Main

Cells behaving as a perfect osmometer respond to osmotic stress by changing their volume due to free water transfer across the cell membrane. Water movement is generated by osmolal gradient between the extra- and intracellular fluid and maintained until the gradient is dissipated (1).

The brain is considered an imperfect osmometer, because perturbation in the osmolal equilibrium between the plasma and cellular fluid activates cell volume regulatory mechanisms to limit osmotic water shift, to mitigate cell shrinkage or swelling, and to restore brain water content to normal (2, 3).

Cerebral adaptation to hypoosmolar conditions has been extensively studied, and it has been shown to involve a transient increase in brain water and a protracted decrease in electrolyte and organic osmolyte contents. The adaptive response is greatly influenced by the extent and the rate of development of hypotonicity (4–10).

To further define the nature of the volume regulatory adaptation of the brain, the present study using serial in vivo and in vitro proton magnetic resonance (MR) relaxation time measurements was undertaken in adult rabbits during the course of induced hyponatremia. An attempt was made to determine tissue water mobility and to relate it to the progressing hyponatremia. The question was also addressed whether restructuring of brain water is a contributory factor to the successful adaptation to the hypotonic environment.

METHODS

Experiments were performed in four groups of adult Pannon white rabbits. The present study was approved by the Animal Ethical Committee of the Medical University of Pécs.

Each group consisted of five animals. Rabbits on normal diet and fluid intake served as control (group I). In the experimental animals, hyponatremia was induced by simultaneous water loading (140 mmol/L dextrose solution) and DDAVP as follows: group II: hydration period, 3 h; treatment, 0 min, 200 mL (∼10% body weight) dextrose solution intraperitoneally and 2 μg DDAVP in water (Minirin) subcutaneously; at 30 min, 100 mL (∼5% body weight) dextrose solution intraperitoneally and 2 μg DDAVP (Minirin) subcutaneously; group III: hydration period, 24 h; treatment, 0 min, 200 mL dextrose solution intraperitoneally and 2 μg DDAVP in oil (Pitressin) intramuscularly; at 2–3, 8, and 16 h, repeated doses of 100 mL dextrose solution intraperitoneally together with 2 μg DDAVP (Minirin) subcutaneously and 2 μg DDAVP in oil (Pitressin) intramuscularly; group IV: hydration period, 48 h; treatment, 0 min, 200 mL dextrose solution intraperitoneally and 2 μg DDAVP (Minirin) subcutaneously and 2 μg DDAVP in oil (Pitressin) intramuscularly; at 2–5, 8, 16, 24, 32, and 40 h, repeated doses of 100 mL dextrose solution intraperitoneally together with 2 μg Minirin and Pitressin as in group III.

After the hydration protocols were completed, venous blood was taken for assessment of plasma osmolality (freezing-point depression) and sodium concentration (ion-selective electrode). At approximately the same time, T2-weighted MR images were made. After the in vivo studies, the animals were killed, their brains were rapidly removed, and tissue samples of approximately 200 μL were prepared from the parietal cortex, the temporal pole, and from the brainstem for MRS and water content measurements. The tissue water content was determined as the difference between the wet and desiccated tissue weight. Tissue samples were dried at 105°C until no further decrease in weight could be detected.

Magnetic resonance imaging.

In vivo T2-weighted imaging was performed on a Siemens Magnetom SP63 1.5 Tesla whole body MR tomograph with a circular polarized knee coil. In all MR experiments, the effective slice thickness was 3 mm, the image matrix was 256 × 256, and resolution was 0.4 × 0.4 mm. T2-weighted images were made using 16 echoes; the interecho interval was 20 ms, and repetition time was 2000 ms. The T2 map was calculated from the 16 T2-weighted images. T2 times were measured on the T2 map by placing the region of interest on the parietal cortex, temporal pole, and brainstem.

MRS.

Tissue samples were placed in 5-mm diameter NMR glass tubes and were incubated at 40°C for 5 min to reach thermal equilibrium with the magnet temperature. MRS was performed on a Bruker Minispec PC 140 portable MR spectroscope operating at 0.96T (40 MHz).

A 386 AT personal computer was used as a storage scope to adjust the 90 and 180o pulses. T1 relaxation time was measured by an inversion recovery method with eight different time intervals between the 180 and 90o pulses. Repetition time was 5 × T1. T2 relaxation time was obtained by using the Carr-Purcell-Meiboom-Gill sequence: 1000 echoes with 1-ms echo time was applied. Each point was the average of five measurements. The data were transferred to the PC for storage and analysis.

Mathematical analysis.

To quantitatively approximate tissue water fractions according to their mobility, multicomponent analysis of the T2 relaxation decay curves was applied by the nonlinear least squares method (11). The free induction decay of the proton relaxation process follows an exponential function. This function can be described by a multiexponential equation provided that in the tissue studied, there are water compartments with different rates of relaxation and these compartments are not interdependent at the time of measurements. We applied the biexponential fitting for the estimation of the bound and free water compartments.

Two components of the T2 relaxation curve were derived from the following expression:EQUATION 1 where k 1 and k 2 represent the relative contribution of the two sets of protons, and T21 and T22 are the relaxation times of the different components.

Data are expressed as mean ± SEM. For statistical analysis, the nonparametric Wilcoxon test was used.

RESULTS

Our water loading protocol of combined administration of hypotonic dextrose in water and DDAVP induced a progressive decrease in plasma sodium concentration from 140.0 ± 3.5 mmol/L in the control animals to 125.0 ± 5.8 mmol/L (p< 0.01), 114.4 ± 4.9 mmol/L (p< 0.001), and 107.2 ± 3.6 mmol/L (p< 0.001) by the end of the hydration period of 3, 24, and 48 h, respectively. This decrease of plasma sodium level was associated with a parallel decline in plasma osmolality with the corresponding values of 284.2 ± 5.4 mOsm/L (control), 263.6 ± 4.0 mOsm/L (3 h, p< 0.05), 254.4 ± 4.5 mOsm/L (24 h, p< 0.001), and 218.0 ± 5.1 mOsm/L (48 h, p< 0.001).

As no significant difference could be detected in the results of tissues water and in the results of in vitro and in vivo MR relaxation measurements among specimens obtained from the three various brain areas at any given time period, the three individual data were averaged and used as one value per animal in further evaluation.

Despite development of marked hypotonicity, water content expressed as percent of total brain weight remained practically unchanged; it varied only between 80.45 ± 1.85% in control and 80.11 ± 2.57% at 24 h. The animals did not show any clinical signs of elevated intracranial pressure. Essentially the same pattern was seen for T1 relaxation time that averaged 898.5 ± 124.2 ms in the controls followed by a slight insignificant decrease to 844.0 ± 136.5 ms during the period of 24 h of water loading. By contrast, T2 relaxation time increased steadily from the baseline value of 100.2 ± 7.7 ms to attain its maximum of 107.5 ± 8.5 ms (p< 0.05) after 48 h of hydration (Fig. 1, a - c ).

Figure 1
figure 1

Water content (a) and T1 (b) and T2 (c) relaxation times in rabbits exposed to water loading. ★p< 0.05.

Full size image

This response of T2 relaxation to hypotonicity proved to be even more pronounced when measured in vivo; it increased gradually from the control value of 92.7 ± 8.5 to 102.2 ± 9.2 ms at 24 h (p< 0.05) and to 107.9 ± 10.5 ms at 48 h (p< 0.001).

Biexponential analysis of the T2 relaxation curve made it possible to distinguish the fast (T21) and slow (T22) components that corresponded to the bound and the free water fractions.

It was observed that the absolute value of T21 had a transient increase of approximately 19% from the initial value of 37.1 ± 16.1 to 44.1 ± 15.3 ms at 3 h and then returned to below the baseline for the further hydration periods. In contrast, after a slight depression during the first 3 h, T22 increased steadily from 101.0 ± 5.5 ms (3 h) to 106.3 ± 10.8 ms (24 h) and 110 ± 8.7 ms (48 h, p< 0.05) as the hypotonicity progressed (Fig. 2, a - c ).

Figure 2
figure 2

T2 relaxation times (a) and their fast (b) and slow (c) components in rabbits exposed to water loading. ★p< 0.05, ★★p< 0.001.

Full size image

As a consequence, the percent contribution of the bound water fraction represented by T21 was markedly depressed from 6.5 ± 3.0% to 3.6 ± 0.9% (3 h, p< 0.05) and 3.9 ± 0.8% (24 h, p< 0.05); then it approached the initial value of 5.3 ± 2.5% by the end of the study (Fig. 3). It could also be seen that the free water fraction (T22) of 93–96% predominated over the motion-constrained bound water regardless of the state of hydration, and the percent T22 responded to water loading with changes that just mirrored those occurring in percent T21 (Fig. 3).

Figure 3
figure 3

Changes in the ratio of the fast to slow component of T2 relaxation times in rabbits exposed to water loading. ★p< 0.05.

Full size image

DISCUSSION

The present study in which in vivo and in vitro lH-NMR relaxation measurements were used provided evidence that cerebral adaptation to hypoosmolar stress can maintain brain water content and prevent edema formation during development of progressive hyponatremia. Further, we showed that despite the unchanged total brain water, there are marked changes in water fractions with different mobility as indicated by the increasing T2 relaxation time and by the transient reduction in the fast and simultaneous elevation in the slow components of the T2 decay curve. It is tempting to propose that restructuring of brain water can serve as one of the first-line defense mechanisms of the volume regulatory adaptation.

In a previous study, we applied the same lH-NMR technique to quantify the distribution of brain water among physical compartments in newborn rabbit pups by measuring the proton relaxation rate. It was shown that during the first 4 d of life, the bound water fraction declined progressively with age and this decline could be further accelerated in pups completely deprived of fluid intake. Because the changes in water mobility greatly exceeded the moderate reduction in brain hydration, we concluded that brain water had to be restructured when newborn pups were challenged by fluid deprivation (12). It is conceivable that in rabbit brain exposed to hypoosmotic environment, adaptive changes can also take place in the distinctly mobile water fractions.

Magnetic resonance imaging is widely used to investigate the time course, extent, and nature of experimental brain edema (13–18). Prolongation of T1 and T2 relaxation times is regarded as an approximate estimate of the increase of brain water content (16); however, determination of diffusion-weighted and T2-weighted images provides a more reliable approach to characterize the physical state of brain water and the underlying pathophysiologic processes (17–22). Decomposing the T2 relaxation curve into fast and slow components that correspond to the bound and free water fractions has been proven to be a useful method to gain new insight into the brain water metabolism (12). In this regard, it is worth noting that the time course of apparent diffusion coefficient in rat brain after hypoxia-ischemia seemed to be similar to that we observed in the percent values of the bound water fraction during development of hyponatremia (18).

The constant brain water in the face of progressing hyponatremia over a period of 48 h is an unexpected finding of the present study. However, the following possibilities may explain this observation:1) Early in the course of hyponatremia, there may be a water shift between the extra- and intracellular compartment without discernible increase of total brain water (3). 2) Cerebral extracellular volume may decrease by movement of fluid into the cerebrospinal fluid that is then shunted into the systemic circulation (3). 3) After induction of acute dilutional hyponatremia, cellular electrolytes are rapidly lost; the highest rate of electrolyte flux is reached within 3 h (7). 4) Cellular hydration may induce altered conformation of protein molecules, and protein-related entrapment and compartmentation of intracellular potassium has been described (23, 24). This suggests that the reduced cerebral osmolyte content is due partly to direct electrolyte transfer from cell to plasma and also to the inactivation of inorganic cerebral osmolytes (25). 5.) It is an additional possibility that in our experimental settings, the rate of development of hyponatremia is slow enough that the cerebral volume regulation could operate efficiently.

The increasing T2 relaxation time during the hydration process is indicative of greater mobility of water molecules, which means that the brain parenchyma is tending to attain equilibrium with the hypoosmolar plasma. The same elevation can be seen in the slow relaxing component as well, as this component is the major contributor to the net T2 relaxation.

Recent distributional studies provided indirect evidences that aquaporin-4, the major water channel of the brain, may participate in the control of intracranial fluid movement (26–29) and may contribute to the pathogenesis or amelioration of brain edema. One can assume that aquaporin-4 may serve as an entrance port for water, resulting in brain edema during hypoosmolar stress. On the other hand, it may provide exit sites for excess brain fluid driven out of the brain by the increased intracranial pressure (30). The involvement of aquaporin-4 in the brain adaptation to progressive hyponatremia, however, has not been defined.

The major issue of the present study is that in response to water loading and progressive hyponatremia, the bound water fraction of the brain was rapidly reduced, stayed at this low level for 24 h, and then approached the baseline value by the end of the hydration period of 48 h. This observation can be interpreted as indicating that in addition to the loss of all classes of osmotically active solutes, the brain tissue responds to hypoosmolar challenge by liberating free water from the cellular water reservoir to achieve osmotic equilibrium. When the reduction of cellular osmolytes fully compensates for the osmotic stress, no more free water generation is needed and water molecules are bound again to electrically charged cellular macromolecules in a highly ordered manner and create the polarized water multilayer (31). It is tempting to postulate that the reversible water-macromolecular interaction is an important adaptive mechanism to protect brain volume during disturbances in plasma osmolality.