Effects of NH4CL application and removal on astrocytes and endothelial cells

Hepatic encephalopathy (HE) is a complex disorder associated with increased ammonia levels in the brain. Although astrocytes are believed to be the principal cells affected in hyperammonemia (HA), endothelial cells (ECs) may also play an important role by contributing to the vasogenic effect of HA. Following acute application and removal of NH4Cl on astrocytes and endothelial cells, we analyzed pH changes, using fluorescence imaging with BCECF/AM, and changes in intracellular Ca2+ concentration ([Ca2+]i), employing fluorescence imaging with Fura-2/AM. Using confocal microscopy, changes in cell volume were observed accompanied by changes of [Ca2+]i in astrocytes and ECs. Exposure of astrocytes and ECs to 1 – 20 mM NH4Cl resulted in rapid concentration-dependent alkalinization of cytoplasm followed by slow recovery. Removal of the NH4Cl led to rapid concentration-dependent acidification, again followed by slow recovery. Following the application of NH4Cl, a transient, concentration-dependent rise in [Ca2+]i in astrocytes was observed. This was due to the release of Ca2+ from intracellular stores, since the response was abolished by emptying intracellular stores with thapsigargin and ATP, and was still present in the Ca2+-free bathing solution. The removal of NH4Cl also led to a transient concentration-dependent rise in [Ca2+]i that resulted from Ca2+ release from cytoplasmic proteins, since removing Ca2+ from the bathing solution and emptying intracellular Ca2+ stores did not eliminate the rise. Similar results were obtained from experiments on ECs. Following acute application and removal of NH4Cl no significant changes in astrocyte volume were detected; however, an increase of EC volume was observed after the administration of NH4Cl, and EC shrinkage was demonstrated after the acute removal of NH4Cl. This study reveals new data which may give a more complete insight into the mechanism of development and treatment of HE.

of HE. Neurotoxins, alteration of the blood-brain barrier (BBB), impairment of neurotransmission, altered brain energy metabolism and systemic response to infections and neuroinflammation are all factors playing a role in the pathogenesis of HE [1][2][3].
Ammonia is the most studied and well-described neurotoxin that precipitates HE [3][4][5], since an increase in brain ammonia level plays a critical role in its manifestation [5]. Although the correlation between venous ammonia levels and the severity of HE is still uncertain, lowering the blood ammonia concentration is the basis of present treatment of HE [6,7].
Strong evidence suggests that astrocytes, which form the predominant cellular compartment in the brain [8], are the principal neural cells affected by ammonia toxicity [9]. Hyperammonemia (HA) is known to induce astrocyte swelling, which may be a key event in the development of HE [10]. HA, especially the acute elevation of ammonia level, leads to brain edema [11,12]. Mechanisms underlying astrocyte swelling are not fully understood, although the metabolism of ammonia in astrocytes and the production of glutamine, which increases intracellular osmolarity, appear to be involved [13,14]. Many studies have also shown that one of the earliest events in astrocytes exposed to ammonia is a rise in [Ca 2+ ] i [9,13,15,16] that correlates with ammonia concentration [13]. Changes in astrocyte [Ca 2+ ] i may also play an important role modulating brain microcirculation [17]. However, little is known about how the removal of NH 4 Cl, as a treatment of HE, influences the affected cells [18].
Apart from affecting central nervous system cells directly, ammonia may also affect the blood-brain barrier (BBB) [19]. The associated endothelial cells (ECs) are the first cells in contact with blood containing ammonia during HA [20,21]. Although the BBB is, structurally, made of endothelial cells connected by tight junctions and resting on the basal lamina, its function is far more complex, since it is influenced by interactions between endothelial cells, pericytes, smooth muscle cells and astrocytes and, occasionally, neuronal terminals [22]. The BBB is therefore difficult to study, and the effects of ammonia on endothelial cells, whether isolated or in contact with adjacent cells, have not been studied extensively. Data from the literature are controversial due to the use of different models and different NH 4 Cl concentrations. Hyperammonemia did not produce brain edema in experimental animals in vivo [23], but it caused profound swelling of brain slices exposed to ammonia [24]. Another phenomenon that has been suggested as the causative agent in the development of brain edema is change in the cerebral blood flow, resulting in altered intracranial blood volume [21,[25][26][27]. Both mechanisms are strongly influenced by the state of the capillary endothelial cells [20].
Changes of pH and [Ca 2+ ] i in astrocytes following the application of ammonia have been studied [13,18], but there is no information on the acute effect of ammonia removal. The aim of the present study therefore was to further analyze pH and [Ca 2+ ] i changes after the acute removal of ammonia in order to give a more complete insight into the mechanisms of HE treatment. Additionally, the role of changes in EC morphology in the development of brain edema has been evaluated by analyzing changes in pH, [Ca 2+ ] i and volume in ECs following the addition and subsequent removal of ammonia.

Solutions
All solutions were prepared immediately prior to the experiment. The standard bathing solution (SBS) consisted of 150 mM NaCl, 5.4 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES and 10 mM glucose; the pH was adjusted to 7.4 with NaOH. Ca 2+ -free solutions were prepared by substituting the CaCl 2 with 2 mM EGTA as a calcium chelator. For the ammonia solutions, NH 4 Cl substituted the equivalent amount of NaCl in SBS in order to maintain the same osmolarity. Although the exact concentration of NH 4 + in the cerebrospinal fluid of patients with HA is not known, NH 4 Cl levels in the arterial blood of patients with liver dysfunction reach roughly 1 mM and in patients with primary urea cycle disorders NH 4 Cl concentration can reach more extreme values [24], but values higher than 20 mM are improbable. In animal models of acute hyperammonemia, the NH 4 Cl concentration range was between 1.5 and 5 mM [23,24].
To be able to compare our results to data from other authors [13,25] [28] was used. 1-3 day old rats were anesthetized, decapitated and the brains removed aseptically. The meninges were removed and the tissue dissociated by passage through sterile Nitex nylon screens (75 μm mesh size) into 10 ml of culture medium composed of DMEM/F12 (1:1), 10 % (vol/vol) FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. The cell suspension was diluted and plated into 75 cm 2 tissue-culture flasks. Cells were grown at 37°C in a watersaturated air environment containing 5 % CO 2 until confluence. The culture medium was changed every three days. The cell cultures were shaken at 200 rpm overnight and the culture medium changed the following day. The procedure was repeated 3 times to remove small process-bearing cells on the protoplasmic cell layer (mainly oligodendrocyte progenitors and microglia). The cell cultures were trypsinized and plated onto 100-mm culture dishes and further cultured for 7-10 days. Under these conditions, astroglial-enriched cultures were obtained (95-98 % of cells showed immunoreactivity for glial fibrillary acidic protein). Analyses were carried out on astrocytes plated on poly-L-lysine-coated coverslips at a concentration of 50 000 cells per ml that had been grown for 5 days until reaching a sub-confluent state [29].

Endothelial cell cultures
We used the endothelial cell line T24 (ECs) derived from human bladder. Cells were grown in cell flasks in cell medium at 37°C in an environment of water-saturated air/ 5 % CO 2 . Every 3 -4 days (depending on cell density and state of cell confluence) the cell medium was removed and the cell cultures were trypsinized and again plated in cell flasks with cell medium under the same conditions. For experimental purposes ECs were plated on a coverslip.

Measurement of intracellular H + and intracellular Ca 2+ concentrations
Fluorescence imaging was performed using an Axiovert S100 microscope (Zeiss) with Fluar 10x/0.5 and Fluar 40x/1.3 oil objectives combined with a Polychromator II (T.I.L.L. Photonics GmbH, Germany) light source and Coolsnap HQ2 camera (Photometrics). The Visiview 2.0.1 (Visitron Systems GmbH, Germany) computer software was used to acquire images and to measure intensities at different wavelengths. Before the beginning of any experiment the viability of cells on coverslips was checked under a light microscope. The cells were incubated in 0.5 μM BCECF/AM in SBS at room temperature (21°C) in the dark for 15 -20 min, then rinsed with SBS and again incubated in SBS for 10 -15 min for complete de-esterification of intracellular AM esters. Upon completion of labeling, the coverslips with the cells were placed on a coverslip holder on the microscope. A group of cells without visible deformities was located under a light microscope. These cells were studied with microspectrofluorometry, using excitation wavelengths of 490 nm and 440 nm; emitted radiation was passed through a filter that passes wavelengths higher than 530 nm. The images acquired at the two excitation wavelengths were stored at intervals of 5 s to capture the rapid changes during the first minute after application or removal of NH 4 Cl and then at intervals of 60 s for another 9 min. Each cell was outlined manually and its fluorescence intensity under excitation at 490 and 440 nm measured. The background fluorescence intensity was subtracted, and an intensity ratio for each cell calculated. The ratio of fluorescence intensities at 490 nm to that at 440 nm (B 490 /B 440 ) was used to determine [H + ] i . During experiments SBS was exchanged as appropriate for a solution with a different concentration of ammonia which, after 10 min, was again changed for SBS [13].
To determine [Ca 2+ ] i, astrocytes were labeled with Fura-2/AM at 4 μM. The protocol was similar to that for [H + ] i measurement, except that excitation wavelengths of 340 nm (F 340 ) and 380 nm (F 380 ) and a filter passing wavelengths higher than 510 nm were used [13]. The ratio (F 340 /F 380 ) was used to determine [Ca 2+ ] i .

Measurement of cell volume
The imaging system for volume analysis comprised a laser confocal microscope Leica TCS SP (Leica, Heidelberg, Germany) with HC PL Fluotar 20x0.5 and PL APO Fluotar 100x1.4 oil objectives and Leica Confocal Software 2.0 (Leica Microsystems Heidelberg GmbH, Germany). Bitplane Imaris 7.4.2 software (Bitplane AG, Switzerland) was used for image analysis. Before the beginning of each experiment, the viability of cells on coverslips was checked under a light microscope. The cells were incubated in 5 μM membrane stain FM 1-43FX at room temperature (21°C) in the dark for 15 min on the microscope holder and then rinsed with SBS. A group of cells without visible deformities was first identified under a light microscope. 3D images of cells were acquired using confocal microscopy with an excitation wavelength of 488 nm (He/Ne laser). Acquired spectra between 555 nm and 585 nm were used. Series of 32 slices at 100x magnification, with a resolution of 512 x 512, were recorded. Using the Bitplane Imaris 7.4.2 computer software, 3D objects were created from the acquired images and the volume of each cell measured, based on object-containing voxels, at specific time points during the experiment. Relative volume changes were expressed as % change of the original volume.

Measurement of cell areas
Images of those cells acquired during the experiments while measuring [H + ] i were further analyzed with the ImageJ (National Institutes of Health) computer software for measuring cell areas. The contrast of the images was enhanced to better determine cell borders. Each visible cell was then outlined at the beginning of the experiment, before and after each solution exchange and at the end of the experiment. Changes of areas were then calculated.

Statistical methods
Analysis of 101 experiments dealing with changes of fluorescence ratio, area and volume before and after the addition of NH 4 Cl and before and after its removal showed identical directions of changes. The directions of changes are presented as trend plots. We performed a nonparametric Wilcoxon signed rank test for each experiment, to test the hypothesis that there is no change in values of fluorescence ratio recorded before and after the addition of NH 4 Cl, and also before and after the removal of NH 4 Cl. Group analysis was then performed by combining the rank sum tests of all experiments [30]. Weighted means (and pooled SDs) were also calculated for all experiments to assess the average relative change of fluorescence ratio values after each type of treatment. Cell volume and area changes after the application and after the removal of NH 4 Cl were analyzed using the same statistical approach.
N is the number of experiments in one group of experiments (one coverslip = one experiment) and n is the total number of cells studied. All statistical analyses were performed using R computer software. Changes were considered significant at p < 0.01. All numerical results in the text are expressed as weighted means ± pooled standard deviation.

NH 4 Cl triggers intracellular pH changes in astrocytes
Extracellular application of NH 4 Cl triggered a rapid rise in B 490 /B 440 (Fig. 1c). This can be explained by a rapid influx of NH 3 , consuming intracellular H + for NH 4 + formation, thereby increasing the intracellular pH (pH i ). After the initial increase a slow decline in B 490 /B 440 was observed. This recovery of pH i is a consequence of NH 4 + continuing to enter the cells after the NH 3 /NH 4 + equilibrium has been reached, driven by the concentration gradient and membrane potential [31]. After incubation for 10 min in the NH 4 Cl solution, the latter was rapidly exchanged for SBS. The removal of NH 4 Cl resulted in a rapid decrease in B 490 /B 440 , again followed by a slow rise (Fig. 1c). The changes observed after the acute fall of extracellular ammonia level are the result of reversal of the process described above. During these experiments the morphology of the astrocytes remained intact ( Fig. 1a and b). The relative increase of B 490 /B 440 after adding 1 mM NH 4 Cl was 15.2 % ± 2.4 % (p < 0.01; N = 7; n = 80). Addition of 5 mM and 20 mM NH 4 Cl triggered greater increases of 20.1 % ± 2.0 % (p < 0.01; N = 7; n = 79) and 46.3 % ± 6.1 % (p < 0.01; N = 5; n = 60) (Fig. 2a, b and c). Resubstituting the extracellular solutions of 1 mM, 5 mM and 20 mM NH 4 Cl with the standard bathing solution resulted in a relative decrease of B 490 /B 440 of 21.9 % ± 2.5 % (p < 0.01; N = 7; n = 80), 35.9 % ± 2.0 % (p < 0.01; N = 7; n = 79) and 51.6 % ± 2.6 % (p < 0.01; N = 5; n = 60) (Fig. 2d, e and f).
The results clearly show that the exposure of cells to NH 4 Cl leads to a rapid increase of pH i followed by a slow decline and that the removal of extracellular NH 4 Cl leads to a rapid pH i fall and a slow recovery. All pH i changes correlate with the extracellular NH 4 Cl concentration. The results are consistent with the previously described mechanism of intracellular pH i shifts with NH 4 Cl [31,32].  [33]. This mechanism suggests that a rise in pH i should lead to a fall of [Ca 2+ ] i . However our results, as well as those of previous studies [13,15,16], show that [Ca 2+ ] i rises after acute alkalization of cytoplasm by NH 4 Cl (Fig. 3c) and that higher concentrations of NH 4 Cl elicit greater  (Fig. 4a, b and c).
Since the rise in [Ca 2+ ] i following a rise of pH i cannot be explained by an increase of binding sites on cytoplasmic proteins, it must be due either to the influx of Ca 2+ through the cell membrane or to its release from intracellular storage [13]. To check the possibility of Ca 2+ entry into the cells, changes in [Ca 2+ ] i were compared following the exposure of astrocytes to NH 4 Cl in SBS and to astrocytes in the Ca 2+ -free bathing solution. No statistically significant differences were observed (p = 0.21), suggesting that intracellular Ca 2+ stores must be responsible for the rapid increase in [Ca 2+ ] i following exposure to NH 4 Cl. In a Ca 2+ -free bathing solution, exposure to 20 mM NH 4 Cl resulted in an increase of F 340 /F 380 of 54.6 % ± 19.6 % (p < 0.01; N = 4; n = 46).
The suggestion that Ca 2+ is released from intracellular stores was further tested by depleting the intracellular stores. Thapsigargin (500 nM), which blocks Ca 2+ transport into the endoplasmic reticulum (ER) and prevents the refilling of calcium stores, and ATP (100 μM), which stimulates Ca 2+ release from ER stores, were added [13] to the Ca 2+ -free bathing solution before the start of the experiment. Addition of 20 mM NH 4 Cl to the astrocytes pretreated with thapsigargin and ATP resulted in a decrease in F 340 /F 380 of 26.5 % ± 8.5 % (p < 0.01, N = 12; n = 88). After intracellular Ca 2+ stores were depleted, the fall in [H + ] i following the addition of NH 4 Cl resulted in release of H + The release of Ca 2+ from intracellular stores demonstrated after alkalinization of astrocytes by NH 4 Cl is consistent with reported results [13]. Further experiments were therefore performed to determine how the removal of extracellular NH 4 Cl and the acidification of astrocytes influence the intracellular Ca 2+ .
After the astrocytes were incubated for 10 min in a solution containing NH 4 Cl, the solution was rapidly substituted by SBS, at the same time measuring F 340 /F 380 . A temporary rise in F 340 /F 380 was observed that was lower than the rise of F 340 /F 380 seen during the addition of NH 4 Cl (Figs. 3c and 4).
The exchange of the 1 mM NH 4 Cl solution with SBS resulted in a rise in F 340 /F 380 of 12.3 % ± 19.5 % (p < 0.01; N = 7; n = 93). The exchange of bathing solutions of 5 mM and 20 mM NH 4 Cl resulted in rises of 20.4 % ± 16.5 % (p < 0.01; N = 13; n = 171) and 30.4 % ± 19.1 % (p < 0.01; N = 18; n = 266) respectively (Fig. 4d, e and f ). The fall of pH always preceded the increase of intracellular Ca 2+ . These results show that acidification of astrocytes leads to a transient rise in [Ca 2+ ] i , the magnitude being dependent on the concentration of NH 4 Cl in the bathing solution before its washout. This is similar to the results obtained on cultured astrocytes caused by alkalinization and acidification using the weak base trimethylamine [13]. The same authors have shown that, in a different experimental setup using astrocytes within brain slices, alkalinization and acidification by the application and washout of 10 mM trimethylamine does not cause any changes in intracellular Ca 2+ [18]. The difference in the results may, at least in part, be due to the use of brain slices and not isolated astrocytes. Evidently the study on isolated cells is only a minor step in understanding the complex interactions between cells within the nervous system.
To determine the source of the increase of Ca 2+ described above, changes of F 340 / F 380 were also recorded following the removal of NH 4

Changes in astrocyte volume on addition and removal of NH 4 Cl
After determining the changes in H + and Ca 2+ concentrations in astrocytes that occurred on substituting SBS with 20 mM NH 4 Cl, then resubstituting the bathing solution back with SBS, changes in cell volume following these acute events were investigated. Volume changes immediately and approximately 10 min after acute exchange of the bathing solutions were determined by confocal microscopy. However, no significant changes in astrocyte volume during the experiment were detected, which is consistent with other studies, in which a progressive increase of the cell volume of astrocytes exposed to NH 4 Cl was observed only after 1-3 days [34]. Our data are also in agreement with the observation that hyperammonemia does not produce brain edema in vivo [23].

NH 4 Cl triggers intracellular pH changes in endothelial cells
Endothelial cells are an essential part of the blood brain barrier, and their function is regulated by astrocytes as well as by metabolites, cytokines and other biologically active molecules [22,35]. These endothelial cells are involved in the regulation of transport of molecules between the blood and the cerebrospinal fluid, and their function appears to serve specific roles within the nervous system. Endothelial cells in the brain are difficult to study due to their small size and their complex interaction with other cells such as pericytes, smooth muscle cells and astrocytes. Primary endothelial cells from other sources and endothelial cell lines differ from those in the brain; therefore the results obtained on endothelial cell lines must be interpreted cautiously.
Experiments for measuring intracellular H + changes were repeated here with ECs, and similar results to those from experiments with astrocytes were found. Extracellular application of NH 4 Cl triggered a rapid rise in B 490 /B 440 , after which a slow decline in B 490 /B 440 was observed. After 10-min incubation with the NH 4 Cl solution, the solution was rapidly exchanged with SBS. The removal of NH 4 Cl resulted in a rapid decrease in B 490 /B 440 , again followed by a slow rise, due to the same mechanism as in experiments with astrocytes (Fig. 5c). An increase in B 490 /B 440 of 41.8 % ± 6.4 % (p < 0.01;  4 Cl and a decrease of 56.2 % ± 1.6 % (p < 0.01; N = 6; n = 89) after removal of NH 4 Cl were observed (Fig. 5d and e).

Addition and removal of NH 4 Cl stimulates changes in [Ca 2+ ] i in endothelial cells
Changes of [Ca 2+ ] i in ECs were observed using the same protocol as in the experiments with astrocytes. In the SBS both addition and removal of NH 4 Cl resulted in an increase in F 340 /F 380 followed by a slow decline (Fig. 6c). However, the first peak of F 340 /F 380 following the addition of 20 mM NH 4 Cl was much higher (99.8 % ± 27.6 % p < 0.01; N = 4; n = 72) (Fig. 6d) than that after the removal of NH 4 Cl (30.2 % ± 8.3 %; p < 0.01; N = 4; n = 72) (Fig. 6e). Similar results were also expected for the Ca 2+ -free bathing solution, but the increase in [Ca 2+ ] i after the addition of NH 4 Cl was much smaller (34.0 % ± 16.0 %; p < 0.01; N = 2; n = 29). Such a difference was not detected in astrocytes. This could suggest that NH 4 Cl stimulates not only release of Ca 2+ from  4 Cl addition plotted as trends. e -Changes after NH 4 Cl removal plotted as trends. Boxplots on each side present median, upper and lower quartile, minimum and maximum and outliers. Experiments are numbered using consecutive numbers as performed intracellular stores but also an influx of extracellular Ca 2+ into ECs. However, the efflux of Ca 2+ from cells in the Ca 2+ -free bathing solution [36], which could be expressed differently in ECs and in astrocytes, must be taken into consideration. The removal of NH 4 Cl in the Ca 2+ -free bathing solution resulted in an increase of F 340 / F 380 of 40.9 % ± 10.2 % (p < 0. 05; N = 2; n = 29).
Intracellular Ca 2+ stores were further depleted by the addition of thapsigargin to a final concentration of 500 nM and ATP to a final concentration of 100 μM [13] before measuring changes of [Ca 2+ ] i . After the addition of NH 4 Cl, a small increase in F 340 /F 380 (3.9 % ± 2.9 %; p < 0.01; N = 4; n = 73) was observed but, after its removal, a rise in F 340 /F 380 (43.0 % ± 17.0 %; p < 0.01; N = 4; n = 73) similar to that observed in ECs in Ca 2+ -free solution without intracellular Ca 2+ stores being depleted (p = 0.092) was measured.
Our results suggest that the addition of NH 4

The volume of ECs changes following the addition and subsequent removal of NH 4 Cl
Endothelial cells may play a major role in vasogenic mechanisms of HE. It has been shown that brain slices swell after exposure to 10 mM or higher concentrations of NH 4 Cl [24]. In our experiments astrocyte volume did not change significantly even after exposure to 20 mM NH 4 Cl. Therefore experiments on ECs were performed to test the hypothesis that ammonia triggers volume changes in ECs. EC volume changes could result in brain blood flow changes and consequently in changes of intracerebral blood volume or cerebral edema [21,[25][26][27].
Volume changes of ECs were measured after the administration and subsequent removal of NH 4 Cl, using confocal microscopy and computer software to calculate maximum volume changes. The time points at which the solutions were exchanged were the same as in the previous experiments when pH and [Ca 2+ ] i were measured. Exposure of EC to different solutions was relatively short, and it is not possible to exclude further volume changes of ECs that might have appeared after a longer exposure to NH 4 Cl.
Exchanging the SBS with 20 mM NH 4 Cl (Fig. 7e) resulted in an average increase of EC volume of 35.1 % ± 27.6 % (N = 8; n = 37; p < 0.01), and returning to SBS resulted in a markedly smaller cell shrinkage of 11.0 % ± 13.0 % (N = 8; n = 37; p < 0.01) (Fig. 7f ). While our experiments were of relatively short durationor at least too short to detect astrocyte volume changes [34] the volume changes in ECs were demonstrated 10 min after the bathing solution exchange. To obtain a better insight into the changes in cell morphology, EC surface projected areas were also measured. An EC area expansion of 2.6 % ± 6.4 % (N = 4; n = 59; p < 0.01) was recorded after changing SBS for increase of ECs is greater than the area increase of ECs, which suggests that the ECs have become taller. Such morphological changes would decrease the vessel diameter, resulting in altered blood flow. The [Ca 2+ ] i of the astrocytes plays an important role in glial control of brain microcirculation [17], but our study also demonstrates that NH 4 Cl influences EC morphology directly, which could also contribute to the vasogenic effect of HA. However, the changes of EC area could also have an important influence on BBB permeability. We also demonstrated shrinkage of EC following acute removal of NH 4 Cl, which could be an important mechanism in the treatment of HE. It has been shown that the BBB permeability remains unchanged during hyperammonemia [23]. This is in agreement with our data showing unchanged astrocyte morphology and an Fig. 7 Changes in volume of ECs following the addition and subsequent removal of NH 4 Cl. a -A group of ECs observed under confocal microscopy and presented in Imaris Section View in three perspectives. b -The same group of ECs presented as 3D objects. The volume for each cell was calculated based on voxel count. c -Relative area changes after NH 4 Cl addition plotted as trends. d -Relative area changes after removal of NH 4 Cl plotted as trends. e -Relative volume changes after NH 4 Cl addition plotted as trends. f -Relative volume changes after NH 4 Cl removal plotted as trends. Experiments are numbered using consecutive numbers as performed increase in endothelial cell volume at a relatively small change of its surface projected area, but the results must be interpreted with caution as hyperammonemia was only transient. Other authors have reported significant swelling of brain tissue slices exposed to ammonia for a prolonged time up to three day [34]. It is possible that this effect of the prolonged exposure to ammonia results from additional biochemical and pathophysiological mechanisms triggered by elevated ammonia levels in the brain [23].

Conclusion
Acute addition of NH 4 Cl elicits a transient rise of intracellular Ca 2+ concentration in astrocytes and ECs. This is consistent with the results of previous studies [13]. The acute removal of NH 4 Cl leads to a transient rise of intracellular Ca 2+ in both cell types, and the observed increase of [Ca 2+ ] i is shown to be due to release of Ca 2+ from cytoplasmic proteins. The present study also shows that the addition and removal of NH 4 Cl results in an increase and decrease respectively of EC volume. Such effects were, however, absent in astrocytes. Moreover, the removal of NH 4 Cl, which is the basis of modern treatment of HE, resulted in a decrease in EC volume despite an accompanying transient increase in [Ca 2+ ] i . Further studies are needed to determine the connection between these events.