Impact of Experimental Acute Hyponatremia on Severe Traumatic Brain Injury in Rats: Influences on Injuries, Permeability of Blood–Brain Barrier, Ultrastructural Features, and Aquaporin-4 Expression

Experimental Neurology 178, 194 –206 (2002) doi:10.1006/exnr.2002.8037

Impact of Experimental Acute Hyponatremia on Severe Traumatic Brain Injury in Rats: Influences on Injuries, Permeability of Blood–Brain Barrier, Ultrastructural Features, and Aquaporin-4 Expression 1 Changshu Ke,* Wai Sang Poon,* Ho Keung Ng,† Fernand M. M. Lai,† Nelson L. S. Tang,‡ and Jesse C. S. Pang† *Division of Neurosurgery, Department of Surgery, †Department of Anatomical and Cellular Pathology, and ‡Department of Chemical Pathology, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong Received June 22, 2001; accepted July 29, 2002

The effects of acute hyponatremia on severe traumatic brain injury (TBI) in 35 adult male Sprague– Dawley rats were studied in a replicated focal and diffuse injury rat model. Such effects were assessed by the cerebral contusion volume and axonal injury (AI) densities, determined by quantitative immunoreactivity of ␤-amyloid precursor protein, by blood– brain barrier (BBB) permeability based on endogenous IgG immunostaining, and by ultrastructural features. Significant increase of contusion volume (P < 0.05) and of AI in the segment of corpus callosum beneath the contusion (P < 0.05) and ipsilateral thalamus (P < 0.05) were observed at 4 h postinjury during the hyponatremic phase. No change in BBB permeability was observed in the hyponatremia ⴙ TBI (HT) groups. Significant swelling of perivascular astrocytic foot processes in the HT groups was seen at 4 h (P < 0.01) and 1 day postinjury (P < 0.01) by quantitative image analysis of ultrastructures. However, attenuated swelling of perivascular astrocytic foot processes in severely edematous medulla oblongata with simultaneous swelling of perikaryal astrocytic processes was observed in the HT 1-day group. The ultrastructural features were also correlated with the down-regulation of aquaporin-4 (AQP4) mRNA expression (P < 0.05). Results suggest that acute hyponatremia acts as one of the secondary insults following severe TBI. Such exacerbation may not be attributable to further disruption of BBB permeability, but rather to the ischemia resulting from the swelling of perivascular astrocytic foot processes impeding microcirculation. Down-regulated AQP4 mRNA expression may be one of the molecular mechanisms maintaining water homeostasis in diffusely injured brain exposed to acute hyponatremia. © 2002 Elsevier Science (USA)

1

This project was supported by a direct grant from The Chinese University of Hong Kong (Project Code: 2040579). 0014-4886/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

Key Words: head injury; hyponatremia; aquaporin-4; ultrastructure.

INTRODUCTION

Hyponatremia is a common disturbance in head injury patients, with an incidence ranging from 4.6 to 34%, and has been associated with significant mortality and permanent brain damage (5, 7, 14). Hyponatremia has been suggested to represent a secondary insult in traumatic brain injury (TBI), although morphological evidences of significant brain damage have not been demonstrated in previous studies (40, 41). The effects of hyponatremia on severe TBI remain to be established. Acute hyponatremia with a rate exceeding 0.5 mmol/ L/h can easily induce neurological symptoms. This is related to the extent and rapid fall of serum sodium concentration and carries a high risk of mortality (12, 24). Therefore, we decided to reproduce the experimental acute hyponatremia by application of desmopressin acetate (dDAVP) plus infusion of hypoosmolar solution on a replicated focal and diffuse injury rat model. The injury and brain edema levels were observed to be exacerbated by the superimposed acute hyponatremia (22). Previous studies also showed that water accumulates in the intracellular compartment during the hyponatremic phase in uninjured brain and that damage to the blood– brain barrier (BBB) is initiated by subsequent rapid infusion of hyperosmolar solution, not by acute hyponatremia per se (37). The precise pathogenesis of exacerbation induced by acute hyponatremia on severe TBI remains unknown. Aquaporin-4 (AQP4), the most abundant water channel in the central nervous system (CNS), was first cloned from lung and brain (20) characterized by a tetrameric conformation (33). The highly polarized location of AQP4 on the membrane of astrocytic foot

194

195

EFFECT OF ACUTE HYPONATREMIA ON TRAUMATIC BRAIN INJURY

TABLE 1 Comparison of Blood Gas Physiological Parameters between the Traumatic Brain Injury (TBI) Group and Hyponatremia ⫹ TBI Group Groups

pH

PCO 2 (kPa)

PO 2 (kPa)

O 2SAT

MABP (mm Hg)

TBI group HT group

7.440 ⫾ 0.031 7.429 ⫾ 0.046 P ⬎ 0.05

3.129 ⫾ 0.152 3.591 ⫾ 0.079 P ⬎ 0.05

20.628 ⫾ 0.775 20.301 ⫾ 0.525 P ⬎ 0.05

0.989 ⫾ 0.001 0.988 ⫾ 0.003 P ⬎ 0.05

91.925 ⫾ 2.577 92.313 ⫾ 1.504 P ⬎ 0.05

Note. All data are expressed as means ⫾ standard error of mean. The t test (equal variance not assumed) (SPSS 9.0) was used for assessment of differences of physiological parameters between the TBI group (n ⫽ 62) and the HT group at 30 min postinjury (n ⫽ 34). No significant difference was observed. Abbreviations: PCO 2, partial pressure of carbon dioxide; PO 2, partial pressure of oxygen; O 2SAT, oxygen saturation rate; MABP, mean arterial blood pressure.

processes opposed to the brain capillaries, pia, and ependymal epithelium points to its critical function in water transport across the blood– brain and brain–CSF interfaces (34, 45). Increase in AQP4 immunoreactivity but not in mRNA expression by hyponatremia has been reported in nontraumatized rat brain (44). We have previously shown that AQP4 mRNA expression decreased in focal contusion associated with edema formation, but not in other edematous diffusely injured regions (21). Given the changes in brain water and electrolytes during acute hyponatremia, we hypothesized that the effect of acute hyponatremia may be mediated through the changes of AQP4 expression in injured brain. Accordingly, our experiments focused on the effect and possible mechanisms of acute hyponatremia in brain injuries by the following assessment: (1) the changes in corrected contusion volume and axonal injury (AI), quantitated by immunoreactivity of ␤-amyloid precursor protein (␤-APP) (36); (2) the permeability of BBB as evaluated by the immunohistochemistry of endogenous IgG (18, 19); (3) the quantitative ultrastructural changes by image analysis; and (4) the pattern of AQP4 protein immunoreactivity and AQP4 mRNA expression level by semi-quantitative reverse transcription polymerase-chain reaction (RT-PCR). MATERIALS AND METHOD

The Replicated Focal and Diffuse Injury Rat Model and Acute Hyponatremia The procedures used were approved by the University Animal Care and Use Committee. The rat trauma model with severe TBI was modified from the impactacceleration model (21, 22, 27). Adult male Sprague– Dawley rats (340 – 400 g) underwent intraperitoneal general anesthesia using ketamine (50 mg/kg) and xylazine (10 mg/kg) and were intubated for mechanical ventilation. A 2-mm burr hole was made 5 mm posterior to the bregma and 3.5 mm right lateral to the midline. A peg-shaped projection was inserted through

the burr hole in contact with the dura mater. The rats were then exposed to a weight-drop trauma device. A 450-g brass weight was allowed to fall freely from a height of 2 m onto the metal helmet to produce focal contusion and axonal injury. Mechanical ventilation of 10 –20 min was immediately started until spontaneous respiration recovered. A right femoral arterial catheter was inserted to monitor blood pressure and other physiological parameters. Blood pH, partial pressure of oxygen and carbon dioxide, and oxygen saturation rate were measured (Table 1) (288 Blood Gas System analyzer, Bayer). The body temperature was maintained between 37 and 38°C by a rectal probe and a heat lamp. The animals were sacrificed at designated intervals by intraperitoneal overdose of pentobarbital (200 mg/kg). The brain was taken out while still fresh or perfused with fixative before retrieval. Acute hyponatremia was induced immediately after the injury by a 0.5-␮g subcutaneous dose of dDAVP (Ferring, Sweden) and intraperitoneal infusion of 12 ml ⫻ 3 140 mM glucose at 30-min intervals to mimic the syndrome of inappropriate secretion of antidiuretic hormone (SIADH) (22, 46). A total of 6 –7 ml blood was collected from each animal; the separated serum was submitted for measurement of serum sodium (CibaCorning 614 analyzer, Japan). Histological, Histochemical, and Immunohistochemical Methodology Anesthetized animals were perfused transcardially, first with normal saline for 5 min and then with 10% buffered formalin for 20 min. Brains were carefully removed and immersed in the same fixative. Two 1.5-mm coronal sections were taken from each brain and routinely processed for paraffin sections. One section included cortical contusion and the other the level of medulla oblongata, 3 mm posterior to the pineal gland. Quantification of contusion. The focal contusion volume was based on a method previously reported (23). Ten serial 10-␮m sections were cut with a mic-

196

KE ET AL.

rotome in 100-␮m increments from the cerebral section for combined histochemical staining of Nissl (NS) ⫹ acid fuchsin (AF) (AF–NS) to identify irreversible neuronal damage and outline the edge of cortical contusion. The brain serial sections were stained with cresyl violet. The acidophilic AF-positive and/or NS-negative neurons were regarded as being at an irreversible stage of neuronal damage. A CAS 200 computerized image analyzer (Micrometer Application v.1.0, Becton– Dickinson) was used to measure the contusion area (in square millimeters) of each serial section, and summed. The contusion volume (in cubic millimeters) was defined as the sum of all contusion areas multiplied by 0.1 mm. To exclude possible interference from the diffuse postinjury cerebral swelling at different times, the volume of the ipsilateral cerebral hemisphere was also calculated. The corrected contusion volume ⫽ (contusion volume)/ (volume of the corresponding section of ipsilateral cerebral hemisphere) ⫻ 100%. Quantification of densities of AI profiles. Two coronal 10-␮m sections were selected: one from the central part of a cerebral section to observe AI profiles within the corpus callosum and bilateral thalami and the other from the medulla oblongata (3 mm posterior to the pineal gland) to quantify the AI profiles in bilateral corticospinal tracts at the medulla oblongata. Immunostaining for ␤-APP was used for the detection of damaged axons in animal brains (36). Sections counterstained with methyl green were examined at 620 nm under a light microscope. Positively stained AI profiles including the axonal swellings and axonal retraction balls were all counted as AI profiles. The sums of AI profiles in two 200⫻ fields at the segment of corpus callosum beneath the cortical contusion and in one 100⫻ field at thalamus region with dense AI were counted, respectively. AI profiles in the whole corpus callosum and bilateral corticospinal tracts at the medulla oblongata were also counted at 100⫻ magnification. Data were presented as the number of AI profiles within each observation field. Immunohistochemical method. The paraffin sections were subjected to microwave antigen retrieval and 0.3% hydrogen peroxide to remove endogenous peroxidase activity. Monoclonal antibody against ␤-APP (Boehringer, Germany; clone 22C11, 1:80), rabbit anti-rat IgG antibody (Vector, Burlingame, CA, 1:600), and rabbit anti-rat AQP4 antibody (Chemicon, Temecula, CA, 1:250), followed by secondary biotinconjugated antibody and streptavidin, were used from the kit, while 3,3⬘-diaminobenzidine (DAB) served as the chromonogen. Tris-buffer solution replaced the primary antibodies in the negative control study. Sections were slightly counterstained by methyl green solution. For immunostaining of endogenous IgG, the blocking solutions A and B from the kit were omitted.

Methodology for Electron Microscopy (EM) For EM, 1 mm 3 fresh brain tissue was taken from the contusion cortex, contralateral cortex, ipsilateral thalamus, contralateral thalamus, and medulla oblongata at 4 h and 1 day postinjury, from both the TBI and the HT groups. Brain samples were fixed in 2.5% glutaraldehyde, postfixed in 1% osmium tetroxide, dehydrated, and embedded in Epon 812. Sixty-nanometer ultrathin sections were stained with uranyl acetate and lead citrate before examination (EM Philips CM 100) at 80 kV. The swelling of perivascular astrocytic foot processes was quantified by the Quantimet 500 ⫹ Processing and Analysis System (Leica, Germany). The sum area of surrounding perivascular astrocytic foot processes, lumenal area of the enclosed capillary, and area of the enclosed capillary were quantified. The corrected area of perivascular astrocytic foot processes ⫽ (sum of areas of surrounding perivascular astrocytic foot processes)/(area of the enclosed capillary). The corrected lumenal area of enclosed vessel ⫽ (lumenal area of capillary)/(area of the corresponding capillary). Data from the bilateral cortex and thalami were collected together as the cerebral hemisphere group and data from the medulla oblongata were regarded as the medulla oblongata group during statistical analysis. Semi-quantification of AQP4 mRNA Expression by RT-PCR The expression of AQP4 mRNA was quantified as previously described (21). Total RNA was isolated from the contusion site, contralateral cortex, ipsilateral thalamus, contralateral thalamus, and the medulla oblongata using Trizol reagent (Life Technologies). One microgram of total RNA was reverse transcribed using the first-strand cDNA synthesis kit (Amersham Pharmacia Biotech). Polymerase chain reaction was performed in a 20-␮l reaction volume containing 2 ␮l first-strand cDNA, 2.5 mM MgCl 2, 200 ␮M dNTPs, 10 pmol of sense and anti-sense primers, 1⫻ PCR buffer II, 0.8 units AmpliTaq Gold DNA polymerase (Perkin– Elmer), and 1 ␮Ci [␣- 32P]dCTP (Amersham Pharmacia Biotech). Primers for the AQP4 gene (GenBank Accession No. U14007) and the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene (GenBank Accession No. M17701.1) were applied and PCR was also performed. Radioactivity of each PCR product was quantified by Instant Imager (Packard, Australia). The normalized AQP4 expression level was formulated as (incorporated radioactivity of AQP4 product)/(incorporated radioactivity of corresponding GAPDH product). Animal Groups Groups were selected at 4 h and 1 day postinjury based on the duration of acute hyponatremia. Groups divided for histochemical and immunohistochemical

EFFECT OF ACUTE HYPONATREMIA ON TRAUMATIC BRAIN INJURY

197

FIG. 1. Bar graphs show the temporal change of serum sodium in the control group, (acute hyponatraemia ⫹ TBI 4 h) postinjury, e.g., HT.4h, and HT 1-day group. All data are expressed as means ⫹ standard deviation. **P ⬍ 0.01, analyzed by t test (SPSS 9.0), compared to the control group (n ⫽ 5 per group).

studies on effect of combined hyponatremia included TBI 4 h (n ⫽ 6), TBI 1 day (n ⫽ 6), HT 4 h (n ⫽ 6), and HT 1 day (n ⫽ 5). The animal groups for EM included control, TBI 4 h, TBI 1 day, HT 4 h, and HT 1 day (n ⫽ 1 for each EM group). For mRNA quantification, the TBI 1-day group (n ⫽ 4) and the HT 1-day group (n ⫽ 3) were designated. Statistical Analysis The differences of corrected contusion volume, AI profiles within white matter regions, and mRNA expression levels between groups (TBI 1 day vs HT 1 day) were analyzed by Mann–Whitney U test. Differences in quantitative EM data were analyzed by independentsamples t test (equal variance not assumed) (SPSS 9.0). P values less than 0.05 were considered significant.

RESULTS

The experimental acute hyponatremia was induced following severe TBI by coadministration of dDAVP plus the intraperitoneal infusion of 140 mM glucose solution. Serum sodium measurement demonstrated the occurrence of hyponatremia in the HT 4-h group. However, the serum sodium recovered to within normal range at 1 day postinjury (Fig. 1). Acute Hyponatremia Exacerbates Cortical Contusion and Axonal Injury Increase of corrected contusion volume. Figure 2 shows significant increase in the corrected contusion volume in the HT group at 4 h postinjury when compared to that in the TBI 4-h group (P ⬍ 0.05). However, no significant difference in corrected contusion

FIG. 2. Comparison of corrected contusional volume between the TBI and HT groups. The corrected contusion volume was expressed as the percentage of volume of the corresponding ipsilateral cerebral hemispheric section. At 4 h (4h) and 1 day after injury, the corrected contusional volume was compared between TBI and HT groups. Data are expressed as means ⫹ standard deviation, *P ⬍ 0.05. The Mann–Whitney U test was used (SPSS 9.0). TBI 4-h group (n ⫽ 6), HT 4-h group (n ⫽ 6), TBI 1-day group (n ⫽ 6), HT 1-day group (n ⫽ 5).

198

KE ET AL.

FIG. 3. Comparison of densities of ␤-APP-positive axonal injury (AI) profiles in different regions with consistent AI were made between the TBI and HT groups by image analysis at 4 h (A) and 1 day (B) postinjury. All data were expressed as means ⫹ standard deviation. The Mann–Whitney U test was used. Significant P values, P ⬍ 0.05, are shown. TBI 4-h group (n ⫽ 6), HT 4-h group (n ⫽ 6), TBI 1-day group (n ⫽ 6), HT 1-day group (n ⫽ 5). rt., right; lt., left.

volume was found between the HT and TBI groups at 1 day postinjury, while the serum sodium concentration had also recovered to the normal range. Increase of ␤-APP positive AI profiles in specific regions. Consistently dense AI positive ␤-APP profiles were seen in the segment of corpus callosum beneath the contusion, whole corpus callosum, bilateral thalami, and corticospinal tracts at the medulla oblongata of the brain stem between the TBI and HT groups at 4

h and 1 day postinjury. Figure 3 demonstrates the exacerbation of acute hyponatraemia on AI, revealing significant increase in the density of AI profiles within the segment of corpus callosum beneath the contusion (P ⬍ 0.05) and the ipsilateral thalamus (P ⬍ 0.05) in the HT 4-h group compared to the corresponding areas in the TBI groups. The AI density did not change significantly in bilateral corticospinal tracts at the medulla oblongata and the whole corpus callosum, where

EFFECT OF ACUTE HYPONATREMIA ON TRAUMATIC BRAIN INJURY

199

FIG. 4. Photomicrographs illustrate the extravasation of endogenous IgG in injured brains from the traumatic brain injury (TBI) and hyponatremia ⫹ TBI (HT) groups detected by immunohistochemistry. (A) Positive IgG staining located in the contusion site at 1 day postinjury from TBI group. The area with positive staining also involves the bordering area of the ipsilateral hippocampus (arrow), ⫻40. (B) Immunoreactivity of endogenous IgG is observed in the cortical contusion and the bordering region of ipsilateral hippocampus (arrow) from the HT 1-day group, ⫻40.

AI in those segments distant to the contusion was also counted. There were no significant differences in AI density between the TBI 1-day and HT 1-day groups. Effect of Acute Hyponatremia on Permeability of BBB in Injured Brain The changes of BBB permeability in cortical contusion, assessed by the extravasation of positive endogeneous IgG staining, were demonstrated and occurred at 4 h and 1 day postinjury. The massive IgG-positive area involved the entire contusion site, bordering areas of the ipsilateral hippocampus and the medulla oblongata (Fig. 4A). However, the distribution pattern of extravasated IgG was not altered prominently in the HT groups compared to that of the TBI groups (Fig. 4B).

Effect of Acute Hyponatremia on Ultrastructural Features in Injured Brain In the TBI and HT groups, degenerated neurons were more easily observed in the contused cortex than in the contralateral noncontused cortex, presenting as peripherally displaced nucleus, increased aggregation of chromatin, and coarse clumping of chromatin. Disappearance of organized Nissl bodies, loss of cristae in the swollen mitochondria, increased dissociated ribosomes within the perikarya, and pyknosis and shrinkage of neurons were also easily observed (Fig. 5). The appearance of numerous electron lucent swollen perivascular astrocytic foot processes was the distinct feature in the injured brains in both TBI and HT groups. Reactive increase of glycogen particles within the swollen perivascular astrocytic foot processes was

FIG. 5. Electronmicrographs show ultrastructural features of injured brain in the TBI group. (A) Degenerated neurons in the contusional cortex at 4 h postinjury show the marked mitochondrial martical swelling and vacuolation, irregular nuclear membrane, and loss of rough endoplasm reticulum (RER), etc., ⫻11,500. (B) The interendothelial tight junction of capillary is not open and microvilli at the vascular lumenal surface are seen (arrow). There are many reactive glycogen particles present within the intracellular space of swollen astrocytic perivascular foot processes in injured brain at 1 day postinjury, ⫻11,500.

200

KE ET AL.

FIG. 6. Electron micrographs illustrate the swelling of the perivascular astrocytic foot processes in cerebral hemispheres in the TBI (traumatic brain injury) and HT (acute hyponatremia ⫹ TBI) groups. Endothelia (arrow), perivascular astrocytic foot processes (*). (A) Swollen perivascular astrocytic foot processes in the ipsilateral thalamus at 4 h postinjury in the TBI group, ⫻5200. (B) Similar pattern of swollen perivascular foot processes in the contralateral thalamus at 4 h postinjury of the TBI group, ⫻5200. (C) Prominent swelling of perivascular astrocytic foot processes investing the whole capillary and compressing the lumen of the enclosed capillary in HT groups, compared to that in the TBI groups, ⫻5200. (D) Prominent swelling of perivascular astrocytic foot processes in the contralateral thalamus in the HT group at 4 h when compared to that in the TBI group, ⫻5200.

present. The endothelial tight junctions and basement membrane were morphologically normal in the brain capillaries. Increased intraendothelial pinocytotic vesicles among lumenal microvilli were also noted (Fig. 5B). In cerebral hemispheres, the swelling of perivascular foot processes started 4 h after injury and persisted 1 day postinjury in the TBI groups. With acute hyponatremia, a significant increase in sum area of perivascular astrocytic foot processes was found in HT 4-h (P ⬍ 0.01) and 1-day (P ⬍ 0.01) groups, compared to the findings in TBI 4-h and 1-day groups. A simultaneous decrease of luminal area of the enclosed capillary was also found in the HT 1-day group (P ⬍ 0.01) when compared to that in the TBI 1-day group (Fig. 6 and Table 2). However, in the medulla oblongata, swelling of perivascular foot processes developed slowly in the TBI

groups, starting from 1 day postinjury (Fig. 7). When combined with acute hyponatremia, excess edematous fluid accumulated within the intracellular space of perivascular astrocytic foot processes in the HT 4-h group, a significant increase of sum area of perivascular astrocytic foot processes was demonstrated in the HT 4-h group (P ⬍ 0.01), compared to that in the TBI 4-h group (Fig. 8A). At 1 day postinjury, swelling of the perivascular astrocytic foot processes in the HT groups was attenuated and no different to that in the TBI 1-day group (Fig. 8B and Table 3). Effect of Acute Hyponatremia on AQP4 Expression in Injured Brain Absence of AQP4 protein expression in the contusion site occurred in the TBI 1-day group (6/6), but had not

EFFECT OF ACUTE HYPONATREMIA ON TRAUMATIC BRAIN INJURY

TABLE 2 Comparison of Corrected Sum Area of Perivascular Astrocytic Foot Processes and Lumenal Area of the Enclosed Capillary in the Injured Cerebral Hemispheres between the TBI and HT Groups Corrected area of perivascular astrocytic foot processes

Corrected lumenal area of the enclosed capillary

TBI 4 h HT 4 h P values

0.8042 ⫾ 0.0713 1.5015 ⫾ 0.1107 0.000

0.4393 ⫾ 0.0209 0.4393 ⫾ 0.0245 0.999

TBI 1 day HT 1 day P values

1.0607 ⫾ 0.0683 1.8843 ⫾ 0.1776 0.000

0.4734 ⫾ 0.0226 0.3700 ⫾ 0.0256 0.003

Group

Note. All data are expressed as means ⫾ standard error of mean. The t test (equal variance not assumed) (SPSS 9.0) was performed to evaluate the differences of corrected sum area of perivascular astrocytic foot processes and lumenal area of the enclosed capillary between the TBI and HT groups in cerebral hemispheres. Abbreviations: TBI, traumatic brain injury; HT, hyponatremia ⫹ TBI. TBI 4-h group (n ⫽ 41), HT 4-h (n ⫽ 40), TBI 1-day (n ⫽ 44), HT 1-day (n ⫽ 40).

changed noticeably in other noncontused regions (Figs. 9A and 9B) when detected by immunohistochemistry. A similar distribution pattern was observed in the HT 1-day group (Figs. 9C and 9D). Using semi-quantitative RT-PCR, AQP4 mRNA expression level was assessed in all samples from the cortex, thalamus, and medulla oblongata of brain stem in the TBI 1-day and HT 1-day groups. The normalized AQP4 mRNA expression levels are shown in Fig. 10. The AQP4 mRNA expression levels in all five regions were evaluated. A significant decrease of AQP4 mRNA expression in the medulla oblongata region was found

201

in the HT 1-day group (P ⬍ 0.05) when compared to that in the TBI 1-day group. No significant difference of AQP4 mRNA expression was found in the other four regions between the HT 1-day and TBI 1-day groups. DISCUSSION

Hyponatremia is a common complication of intracranial disease, especially following head injury or stroke, and has an incidence ranging from 4.6 to 34% (10, 16). This study has demonstrated the role of acute hyponatremia in the exacerbation of injury in focal contusion and axonal injury with severe experimental TBI. The severity of edema in injured brain was also aggravated by acute hyponatremia using the same model (22). Brain cells use two mechanisms to adapt to hyponatremia, namely an increase in CSF flow as a result of increased hydrostatic pressure, and loss of intracellular osmolytic solutes (29, 41). Among the extruded organic solutes, glutamate, glutathione, and taurine are among the most important for neuronal survival. The role played by glutamates is particularly important in cerebral adaptation (31). However, in TBI complicated by hyponatremia, the uptake of glutamate from the extracellular space might be impaired in the injured astrocytes, and it is possible that glutamate induces excitotoxicity in the damaged neurons. Taurine, an osmolyte for volume regulation, has been shown to serve as an antioxidant and to be redistributed in neurons and glia with hypoosmotic stress (6). Hypoosmolar stress also induces the depletion of glutathione, an antioxidant in the brain, and the reduction of antioxidant content will render neurons more susceptible to TBI (11, 17). Furthermore, the impaired active brain buffering, ascribed to active movement of hydrogen ions out of cells and/or move-

FIG. 7. Electron microphotographs show the features of perivascular astrocytic foot processes swelling in the medulla oblongata in traumatic brain injury (TBI) groups. (A) Less swelling of astrocytic perivascular foot processes in the medulla oblongata is observed at 4 h postinjury; swelling of myelinated axons is also present, ⫻5200. (B) Progressive swelling of perivascular foot processes (*) in the medulla oblongata is more severe at 1 day postinjury compared to that in (A), ⫻5200.

202

KE ET AL.

FIG. 8. Electron microphotographs show the different ultrastructural features in the medulla oblongata in hyponatremia ⫹ TBI (HT) groups. (A) Prominent swelling of perivascular foot processes occurs in the medulla oblongata at 4 h postinjury in the HT group, compared to that in TBI group at the same time point, ⫻5200. (B) The swelling of perivascular astrocytic foot processes (*) in the medulla oblongata was attenuated at 1 day postinjury in the HT group. However, the excess edema fluid gradually diffuses into the perikaryal astrocytic processes (arrow) resulting in swelling, ⫻5200.

ment of hydroxyl or bicarbonate ions into cells may also be induced by acute hyponatremia (3). Therefore, acute hyponatremia induces a series of adverse metabolic processes and an unfavorable milieu for damaged neurons, which are concentrated at the penumbra of the contusion. Retrograde degeneration may occur in cortical neurons connected with disrupted axons within the corpus callosum and ipsilateral thalamus. Retrograde degeneration of magnocellular neurons occurs 21 days after axotomy, and neuronal survival is worsened by superimposed chronic hyponatremia (15). Retrograde neuroTABLE 3 Comparison of Corrected Sum Area of Perivascular Astrocytic Foot Processes and Lumenal Area of the Enclosed Capillary in the Injured Medulla Oblongata between the TBI and HT Groups Corrected area of perivascular astrocytic foot processes

Corrected lumenal area of the corresponding capillary

TBI 4 h HT 4 h P value

0.4796 ⫾ 0.0567 1.1467 ⫾ 0.1413 0.001

0.4126 ⫾ 0.0502 0.4635 ⫾ 0.0304 0.399

TBI 1 day HT 1 day P value

0.9757 ⫾ 0.1450 0.7364 ⫾ 0.1300 0.235

0.4660 ⫾ 0.0502 0.5123 ⫾ 0.0309 0.442

Groups

Note. All data are expressed as means ⫾ standard error of mean. The t test (equal variance not assumed) (SPSS 9.0) was performed to evaluate the differences of corrected sum area of perivascular astrocytic foot processes and lumenal area of the enclosed capillary between the TBI and HT groups in the medulla oblongata. Abbreviations: TBI, traumatic brain injury; HT, hyponatremia ⫹ TBI. TBI 4-h group (n ⫽ 10), HT 4-h (n ⫽ 10), TBI 1-day (n ⫽ 10), HT 1-day (n ⫽ 10).

nal death from axotomy ranges from 3 to 11 days postinjury (4, 38). In this injury model, dense secondary axonal injury in the corpus callosum and ipsilateral thalamus was observed at 6 days postinjury in the TBI group (data not shown). The significant increase in contusion volume at 4 h postinjury during acute hyponatremia appears not to be related to retrograde neuronal degeneration, but rather to the effect of acute hyponatremia. While injured brain is evenly exposed to hyponatremia, the distribution of AI and its reactive changes are not uniform. Depending upon the intensity of mechanical forces, two types of axonal injury are recognized. The primary axotomy occurs if the axonal deformation affects more than 20% of the axons, and the secondary axotomy occurs for the strain deformation of 5–10% of axons associated with axonal swelling from 2– 6 h (28). Moreover, small axons usually undergo paranodal swelling without disconnection (42). This study, including both the disconnected and the swollen ␤-APP positive AI profiles, showed their concentration within the segment of corpus callosum beneath the contusion and the ipsilateral thalamus significantly increased in density during the hyponatremic phase (4 h postinjury). It is known that abundant reciprocal projections exist between the cortex and the ipsilateral thalamus (30). Considering that injured axons in the segment of corpus callosum beneath the contusion and in the ipsilateral thalamus contain direct projections, we postulate that such projections may conduct more mechanical forces by impact, resulting in additional axonal damage. This may represent the structural basis for the exacerbation of injury in acute hyponatremia. This hypothesis is further supported by the absence of exacerbation of injury in regions devoid of direct pro-

EFFECT OF ACUTE HYPONATREMIA ON TRAUMATIC BRAIN INJURY

203

FIG. 9. Photomicrographs illustrate the distribution of aquaporin-4 (AQP4) in brains from the traumatic brain injury (TBI) and hyponatremia ⫹ TBI (HT) groups detected by immunohistochemistry. (A) Absence of AQP4 expression occurrs in the contusion (arrow) in the TBI 1-day group. However, positive staining at the microvessels of pericontusional area is observed, ⫻100. (B) Expression of AQP4 in the medulla oblongata in the TBI 1-day group; positive staining is located at the pial surface and microvessels within the neuropil, ⫻40. (C) Absence of AQP4 expression (arrow) is also observed in the contusion site in the HT 1-day group, ⫻100. (D) Expression of AQP4 is located at the pial surface and the capillaries of the medulla oblongata in the HT 1-day group, ⫻200.

jections, such as the left thalamus and bilateral corticospinal tracts at the medulla oblongata during acute hyponatremia. This observation provides supporting evidence and significance for the classification of focal AI and nonfocal AI in severe TBI suggested by AbouHamden et al. (1). According to their work, AI around the focal lesion has been termed focal AI related to the secondary insults, whereas AI distant to the focal lesion was regarded as nonfocal AI or real diffuse axonal injury (DAI). From this viewpoint, our results are consistent with the findings in severe clinical TBI where AI in the segment of corpus callosum and the ipsilateral thalamus constitutes the focal AI exacerbated by acute hyponatremia. The AI in contralateral corpus callosum and bilateral corticospinal tracts at the medulla oblongata is comparable to the nonfocal AI as described (1). Other reports also support the point that differential responses of AI are initiated. No potentiation of AI in the lower medulla and upper cervical cord was observed when secondary insults such as hypotension or pyrexia were superimposed at 4.5 h postinjury in a

diffuse-injury model (25). In a fluid-percussion model, increased AI in the cortex, subcortical white matter, and internal capsule was observed with hyperthermia, but not in the distant striatal region (13). Therefore, the distinction of focal and nonfocal AI is pathophysiologically important, especially when secondary insults may result in different clinical outcomes. Further studies on its pathogenesis may improve treatment of secondary AI in severe TBI patients. The breakdown of BBB determined by leakage of endogenous IgG is mainly located at the contusion site and bordering area of the ipsilateral hippocampus and medulla oblongata. No obvious difference was identified between the trauma group, with or without hyponatremia, indicating that acute hyponatremia does not further alter BBB permeability in injured brain. The effect of hyponatremia on intact BBB function assessed by MRI showed that acute hyponatremia did not increase the permeability in the cortex, white matter, or cerebellum (2). Our observation is consistent with this finding.

204

KE ET AL.

FIG. 10. Comparison of aquaporin-4 (AQP4) mRNA expression levels between the traumatic brain injury (TBI) 1-day group and hyponatremia ⫹ TBI (HT) 1-day group detected by semi-quantitative RT-PCR. (A) Autoradiograph of 32P-incorporated RT-PCR products separated by 4% polyacrylamide gel electrophoresis. (B) Bar chart graph illustrates the normalized AQP4 mRNA expression levels in the TBI 1-day and HT 1-day groups. A significant decrease of AQP4 mRNA expression in the medulla oblongata of HT 1 day group is observed (P ⬍ 0.05) compared to that in the TBI 1-day group. Data are presented as means ⫹ standard deviation, *P ⬍ 0.05; the Mann–Whitney U test was used (SPSS 9.0) and only the significant P value is shown. TBI 1d, TBI at 1 day postinjury (n ⫽ 4), HT 1d, acute hyponatremia ⫹ TBI at 1 day postinjury (n ⫽3). Lanes: 1, contusional cortex; 2, ipsilateral (right) thalamus; 3, contralateral (left) cortex; 4, contralateral (left) thalamus; 5, medulla oblongata of the brain stem; (⫺), negative control (H 2O); (⫹), positive control (cDNA had been reverse transcribed from the total RNA extracted from the normal cortex in the pilot study).

Significant increase in swelling of perivascular astrocytic foot processes is the main ultrastructural feature of superimposed acute hyponatremia in injured brain. In focal contusion with vasogenic edema, the AQP4 expression has been down-regulated in TBI (21) and kept at a low expression level in the HT groups. Therefore, other diffuse injured regions could be the targets affected by acute hyponatremia. Based on the findings of coexistence of AQP4 and one of K ⫹ channels (Kir 4.1), it has been suggested that water transportation across the BBB is mediated by AQP4 coupled with the active excretion of excess extracellular K ⫹ (32). High extracellular potassium also occurs following severe TBI. Thus, in diffusely injured areas with edema, the presence of AQP4 is essential for the resolution of edema fluid. AQP4 expression in those cerebral regions except for the contusion site was not significantly altered by acute hyponatremia. In AQP4 knockout mice, deletion of the AQP4 gene results in slower development of edema and less swelling of astrocytic foot pro-

cesses (26). Therefore, water permeability of BBB within these regions is preserved. Therefore, the significant swelling of foot processes in cerebral regions was observed at 4 h and persisted to 1 day postinjury. In medulla oblongata, water content was significantly increased when combined with acute hyponatremia (22). AQP4 expression was not changed at 4 h, being associated with swelling of perivascular astrocytic foot processes of the astrocyte. However, swelling of perivascular foot processes is alleviated when the down-regulatory effect on AQP4 expression appears at 1 day. Thus, down-regulation of AQP4 mRNA expression in the medulla oblongata may represent a regulatory response induced by acute hyponatremia. The attenuated swelling of astrocytic perivascular foot processes in the medulla oblongata of HT 1 day group supports the hypothesis that down-regulation of AQP4 mRNA is associated with the decrease of water permeability of BBB for the purpose of preventing further water accumulation from hyponatremia. The water

EFFECT OF ACUTE HYPONATREMIA ON TRAUMATIC BRAIN INJURY

content in the medulla may not be changed since the edematous fluid shifts into the perikaryal astrocytic processes. In a recent study on the effect of hyponatremia on intact rat brain, an increase of AQP4 immunoreactivity but not mRNA expression was observed and this phenomenon was ascribed to secondary conformational modifications of the AQP4 protein (44). This may reflect the AQP4 expression in response to systemic hyponatremia in uninjured brain, although much of the mechanism involved in water homeostasis of the injured edematous brain tissue remains to be clarified. The swelling of perivascular astrocytic foot processes and the ensuing volumetric change significantly affect the exchange among cellular compartment, nutrient vessels, and oxygen tension, causing hypoxia (9). The hypoxic/ischemic condition resulting from acute hyponatremia will exacerbate neuronal death in the penumbra of contusion and axonal injury in the specific regions of the corpus callosum and ipsilateral thalamus, which are dependent on a continuous supply of energy generated from oxidative phosphorylation (43). As to why the exacerbating effect of hyponatremia on focal contusion and focal AI was only present at 4 h postinjury, we speculated that it depends upon the main secondary factors. When acute hyponatremia recovered to normal range at 1 day postinjury, other secondary factors turned out to be the main injury mechanism, which was similar in both groups. The compromised corrected luminal area of cerebral capillaries occurred at 1 day. This phenomenon may be ascribed to the effect of brain trauma on vascular tone. As reported previously, even mild concussive injury results in activation of the endothelium (8). In vitro experiments also demonstrated that NO acts as a modulator on vessels at high extracellular potassium concentrations (39). Evaluation of the biosynthetic equilibrium between endothelin-1 (vasoconstricting agent) and nitric oxide (vasodilating effects) in an impact acceleration injury model was performed. A differential expression in time indicates that the reciprocal interaction of these two molecules may control microvascular autoregulation (35). It can be postulated that the time-point of appearance of the compromised luminal area is associated with the above pathophysiological processes, although this remains to be established in the future. The present study demonstrates that acute hyponatremia acts as a secondary insult following severe TBI, exacerbating contusion and AI. Such an exacerbation is not attributable to the deterioration of BBB, but appears to be associated with hypoxic/ischemic factors due to the swelling of perivascular astrocytic foot processes. Down-regulation of AQP4 mRNA expression may represent a molecular adaptation to maintain water homeostasis in injured brain with cytotoxic edema when exposed to acute hyponatremia.

205

ACKNOWLEDGMENTS We express our gratitude to Mr. Kenny C. H. Ho, Mr. Rocky L. K. Ho, and Mr. Ernest C. W. Chak from the Surgical Laboratory and also to Ms. Janet S. K. Tang, Ms. Dong Shu-min, Mr. David S. Y. Lo, and Mr. Hardy C. W. Ko from the Histopathological Laboratory of Anatomical & Cellular Pathology for their technical assistance and kind help in this study. We also thank Ms. Gillian Kew for proofreading the manuscript.

REFERENCES 1.

2.

3.

4.

5. 6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

Abou-Hamden, A., P. C. Blumbergs, G. Scott, J. Manavis, H. Wainwright, N. Jones, and J. McLean. 1997. Axonal injury in falls. J. Neurotrauma 14: 699 –713. Adler, S., D. Williams, and J. G. Verbalis. 1993. Effect of acute and chronic hyponatremia on blood– brain barrier function in the rat. NMR Biomed. 6: 119 –124. Adler, S., and V. Simplaceanu. 1989. Effect of acute hyponatremia on rat brain pH and rat brain buffering. Am. J. Physiol. 256: F113–F119. Agarwala, S., and R. E. Kalil. 1998. Axotomy-induced neuronal death and reactive astrogliosis in the lateral geniculate nucleus following a lesion of the visual cortex in the rat. J. Comp. Neurol. 392: 252–263. Arieff, A. I. 1987. Hyponatremia associated with permanent brain damage. Adv. Intern. Med. 32: 325–344. Aruoma, O. I., B. Halliwell, B. M. Hoey, and J. Butler. 1988. The antioxidant action of taurine, hypotaurine and their metabolic precursors. Biochem. J. 256: 251–255. Atchison, J. W., J. Wachendorf, D. Haddock, W. J. Mysiw, M. Gribble, and J. D. Cornigan. 1993. Hyponatraemia-associated cognitive impairment in traumatic brain injury. Brain Inj. 7: 347–352. Balabanov, R., H. Goldman, S. Murphy, G. Pellizon, C. Owen, J. Rafols, and P. Dore-Duffy. 2001. Endothelial cell activation following moderate traumatic brain injury. Neurol. Res. 23: 75– 82. Bourke, R. S., H. K. Kimelberg, L. R. Nelson, K. D. Barron, E. L. Auen, A. J. Popp, and J. B. Waldman. 1980. In Biology of Glial Swelling in Experimental Brain Edema. Advances in Neurology. Brain Edema (J. Cervo´ s-Navarro and R. Ferszt, Eds.), Vol. 28, pp. 98 –109. Raven Press, New York. Bussmann, C., T. Bast, and D. Rating. 2001. Hyponatraemia in children with acute CNS disease: SIADH or cerebral salt wasting? Childs Nerv. Syst. 17: 58 – 62. Clark, E. C., D. Thomas, J. Baer, and R. H. Sterns. 1996. Depletion of glutathione from brain cells in hyponatremia. Kidney Int. 49: 470 – 476. Cluitmans, F. H. M., and A. E. Meinders. 1990. Management of severe hyponatremia: Rapid or slow correction? Am. J. Med. 88: 162–166. Do´ czi, T., J. Tarja´ nyi, E. Huszka, and J. Kiss. 1982. Syndrome of inappropriate secretion of antidiuretic hormone (SIADH) after head injury. Neurosurgery 10: 685– 688. Dohanics, J., G. E. Hoffman, and J. G. Verbalis. 1996. Chronic hyponatremia reduced survival of magnocellular vasopressin and oxytocin neurons after axonal injury. J. Neurosci. 16: 2373– 2380. Dietrich, W. D., O. Alonso, M. Halley, and R. Busto. 1996. Delayed posttraumatic brain hyperthermia worsens outcome after fluid percussion brain injury: A light and electronmicroscopic study in rats. Neurosurgery 38: 533–541.

206 16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

26.

27.

28.

29.

30.

31.

KE ET AL. Fox, J. L., J. L. Falik, and R. J. Shaloub. 1971. Neurosurgical hyponatremia: The role of inappropriate antidiuresis. J. Neurosurg. 34: 506 –514. Halliwell, B. 1989. Protection against oxidants in biological systems. The superoxide theory of oxygen toxicity. In Free Radicals in Biology and Medicine, 2nd ed. (B. Halliwell, and J. M. C. Gutteridge, Eds.), pp. 86 –187. Clarendon Press, London. Hoshino, S., S. Kobayashi, and S. Nakazawa. 1996. Prolonged and extensive IgG immunoreactivity after severe fluid-percussion injury in rat brain. Brain Res. 711: 73– 83. Jiang, J. Y., B. G. Lyeth, M. Z. Kapasi, L. W. Jenkins, and J. T. Povlishock. 1992. Moderate hypothermia reduces blood– brain barrier disruption following traumatic brain injury in the rat. Acta Neuropathol. 84: 495–500. Jung, J. S., R. V. Bhat, G. M. Preston, J. M. Baraban, and P. Agre. 1994. Molecular characterization of an aquaporin cDNA from brain: A candidate osmoreceptor and regulator of water balance. Proc. Natl. Acad. Sci. USA 91: 13052–13056. Ke, C., W. S. Poon, H. K. Ng, J. C. S. Pang, and Y. Chan. 2001. Heterogeneous responses of aquaporin-4 in oedema formation in a replicated severe traumatic brain injury model in rats. Neurosci. Lett. 301: 21–24. Ke, C., W. S. Poon, H. K. Ng, N. L. S. Tang, Y. Chan, J. Y. Wang, and J. N. K. Hsiang. 2000. The impact of acute hyponatraemia on traumatic brain injury in rats. Acta Neurochir. (Wien) 76: 405– 408. Kochanek, P. M., D. W. Marion, W. Zhang, and J. K. Schiding, M. White, A. Palmer, R. S. B. Clark, M. E. O’Malley, S. D. Styren, and C. Ho. 1995. Severe controlled cortical impact in rats: Assessment of cerebral edema, blood flow and contusion volume. J. Neurotrauma 12: 1015–1025. Kroll, M., M. Juhler, and J. Lindholm. 1992. Hyponatraemia in acute brain disease. J. Intern. Med. 232: 291–297. Lammie, G. A., I. R. Piper, D. Thomson, and F. Brannan. 1999. Neuropathologic characterization of a rodent model of closed head injury—Addition of clinically relevant secondary insults does not significantly potentiate brain damage. J. Neurotrauma 16: 603– 615. Manley, G. T., M. Fujimura, T. Ma, N. Noshita, F. Filiz, A. W. Bollen, P. Chan, and A. S. Verkman. 2000. Aquaporin-4 deletion in mice reduces brain edema after water intoxication and ischemic stroke. Nat. Med. 6: 159 –163. Marmarou, A., A. A. A. Foda, W. Van den Brink, J. Campbell, H. Kita, and K. Demetridou. 1994. A new model of diffuse brain injury in rat: Part I: Pathophysiology and biomechanics. J. Neurosurg. 80: 291–300. Maxwell, W. L., J. T. Povlishock, and D. L. Graham. 1997. A mechanistic analysis of non-disruptive axonal injury: A review. J. Neurotrauma 14: 419 – 440. Melton, J. E., and E. E. Nattie. 1983. Brain and CSF water and ions during dilutional and isomotic hyponatremia in the rat. Am. J. Physiol. 244: R724 –R732. Mitrofanis, J., and L. Mikuletic. 1999. Organisation of the cortical projection to the zona incerta of the thalamus. J. Comp. Neurol. 412: 173–185. Nagelhus, E. A., A. Lehmann, and O. P. Ottersen. 1996. Neuronal and glial handling of glutamate and glutamine during hypoosmotic stress: A biochemical and quantitative immunocy-

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45. 46.

tochemical analysis using the rat cerebellum as a model. Neuroscience 72: 743–755. Nagelhus, E. A., Y. Horio, A. Inanobe, A. Fujita, F. M. Huang, S. Nielsen, Y. Kurachi, and O. P. Ottersen. 1999. Immunogold evidence suggests that coupling of K⫹ siphoning and water transport in rat retinal Muller cells is mediated by a coenrichment of Kir4.1 and AQP4 in specific membrane domains. Glia 26: 47–54. Neely, J. D., B. M. Christensen, S. Nielsen, and P. Agre. 1999. Heterotrtrametic composition of aquaporin-4 water channels. Biochemistry 38: 11156 –11163. Nielsen, S., E. A. Nagelhus, M. Amiry-Moghaddam, C. Bourque, P. Agre, and O. P. Ottersen. 1997. Specialized membrane domains for water transport in glial cells: High-resolution immunogold cytochemistry of an aquaporins-4 in rat brain. J. Neurosci. 17: 171–180. Petrov, T., and J. A. Rafols. 2001. Acute alterations of endothelin-1 and iNOS expression and control of the brain microcirculation after head trauma. Neurol. Res. 23: 139 –143. Pierce, J. E. S., J. Q. Troganowski, D. I. Graham, D. H. Smith, and T. K. McIntosh. 1996. Immunohistochemical characterization of alterations in the distribution of amyloid precursor proteins and ␤-amyloid peptide after experimental brain injury in rat. J. Neurosci. 16: 1083–1090. Rojiani, A. M., J. W. Prineas, and E. S. Cho. 1994. Electrolyteinduced demyelination in rats. I. role of the blood-brain barrier and edema. Acta Neuropathol. 88: 287–292. Ross, D. T., and F. F. Ebner. 1990. Thalamic retrograde degeneration following cortical injury: An excitotoxic process? Neuroscience 35: 525–550. Schuh-Hofer, S. Lobsien, E. Brodowsky, J. Vogt, J. P. Dreier, R. Klee, U. Dirnagl, and U. Lindauer. 2001. The cerebrovascular response to elevated potassium—Role of nitric oxide in the in vitro model of isolated rat middle cerebral arteries. Neurosci. Lett. 306: 61– 64. Soupart, A., R. Penninckx, A. Stenuit, and G. Decaux. 1997. Lack of major hypoxia and significant brain damage in rats despite dramatic hyponatremic encephalopathy. J. Lab. Clin. Med. 130: 226 –231. Sterns, R. H., D. J. Thomas, and R. M. Herndon. 1989. Brain dehydration and neurologic deterioration after rapid correction of hyponatremia. Kidney Int. 35: 69 –75. Stone, J. R., S. A. Walker, and J. T. Povlishock. 1999. The visualization of a new class of traumatically injured axons through the use of a modified method of microwave antigen retrieval. Acta Neuropathol. 97: 335–345. Stys, P. K. 1998. Anoxic and ischemic injury of myelinated axons in CNS white matter: From mechanistic concepts to therapeutics. J. Cereb. Blood Flow Metab. 18: 2–25. Vajda, Z., D. Promeneur, T. Doczi, E. Sulyok, J. Frokiaer, O. P. Ottersen, and S. Nielsen. 2000. Increased aquaporin-4 immunoreactivity in rat brain in response to systemic hyponatremia. Biochem. Biophys. Res. Commun. 270: 495–503. Verkman, A. S., and A. K. Mitra. 2000. Structure and function of aquaporin water channels. Am. J. Physiol. 278: F13–F28. Vexler, Z. S., J. C. Ayus, T. P. L. Roberts, C. L. Fraser, J. Kucharczk, and A. I. Arieff. 1994. Hypoxic and ischemic hypoxia exacerbate brain injury associated with metabolic encephalopathy in laboratory animals. J. Clin. Invest. 93: 256 –264.