than that observed in the control group. Likewise, final body weights in the CRF, CRF-LZ, and CRF-cap groups were significantly lower than that of the control group. Plasma concentration of lipid peroxidation product, MDA, was significantly higher in the CRF group than that found in the control rats. Antioxidant therapy with the lazaroid compound as well as captopril administration restored plasma MDA concentration to normal levels in the rats with CRF. Systolic BP in the CRF group was significantly higher than that found in the control group. Lazaroid therapy significantly ameliorated but did not normalize the CRF-induced hypertension in the CRF-LZ group. However, captopril administration lowered BP to near normal values. Food intake was monitored during the final week of the study period and was found to be slightly lower in the CRF groups (17.6 ± 4 g/24 h) than in the sham-operated control group (18.8 ± 4 g/24 h). However, the difference did not reach statistical significance.
Plasma creatinine concentration, Ccr, body weight, systolic BP, and plasma MDA in the sham-operated control group and CRF, CRF-LZ, and CRF-cap groupsa
Cellular Localization of Nitrotyrosine in the Cerebrum
In general, specific nitrotyrosine immunostaining was found in all cortical layers in the frontal neocortical area in all groups (Figure 1). Although nitrotyrosine immunoreactivity was associated with cell bodies, fibers, and puncta structures, the most abundant and distinctly stained profiles were processes that presumably originated from the cortical neurons. These fibrous structures were most dense in layers IV and VI and were less dense in layers III and V but were sparse in layer I (Figure 1, A through C). The nitrotyrosine-positive fibers were abundant in the white matter. At high magnification, the majority of these fibers were thin and oriented in different directions, forming a complicated plexus in the neuropil in the cortex and underlying white matter. Those in layers IV and VI were more frequently seen to run either vertically or tangentially relative to the pial surface (Figure 1, D through F). These fine processes seemed to be characteristic of axons and axon terminals of the neocortical neurons. The so-called chandelier structure formed by axon terminals that synapse on the initial segment of pyramidal cell axons (39,40) also was distinctly labeled for nitrotyrosine (Figure 1, D through F). These labeled chandelier formations appeared as strong immunoreactive puncta packed in paired lines in layers IV and VI, underneath the somata of cortical pyramidal cells that were immunonegative (Figure 1D).
Distribution of immunolabeling for nitrotyrosine in the neocortex of rats with chronic renal failure (CRF) and antioxidant-treated rats (A through G). (A, B, and C) Low-power views of the immunoreactivity in comparative frontal neocortical regions from normal, uremic, and antioxidant-treated uremic rats, respectively. (D, E, and F) Images enlarged from areas of A, B, and C, respectively. Nitrotyrosine immunoreactivity is localized mainly to processes (D through F). A number of somata are lightly to moderately labeled (G). Most of the labeled cell bodies are multipolar (arrows in G). The proximal dendritic processes of these cells are labeled, and they arise from the upper (open arrow) and basal (solid arrows) poles of the soma. The immunoreactivity is located in the peripheral zone of the cytoplasm (* in G). These somata seemed to be cortical pyramidal cells with labeled apical and basal dendrites. In the neuropil, numerous fine processes were immunoreactive (D through F). Some of them formed vertically oriented “cartridges” (arrows in E) that represented axon terminals synapsing on the initial segment of the pyramidal cell axons. Other labeled fibers run parallel to the cortical surface (arrowheads in E). Note that a heavier immunolabeling in the neuronal processes was present in the cortex of the uremic rat, compared with that in the normal rat. More important, this enhanced expression of nitrotyrosine immunoreactivity in the uremic rat was reduced after antioxidant treatment (compare D, E, and F). Cortical layers are indicated through layers I through VI. Scale bar = 500 μm for A and C, 100 μm for D through F, and 25 μm for G.
Cell bodies of the cortical neurons were lightly stained and found across the depth of the cortex except for layer I (Figure 1A). These labeled somata exhibited a considerable variation in shape and size. Most of the somata were round and medium in size, with a subpopulation being large and triangular or multipolar in shape (Figure 1G). The immunoreaction product in the cell body was localized to the periphery of cytoplasm or subjacent to the plasmalemma (Figure 1G). The central region, especially the nucleus, was almost always devoid of any immunolabeling. Immunolabeling to the dendrites of the cortical neurons was limited to their proximal portion of <100 μm away from the somata. In a few cases, these labeled primary dendrites could be identified as apical and basal dendrites of the pyramidal cells according to their origin and orientation as well as the morphology of their parent somata (Figure 1G).
Effects of CRF and Antioxidant Treatment
The pattern of cellular localization and laminar distribution of immunoreactive somata and processes demonstrated little difference among the normal control, CRF, and CRF-LZ rats (Figure 1, A through F). However, the abundance of these labeled profiles, particularly that of the neuronal processes, was markedly increased in the CRF rats compared with the sham-operated control rats. Thus, the immunoreactive processes in the cortex of the CRF rats were much more numerous and more intensely labeled (Figure 1, D and E), resulting in a densely packed immunoreactive plexus that almost filled the entire neuropil (Figure 1E). The intensity of labeling of the neuronal processes was significantly less in the antioxidant-treated CRF rats, compared with the CRF rats (Figure 1F). In contrast to the neuronal processes, the immunoreactivity in neuronal somata was comparable in the three groups (Figure 1, A through C). However, the labeled neuronal somata in the rats with CRF often were obscured by the intense immunolabeling of the surrounding processes.
Western Blot
Data are illustrated in Figures 2 and 3. Multiple protein bands with nitrotyrosine reactivity were detected by Western blot that reflected nitration of several protein constituents in the cerebral cortex of rats, as has been shown previously in human brain (19). The untreated CRF group exhibited a marked increase in nitrotyrosine abundance in the neocortical tissue preparation when compared with the control group. Antioxidant therapy reversed the CRF-associated elevation of nitrotyrosine abundance to a level that approximated that seen in the controls. Likewise, administration of ACE inhibitor captopril resulted in normalization of nitrotyrosine abundance in the brain of the rats with CRF.
Representative Western blots and group data depicting nitrotyrosine abundance in the cerebral cortex of the placebo-treated CRF (CRF), lazaroid-treated CRF (LZ), and sham-operated normal control (NL) groups. n = 6 rats in each group. *, P < 0.01 (ANOVA).
Representative Western blot and group data depicting nitrotyrosine abundance in sham-operated control, placebo-treated CRF (CRF), and captopril-treated CRF (CRF+Cap) rats. n = 6 in each group. *, P < 0.01 (ANOVA)
Discussion
Western blot analysis revealed a marked increase in nitrotyrosine abundance in the cerebral cortex of the rats with CRF. In an attempt to explore the cellular localization of nitrotyrosine accumulation in the cortical tissue, we conducted an immunohistological examination using nitrotyrosine antibody. The immunoreactivity was present in the somata, proximal dendrites, axons, and axon terminals of the cortical neurons. Most of the labeled somata seemed to be cortical interneurons, according to their somal size and configuration (40). This interpretation was supported by the fact that chandelier formation, a characteristic structure formed by the axon terminals of interneurons (39,40), was visualized by the antibody labeling. A few of the labeled somata could be identified as cortical principal neurons, i.e., the pyramidal cells, on the basis of their triangular somata and the orientation of their apical and basal dendrites. The labeled neuronal processes were spread over the entire cortex and the underlying white matter. The distribution of nitrotyrosine generally seemed to be consistent with the previously reported expression of neuronal NOS in the cerebral cortex (40,41,42). For example, a small number of cortical interneurons contain nNOS (43,44). The NOS-containing dendrites and axons of these interneurons are distributed extensively throughout the grey and white matters. In addition, under certain conditions, the pyramidal neurons in the cerebral cortex may express inducible NOS (42,45).
Although the pattern of nitrotyrosine distribution was similar, the intensity of nitrotyrosine abundance was markedly increased in the uremic brains, compared with that found in the control brains. We believe that augmented cerebral NO production capacity shown by Ye et al. (34) coupled with enhanced ROS activity shown by Vaziri et al. (33) necessarily favors the idea that NO-ROS interaction leads to increased nitrotyrosine generation in the rats with CRF. This viewpoint is supported by the observation that amelioration of oxidative stress by antioxidant therapy minimized nitrotyrosine accumulation in the antioxidant-treated CRF group. It is of note that the immunohistological examination of the brain in the CRF group indicated a close association of nitrotyrosine with neuronal processes and plasma membrane of cortical cells. This finding suggests that the free radicals and byproducts of NO-ROS interaction may preferentially attack the lipid-rich cell membranes and neuronal processes.
Administration of the potent antioxidant lazaroid compound mitigated nitrotyrosine accumulation in the brain of the rats with CRF. These observations are consistent with the results of recent studies by Vaziri et al. (46,47) in rats with lead-induced hypertension that exhibited a marked accumulation of nitrotyrosine and a compensatory upregulation of NOS isotypes in various tissues, including brain. These abnormalities were reversed by antioxidant therapy in rats with lead-induced hypertension, a condition that is marked by oxidative stress (46,47). In confirmation of the earlier study by Vaziri et al. (33), the placebo-treated rats with CRF showed a marked increase in plasma lipid peroxidation product, MDA, denoting the presence of oxidative stress. Antioxidant therapy normalized plasma MDA in the CRF-LZ group. The observed accumulation of nitrotyrosine in the brain of rats with CRF and its reversal by antioxidant therapy represent tangible evidence of the biochemical modification of brain by reactive oxygen and nitrogen species in CRF.
The role of enhanced tyrosine nitration in the pathogenesis of uremic brain dysfunction is uncertain. However, compelling evidence recently emerged to link similar events in the course of numerous central nervous system disorders, such as the transgenic mouse and nonhuman primate models of Huntington's disease (16,48), progressive supranuclear palsy (49), Alzheimer's disease (19,20), the aged brain (18), amyotrophic lateral sclerosis (22), carbon monoxide poisoning (23), and ischemic brain injury (24). The administration of specific nNOS inhibitors to reduce NO generation or antioxidant therapy to reduce ROS activity has been beneficial in some of the above disorders (16,27,28,29). The results of the present study in the uremic rats are consistent with those of the above studies of other central nervous system disorders in humans and experimental animals. We believe that increased abundance of reactive oxygen-nitrogen species in CRF constitutes an important component of uremic toxicity that contributes to numerous short- and long-term systemic consequences of uremia. In addition to causing direct cytotoxicity and biochemical modifications, oxidative stress may indirectly affect the brain function and structure by exacerbating anemia (via shortening erythrocyte life span) and hypertension (via inactivation of NO) and promoting arteriosclerosis, to mention a few possibilities. It is noteworthy that oxidative stress causes hypertension (50), and hypertension augments NO production (via shear stress), a combination that is highly conducive to nitrotyrosine generation. By ameliorating oxidative stress, antioxidant therapy reduces BP (33,50) and reverses the compensatory upregulation of NO production (51), thus limiting nitrotyrosine generation. This phenomenon can, in part, account for the reduction of the nitrotyrosine burden in the brain of lazaroid-treated rats with CRF. This viewpoint is supported by the observation that antihypertension therapy resulted in normalization of nitrotyrosine abundance in the captopril-treated rats with CRF used in this study. It is of note that nicotinamide-adenin dinucleotide phosphate oxidase is thought to be a major source of ROS in the vascular tissue (52,53) and that this enzyme is upregulated by both angiotensin II and cyclical stretch (54,55). Thus, blockade of angiotensin II production and amelioration of hypertension (hence, cyclical stretch) by ACE inhibition can attenuate oxidative stress, ROS-NO interaction, and tyrosine nitration, as was seen in the captopril-treated rats with CRF.
In summary, rats with CRF exhibited a significant increase in nitrotyrosine abundance in the cerebral cortex. Antioxidant therapy reversed this abnormality, thus pointing to the role of oxidative stress in the pathogenesis of enhanced tyrosine nitration in the uremic brain. The observed nitrotyrosine accumulation likely is due to a simultaneous upregulation of brain nNOS expression (34) and increased ROS activity in rats with CRF (33). The applicability of these observations to and usefulness of antioxidant therapy in clinical uremia is uncertain and awaits future investigation.
- © 2001 American Society of Nephrology