Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Biochim Biophys Acta. Author manuscript; available in PMC 2013 November 1.
Published in final edited form as:
PMCID: PMC3444663

Na+ and K+ ion imbalances in Alzheimer’s disease


Alzheimer’s disease (AD) is associated with impaired glutamate clearance and depressed Na+/K+ ATPase levels in AD brain that might lead to a cellular ion imbalance. To test this hypothesis, [Na+] and [K+] were analyzed in postmortem brain samples of 12 normal and 16 AD individuals, and in cerebrospinal fluid (CSF) from AD patients and matched controls. Statistically significant increases in [Na+] in frontal (25%) and parietal cortex (20%) and in cerebellar [K+] (15%) were observed in AD samples compared to controls. CSF from AD patients and matched controls exhibited no differences, suggesting that tissue ion imbalances reflected changes in the intracellular compartment. Differences in cation concentrations between normal and AD brain samples were modeled by a 2-fold increase in intracellular [Na+] and an 8–15% increase in intracellular [K+]. Since amyloid beta peptide (Aβ) is an important contributor to AD brain pathology, we assessed how Aβ affects ion homeostasis in primary murine astrocytes, the most abundant cells in brain tissue. We demonstrate that treatment of astrocytes with the Aβ 25–35 peptide increases intracellular levels of Na+ (~2–3-fold) and K+ (~1.5-fold), which were associated with reduced levels of Na+/K+ ATPase and the Na+-dependent glutamate transporters, GLAST and GLT-1. Similar increases in astrocytic Na+ and K+ levels were also caused by Aβ 1–40, but not by Aβ 1–42 treatment. Our study suggests a previously unrecognized impairment in AD brain cell ion homeostasis that might be triggered by Aβ and could significantly affect electrophysiological activity of brain cells, contributing to the pathophysiology of AD.

Keywords: Alzheimer’s disease, amyloid β, astrocytes, brain, sodium, potassium

1. Introduction

Alzheimer’s disease (AD), the most prevalent form of dementia, is a neurodegenerative disorder characterized by the presence of neurofibrillary tangles, dystrophic neurites and senile plaques composed predominantly of amyloid beta peptide (Aβ) derived from the amyloid precursor protein (1). Aβ is 37–43 amino acids in length and residues 25–35 in the Aβ peptide are the most toxic (2). Among other biochemical aberrations, elevated levels of reactive oxygen species (3), impaired glutamate clearance (4), and depressed Na+/K+ ATPase levels (5) have been reported in AD brain. Also, a decrease in Na+/K+ ATPase activity and protein levels was observed in hippocampus of APP+PS1 transgenic mice, a model for AD (6). These biochemical effects suggest changes in antioxidant capacity, glutamate-based neurotransmission and in membrane polarization-repolarization, processes interconnected at a single metabolic hub. We hypothesized that in AD, impairments in the interconnections extending across astrocytes and neurons linking sulfur metabolism to glutamate-based neurotransmission and ion homeostasis, propagate through this intercellular network and result in coordinated changes in redox and ion regulation affecting both cell types.

Astrocytes are organized in extensive networks poised to dynamically influence neuronal function. The metabolic dependence of neurons on astrocytes renders them vulnerable to astrocytic dysfunction, which can exacerbate neuronal damage under pathological conditions (7, 8). Glutathione (GSH) is a major antioxidant in brain whose synthesis is limited by the availability of cysteine (9). The latter can be synthesized from methionine via the transsulfuration pathway, or derived from imported cystine, which is reduced subsequent to its transport. The twin transporters, xC and XAG, link GSH metabolism to the glutamate-glutamine cycle and connect in turn to Na+/K+ ATPase (10, 11). The XAG system (also known as EAAT transporters) clears glutamate released during synaptic transmission using the sodium gradient in a process that translocates 3Na+ (in) and one K+(out) for every glutamate that is imported into the cell (12). Astrocytes exposed to Aβ show lower glutamate uptake than untreated controls (13, 14) and have a disturbed redox status (15). Oxidative modification and impaired functioning of GLT-1 have been reported in AD brain and in Aβ-treated synaptosomes (4). Glutamate uptake activates the Na+/K+ ATPase, which works to normalize the resting transmembrane ion gradients. Protein levels and the activity of the Na+/K+ ATPase are reported to be lower in AD brains compared to age-matched controls (5). The glutamate gradient in turn drives the xC antiporter, which exchanges extracellular cystine for intracellular glutamate in a stoichiometric process (10). Expression of this antiporter decreases in Aβ treated astrocytes (15). In addition to using the transmembrane sodium gradient, which drives the transport of glutamate and other metabolites, astrocytes take up potassium ions released during the repolarizing phase of action potentials (16) and changes in the astrocytic membrane potential can influence neuronal firing rates (17).

Heterocellular coupling between astrocytes and neurons is indispensable for cycling glutamate, and for redox and ion balance. Impairment at one or more key network nodes in either cell type is expected to have pleiotropic consequences. While perturbations in astrocytic redox balance (15, 18) and glutamate clearance (13, 14) in response to Aβ have received some attention, little is known about the effects of Aβ on the astrocytic cation inventory, which might in turn influence ion conductance changes that underlie many aspects of neuronal function.

To test the hypothesis that impaired glutamate clearance and depressed Na+/K+ ATPase levels observed in AD brain is associated with Na+ and K+ imbalance, we analyzed Na+ and K+ levels in postmortem brain tissues of normal and AD individuals as well as in cerebrospinal fluid (CSF) of AD patients and matched controls. Significant ion imbalances were found in the cortical (Na+) and cerebellar (K+) specimens of AD brain but not in CSF of AD subjects, suggesting that the observed changes in tissue samples reflect changes in the intracellular pool. Mathematical modeling of the experimental data showed that the observed differences in ion concentrations between normal and AD brain tissues can be explained by a 2-fold increase in intracellular [Na+] and a 8–15% increase in intracellular [K+] in AD individuals. To assess whether Aβ might be responsible for the observed ion imbalance in brain cells, we analyzed the effect of Aβ on Na+ and K+ concentrations in cultured primary mouse astrocytes, which are the most abundant cell type in brain. We found that treatment of astrocytes with Aβ 25–35 increased intracellular [Na+] (~2–3-fold) and [K+] (~1.5-fold), which is similar to the increase in intracellular Na+ and K+ levels observed in brain tissues of AD individuals. Similar ion imbalances were observed in astrocytes treated with Aβ 1–40, but not with Aβ 1–42. We speculate that ion imbalance in AD brain might result at least in part, from ion imbalance in astrocytes, which, in turn, might be induced by Aβ. To our knowledge, this study provides the first evidence for significant perturbation of ion homeostasis in AD brain, which could be important for the pathophysiology of brain dysfunction in AD.

2. Materials and methods

2.1. Human Brain and CSF Samples

AD brain and age-matched control samples were from the Michigan Alzheimer’s Disease Research Center Brain Bank. AD samples were derived from clinically well-characterized individuals participating in a prospective brain collection program. At death, one hemisphere was cut into 1–1.5 cm coronal slabs and frozen rapidly over liquid N2 vapor. The other hemisphere was fixed in 10% neutral buffered formalin and used for neuropathologic analysis. All specimens are reviewed by the same neuropathologist and diagnosis established by the Reagan-NIA criteria (19). Slabs were stored in heat-sealed freezer bags at −80 °C until the time of tissue use. Age, gender, and post-mortem delay matched control specimens were derived from individuals with no clinical history of neurologic disease and normal neuropathologic examinations. Slabs were warmed to -20 °C and blocs of frontal cortex, parietal cortex, and cerebellum were prepared. The collection of human brain tissues was approved by the University of Michigan Medical School Institutional Review Board. CSF samples from non-demented individuals (Clinical Dementia Rating (CDR) 0) and those with very mild to moderate Alzheimer’s dementia (CDR 0.5–2) (20) were obtained from Washington University’s Alzheimer’s Disease Research Center. Data for CSF levels of Tau, Ptau181, Aβ 1–42 and data for PIB binding were obtained from the Washington University Biomarker and Imaging Cores, respectively. Details of these procedures have been published previously (21).

2.2. Cell culture

U-87MG (human astrocytoma), Jurkat (human T cell leukemia) and HEK293 (human embryonic kidney) cells were obtained from ATCC. All transformed cell lines were cultured with 100% medium change in 6-well plates with 2 mL medium per well. The following media were used for cultures of transformed cell lines: EMEM (ATCC) for U-87 MG cells, DMEM (Invitrogen) for HEK 293 cells and RPMI-1640 (Invitrogen) for Jurkat cells. All media contained 10% normal FBS (Hyclone), 100 units/mL penicillin and 100 μg/mL streptomycin (Invitrogen).

2.3. Isolation and culture of murine astrocytes

Primary murine cortical astrocyte cultures were prepared from 1–2 day old Balb/C pups as described previously (22, 23). At the end of the third passage, cells were seeded in 6-well (1–2×106 cells/well/2 mL) or 12-well (1×106 cells/well/1 mL) plates and used for experiments after two days. DMEM/F12 mixture containing 10% heat-inactivated FBS, 2 mM additional L-glutamine, 100 units/mL penicillin and 100 μg/mL streptomycin was used as a complete medium for astrocyte cultures. The purity of astrocytes was determined as described previously (23) and was ≥ 93%. The protocols used for handling animals were approved by the University Committee on Use and Care of Animals, University of Michigan.

2.4. Aggregation of Aβ peptide

The lyophilized forms of the triflouroacetate salt of the human Aβ 25–35, Aβ 1–40, Aβ 1–42, and the reversed Aβ 35–25 peptide (Bachem) were aggregated as described previously (24). Briefly, Aβ was dissolved in sterile double-distilled water as a 2 mM stock solution, incubated for 8 days at 37 °C and stored in aliquots at −80 °C. Formation of aggregates was evident from the turbidity of the solution and by observation under a light microscope.

2.5. Aβ treatment of cell cultures

U-87MG and HEK 293 cells were used for experiments at 80–90% confluency. Jurkat cells were used at a cell density of 2×106 cells/well. At the beginning of each experiment, 100 % medium was changed and aggregated Aβ 25–35 was added to a final concentration of 50 μM (concentration estimated prior to aggregation). Cells were collected for ion analysis after 24 h of incubation with Aβ as described below.

Astrocytes were treated acutely or repeatedly with Aβ 25–35 and Aβ 35–25. Aβ 1–40 and Aβ 1–42 were only used in acute treatments. A single bolus of Aβ was added for acute treatment and samples were collected 24 h later. For the repeated treatment model, Aβ was added 3 times at 72 h intervals and samples were collected 24 h after the last treatment. Each time, Aβ was added to astrocytes after 100% change of the complete medium. In each independent experiment (using different preparations of astrocytes), cells were cultured in 2–3 wells in parallel for each experimental condition. Sample from each well was prepared and analyzed separately and the data were averaged. Cell death was observed in <10% of cells in culture following acute or repeated Aβ treatment and was comparable to that seen in the control cell culture as reported previously (15). Cell death was monitored by the PI-AV labeling kit (BioVision) and the TUNEL assay kit (Roche) according to the vendor’s protocols (data not shown).

Ouabaine (500 μM), aspartate β-hydroxomate (400 μM), furosemide (2 mM), bumetanide (10 μM), and dimethyl amiloride (50 μM) (Sigma) when used, were added to astrocytes exposed to repeated treatment with 50 μM Aβ 25–35 at the time of the last Aβ treatment, and cells were collected for ion analysis 24 h later.

2.6. Metabolite analysis

Frozen brain tissue (~100 mg) was pulverized in liquid N2 and the resulting powder (~50 mg) was added to pre-weighed tubes containing 300 μL of 6% TCA (to precipitate protein) for glutamine and methionine analysis or 300 μL of metaphosphoric acid solution (16.8 mg/mL HPO3, 2 mg/mL EDTA and 9 mg/mL NaCl, to precipitate protein) for glutamate analysis. The tubes were weighed again to determine the wet weight of the added tissue and then stored at −80 °C until further analysis. CSF samples were mixed with an equal volume of 10% TCA or metaphosphoric acid solution and stored at −80 °C until further analysis.

Glutamate concentration in CSF and brain samples was measured by HPLC with 2,4-dinitrofluorobenzene derivatization initially described for thiol analysis (25). Concentrations of glutamine and methionine were measured by HPLC using o-phthaldialdehyde derivatization (25).

2.7. Western blot analysis

Astrocytes were collected for Western blot analysis after 24 h of acute treatment or the last repeated treatment with 50 μM of Aβ 25–35. Samples were prepared as described previously (26). Antibodies against GLAST (EAAT1) (Millipore), GLT-1 (SantaCruz Biotech), Na+/K+ ATPase α1 (SantaCruz Biotech) and actin (Sigma) were used to monitor expression of the respective protein antigens and were detected using the Dura chemiluminescent horseradish peroxidase system (Pierce) according to the vendor’s protocol. For quantification of Western blot data, the integrated intensity of protein band of interest was measured using ImageJ software and normalized to the integrated intensity of the actin band in the same samples.

2.8. Measurement of cations

Frozen brain samples (~100 mg) were added to a pre-weighed plastic tube containing 300 μL of deionized water, and the tubes re-weighed. Tissue was homogenized with a plastic pestle on ice, centrifuged at 15000 × g for 5 min and the supernatant was used for Na+ and K+ ion analysis using a flame photometer. CSF samples were used for ion analysis without any modifications. To analyze Na+ and K+ in cells pre-exposed to Aβ, culture medium was aspirated and stored frozen until analysis of the ions in the medium. Cells were washed twice with cold choline chloride solution (150 mM), and re-suspended in 100 μL of choline chloride solution. The cell suspension was freeze-thawed 3 times, cell membranes were removed by centrifugation and ion concentrations in the resulting supernatant were measured using a flame photometer. To measure protein concentration, ~30 μL of the cell suspension was mixed with an equal volume of lysis buffer and protein content was measured using the Bradford method as described previously (25). The resulting ion concentrations were normalized to protein concentration in the cell suspension or normalized to the amount of tissue and presented as mmoles per g of protein or mmoles per kg of tissue, respectively.

2.9. Mathematical model for correlating total, intracellular and extracellular brain [Na+] and [K+]

To estimate intracellular concentrations of Na+ and K+ in brain samples using total (measured) concentration of ions in the tissue, we developed the following model.

Total tissue concentrations of Na+ and K+ are described by equations 1 and 2 for mass conservation:



Here Nat and Kt denote total (measured) Na+ and K+ concentrations in tissue, Nain, Naex, Kin, Kex, denote intracellular and extracellular Na+ and K+ concentrations respectively, Vin, and Vex denote cell volume and extracellular volume in the tissue. The cell volume includes total volume occupied by all cells in the tissue sample and extracellular volume includes all extracellular compartments (interstitial liquid, CSF, blood plasma, etc) in the sample.

For unit volume (Vin+Vex=1), equations (1) and (2) transform to the following form:



Here Vin and Vex are fractions of the tissue volume corresponding to cell and extracellular volumes.

Equations (3) and (4) indicate that total tissue concentrations of Na+ and K+ are linear functions of the relative cell volume. As a consequence, the total K+ concentration in samples of the same tissue with different relative cell volumes must be a linear function of total Na+ concentration. These equations show that if intracellular and extracellular ion concentrations are constant, then an increase in the cell volume fraction in a tissue causes a linear decrease in total (measured) Na+ concentration (because Nain−Naex<0), and a linear increase in total (measured) K concentration (because Kin−Kex>0).

Next, using equation (3) one can express Vin as a function of Nat:


By replacing Vin in the equation (4) with the function in equation (5), we obtain the following equation:


This equation reveals the linear dependence of total K+ concentration on total Na+ concentration in tissue samples if intracellular and extracellular Na+ and K+ concentrations are constant and the cell volume fraction in samples varies. Equation (6) can be rewritten in the following form:




Equations (6)(9) can be used to estimate intracellular tissue Na+ and K+ concentrations using total concentrations of Na+ and K+.

2.10. Statistical Analyses

Statistical analysis and linear fits of experimental data were done using Origin 7 software (OriginLab). Comparison between groups was done using Student’s t test. P ≤ 0.05 was considered as the criterion for significant difference.

3. Results

3.1. Ion imbalances in AD brain

Tissue Na+ and K+ concentrations were determined in human postmortem brain samples of frontal cortex, parietal cortex, and cerebellum from 16 AD subjects and 12 age-matched controls (Table 1). On average, frontal and parietal cortical regions from AD subjects were found to exhibit significantly higher tissue Na+ concentrations, while cerebellum, a region with lower amyloid pathology, exhibited a non-significant trend towards higher tissue Na+ concentrations (Table 2). The average K+ concentrations in frontal and parietal cortex samples of AD patients were not significantly different from K+ concentrations in similar samples of control individuals, while K+ concentrations in cerebellum samples of AD patient were significantly increased (Table 2). Differences in ion concentrations between AD and healthy subjects did not correlate with sex, age and postmortem time of the individual specimens (Table 1, Fig. S1).

Table 1
Postmortem [Na+] and [K+] values in brain tissue of normal individuals and AD patients.
Table 2
Average ion and metabolite concentrations in brain samples from AD patients and age-matched controlsa.

In principle, the observed elevation in Na+ concentration in AD brain samples could reflect increases in Na+ concentration of either the intra- or the extra-cellular (blood, CSF, interstitial) compartment or an increase in the relative size (i.e. volume) of the extracellular (Na+-rich) compartment. An increase in the relative extracellular volume has been reported in a mouse AD model (27). To assess whether ion concentrations change in the extracellular compartment of AD brain tissue, we analyzed Na+ and K+ concentrations in CSF samples from clinically characterized AD subjects and matched controls (Table 3). No difference in Na+ or K+ concentrations between the two groups was observed (Table 3). Also, no difference in CSF Na+ or K+ concentrations was observed between males and females (Table S1). Thus, the increase in Na+ in AD post-mortem tissue samples (Table 2) likely does not reflect changes in the composition of the extracellular compartment. An increase in the relative extracellular volume in AD brain tissue is also inconsistent with the negligible decrease in tissue K+ levels in AD frontal cortex samples and the slight increase in tissue K+ levels in parietal cortex and cerebellum samples. An increase in the relative extracellular volume is expected to result in a significant decrease in K+ levels since the extracellular [K+] [double less-than sign] intracellular [K+]. Based on these lines of reasoning, we conclude that the ion imbalance observed in AD brain tissue is likely to reflect intracellular ion perturbations.

Table 3
Parameters for individual CSF samples, mean and SD values for controls (1–15A) and AD patients (1–15B) used in this study.

3.2. Comparison of glutamate levels between AD and normal subjects

Estimates of glutamate levels in AD versus control brain have produced contrasting results (2830). We measured the glutamate concentration in human brain homogenates and in the CSF. Although slightly higher glutamate levels were observed in all three regions from AD versus control brain samples, the difference did not reach statistical significance (Table 2). In CSF, a modest decrease in glutamate concentration was observed in clinically characterized AD subjects compared to normal controls (Table 3). Differences were not seen in two other metabolites linked to this metabolic hub, methionine and glutamine, in AD versus control cortical samples or in CSF samples, while a significant decrease in methionine was observed in AD cerebellar samples versus controls (Tables 2 and and33).

3.3. Aβ induces Na+ and K+ ion imbalances in astrocytes

Astrocytes are the most abundant cells in brain. To assess whether the ion imbalances seen in brain tissue could be induced by Aβ, we treated primary mouse astrocytes with Aβ 25–35. In response to this treatment, an increase in both intracellular Na+ and K+ content was observed in astrocytes, which was dependent on Aβ concentration (Fig. 1). No changes were observed in the concentrations of these ions in the culture medium. The reverse and inactive Aβ 35–25 peptide had no effect on the cellular Na+ and K+ content (Fig. 1). At 50 μM Aβ 25–35 concentration, where a maximal effect was observed, intracellular Na+ increased ~1.8- and ~2.6 fold while K+ increased ~1.4- and ~1.6 fold following acute and repeated Aβ treatment, respectively (Fig. 2a).

Fig. 1
The effect of different concentrations and forms of Aβ 25–35 on intracellular levels of Na+ and K+ in primary mouse astrocytes
Fig. 2
Changes in ion levels and in the Na+/K+ ATPase expression in astrocytes treated with Aβ 25–35

The Na+/K+ ATPase is primarily responsible for maintaining the Na+ and K+ ion gradients across the plasma membrane and the large ion imbalance observed with Aβ treatment suggested impaired activity of this protein. Western blot analysis showed lower Na+/K+ ATPase levels in astrocytes that were either acutely or repeatedly treated with Aβ 25–35, with the effect being more significant in the latter case (Fig. 2b, c). The decrease in Na+/K+ ATPase expression cannot by itself account for the simultaneous increase in intracellular Na+ and K+ content since inhibition of astrocytic Na+/K+ ATPase with ouabain caused increased intracellular Na+ while concomitantly decreasing intracellular K+ levels (Fig. 3). Addition of ouabain to Aβ 25–35 treated astrocytes selectively increased intracellular Na+ while K+ levels decreased to the levels of ouabain treated controls.

Fig. 3
Effects of ouabain inhibition of Na+/K+ ATPase on intracellular cation levels in astrocytes treated with Aβ 25–35

Since the transport of glutamate and the cations, Na+ and K+, across the plasma membrane are functionally linked, we next examined the effect of Aβ 25–35, if any, on the levels of the astrocytic XAG transporters, GLAST and GLT-1, which were reduced with repeated but not acute Aβ 25–35 treatment (Fig. 4).

Fig. 4
Changes in expression of excitatory amino acid transporters (GLAST and GLT-1) in astrocytes treated with Aβ 25–35

Treatment of astrocytes with one of the following inhibitors: aspartate β-hydroxamate (for xAG), furosemide and bumetanide (for Na/K/Cl co-transport), and dimethyl amiloride (for Na/H exchange), did not significantly influence the cation balance in control cells or affect the cation imbalance in Aβ 25–35 treated cells (data not shown). These results suggest that the corresponding transporters are probably not involved in the ion imbalance induced by Aβ treatment of astrocytes.

The effect of Aβ 25–35 treatment on intracellular Na+ and K+ levels is cell specific. While the human astrocytoma cell line, U-87MG, responded like astrocytes to Aβ treatment, the human embryonic kidney epithelial cell line, HEK 293 cells and human T cell leukemia line, Jurkat cells showed no effect (Fig. 5).

Fig. 5
Cell-specific effect of Aβ 25–35 on intracellular cation levels

We also examined the effects of aggregated the Aβ peptides with longer sequences, 1–40 and 1–42, on Na+ and K+ levels in cultured mouse primary astrocytes. While Aβ 1–42 did not elicit detectable changes in astrocytic ion balance (data not shown), Aβ 1–40 resulted in a concentration-dependent increase in intracellular Na+ and K+ levels (Fig. 6), similar to that observed with Aβ 25–35.

Fig. 6
The effect of Aβ 1–40 on intracellular levels of Na+ and K+ in primary mouse astrocytes

4. Discussion

In this study, we have found differences between Na+ and K+ concentrations in postmortem tissue samples from different segments of AD brain compared to controls and this difference is not reflected in CSF Na+ and K+ concentrations in these two groups. While ion gradients between cells and extracellular compartments disappear after death and intracellular and extracellular Na+ and K+ concentrations equilibrate, the total ion content in a tissue should not be affected because ions are not metabolized and there is no ion exchange with other tissues because of lack of blood circulation. Indeed, the absence of a pronounced correlation between postmortem time of brain collection and tissue ion content (Table 1) confirms that total brain tissue ion content does not change postmortem. A recent study using in vivo MRI sodium imaging revealed that hippocampal formations of AD patients had modestly (~10%) higher [Na+] versus controls (31), which is consistent with our finding of ion imbalance in other brain regions. The increased Na+ concentration was suggested to be primarily within the intracellular compartment (31). Since intracellular [Na+] is >10-fold lower than the extracellular [Na+], even small changes in a total tissue [Na+] are expected to be associated with a significant change in intracellular [Na+].

After excluding the possibility that extracellular ion concentrations contribute to the difference in ion content between normal and AD brain tissue, two other mechanisms were considered: (i) a decrease in the relative cell volume in AD tissue or (ii) an increase in intracellular Na concentration in AD tissue. While considering the first mechanism in more detail in the following section, we note that it alone cannot explain an increase in tissue [Na+] without a corresponding decrease in [K+] or explain the observed increase in cerebellar [K+]. In the first mechanism i.e., a decrease in the relative cell volume in AD brain tissue with a concomitant increase in the extracellular volume, the total Na+ concentration is predicted to increase since the extracellular Na+ concentration is >10-fold higher than its intracellular concentration. The increase in total Na+ concentration in individual samples should be correlated with a decrease in total tissue K+ concentration since extracellular K+ concentration is many-fold lower than its intracellular concentration. To assess this correlation we have employed a mathematical model (equations (1)(9)). The model predicts that if intracellular and extracellular [Na+] and [K+] are constant and only the cell volume fraction in samples vary, then the total tissue K+ concentration will decrease linearly with increasing total Na+ concentration (equations (6) and (7)). The corresponding graph (Fig. 7) shows the dependence of total K+ concentration on total Na+ concentration as the relative intracellular volume varies between 0 and 100%. Hence, at a theoretical zero intracellular cell volume, Na+ concentration is high while total K+ ion is low and corresponds to the extracellular Na+ and K+ ion concentrations. Conversely, at a theoretical zero extracellular volume, Na+ concentration is low while K+ is high and corresponds to the intracellular Na+ and [K+] ion concentrations.

Fig. 7
Modeled dependence of total tissue potassium concentration on total tissue sodium concentration in samples with different relative intracellular volumes

The model predicts that the slope of the graph increases with increasing intracellular cation concentrations (Figs. 7a and b). The lines converge at the same point representing extracellular [Na+] =147 mM and [K+] =2.8 mM and coincide with our experimentally determined values for [Na+] and [K+] in CSF from normal individuals and AD patients. The dotted lines in Fig. 7 correspond to intracellular cation concentrations. As is evident from these plots, the slope is sensitive to increases in intracellular [Na+] (Fig. 7a) or [K+] (Fig. 7b).

To assess how our experimentally derived data obtained in brain fit the model we plotted the tissue [K+] versus tissue [Na+] determined in each sample (Fig. 8). Data points representing control and AD samples segregate to different areas in the graph. Interestingly, despite the relatively small difference in the average ion concentrations between cerebellar samples from AD versus controls, the data points within each group also tended to segregate (Fig. 8). The linear fit for control (solid lines) and AD (dashed lines) data was obtained by constraining the lines to intersect at the point with coordinates of [Na+]=147 mmol/kg tissue and [K+]=2.8 mmol/kg tissue. The fitted lines display an excellent correspondence to the experimental data points for the control set implying that individual variations in total [Na+] and [K+] (ranging from 50 to 90 mmol/kg tissue [Na+] and 89.0 to 49 mmol/kg tissue [K+]) result from variations in the fractional intracellular volume (ranging from 0.71 to 0.42). In contrast, the linear fit to the AD data has a significantly higher slope than the fit to the control data in all three brain regions, consistent with increased intracellular cation levels. The experimental data for AD samples exhibit a much greater scatter than the control data. This implies that in the AD samples, an additional source of individual variations exist, which can be ascribed to variations in intracellular cation concentrations.

Fig. 8
Comparison of Na+ and K+ ion concentrations in brain tissue from AD patients versus matched controls

The linear fit to the control data for frontal cortex yields the following values for parameters representing the constant (A) and slope (B) in equation (7): A=118 mM, B= −0.786. The same values for these parameter can be obtained from equations (8) and (9) using experimentally derived extracellular cation concentrations and at intracellular [Na+] and [K+] concentrations of 10 mM and 110 mM respectively. Thus, the linear fit coincides with the modeled plot in Fig. 7, obtained for normal intracellular Na+ and K+ concentrations, and supports the presence of normal intracellular cation concentrations in control samples. The linear fit to the AD frontal cortex data yields the following parameter values: A=138, B=−0.918. These values can be obtained from equations (8) and (9) at intracellular [Na+] and [K+] of 20 mM and 119 mM respectively. The plot of the AD frontal cortex data reveals a significant elevation in intracellular cation concentrations. The estimates for intracellular cation concentrations for all three brain regions are shown in Table 4 and reveal that the difference in cation concentrations between normal and AD brain tissue samples can be explained by a significant (two-fold) increase in intracellular [Na+] together with a mild (8–15%) increase in intracellular [K+]. Thus, analysis of our data obtained with postmortem brain samples reveals a significant increase in intracellular Na+ concentration and a mild increase in K+ concentration in AD compared to normal brain samples. An obvious limitation of our study is that intracellular Na+ and K+ concentrations cannot be measured directly in frozen postmortem brain samples and we are unable to determine whether the ion imbalance affects all or a subset of brain cell types. Direct measurement of Na+ and K+ ion concentrations in AD brain samples is precluded by loss of the transmembrane ion gradient in postmortem tissue.

Table 4
Estimates of intracellular N+ and K+ concentrations in AD versus control brain samplesa.

A difference in glutamine and glutamate levels was not observed between normal and AD brain samples (Table 2). It seems unlikely that use of postmortem tissue obscured possible changes in these amino acids pool sizes since previous reports indicate that glutamine and glutamate levels in brain do not change significantly up to 48 h postmortem (32, 33). On the other hand, methionine increased significantly in postmortem brain tissue from ~10–50 μmoles/kg of tissue upto 230 μmoles/kg of tissue (32, 33). Hence, the observed difference in methionine levels between normal and AD cerebellum (Table 2) might be related to postmortem methionine accumulation.

We used a cell culture model to assess whether the ion perturbations observed in AD brain samples might be associated with Aβ. Since Aβ induces neuronal death in vitro, it precluded their use in the current experimental protocols, which involve a prolonged exposure of cells to Aβ. Instead, we used astrocytes and primarily, the synthetic Aβ 25–35 peptide. While this peptide is generally considered to be unnatural, there is evidence of its presence in AD brain (34, 35). Its greater affordability has resulted in the widespread use of the Aβ 25–35 peptide as a surrogate for the naturally occurring Aβ 1–40/42 peptides, particularly since it often elicits very similar effects (3641). We note a further caveat i.e., that while the sequences of the human and murine Aβ 25–35 peptide are identical, the sequences of the longer Aβ 1–40/42 peptides are not, raising some questions about the use of the commercially available human Aβ 1–40/42 peptide sequence in ex vivo experiments with murine cells. We note that the Aβ concentration used in our experiments is estimated based on the initial concentration of soluble peptide prior to aggregation. The presence of residual soluble Aβ (monomer, oligomer) in these preparations cannot be ruled out. While the 50 μM Aβ concentration used in our cell culture might be considered to be high, in Alzheimer’s disease, cells are exposed to Aβ in soluble and aggregate form over a period of many years. We speculate that prolonged exposure of tissue to lower Aβ concentrations might elicit some of the same effects as short term exposure to higher Aβ concentrations in cell culture experiments.

Interestingly, in our study, the Aβ 1–40 like the Aβ 25–35 peptide induced astrocytic ion imbalance while Aβ 1–42 had no effect. A difference, albeit in the opposite direction between Aβ 1–40 and Aβ 1–42 has been reported in a study where Aβ 25–35 and Aβ 1–42 caused significant activation of glucose consumption in astrocytes while Aβ 1–40 had no effect (37). These results demonstrate that even a small difference in the length of the Aβ peptide can be associated with a significant difference in the effects that it elicits warranting caution about results obtained with these peptides. Further, we note that the effects of Aβ 1–40 and Aβ 1–42 on ion concentrations in astrocytes were only studied under conditions of acute treatment. Our studies with Aβ 25–35 show that repeated treatment can elicit significantly larger effects on ion levels than acute treatment (Fig. 1, ,2a).2a). Hence, our data do not allow us to exclude the possibility that repeated exposure of brain cells to Aβ 1–42 might also elicit ion imbalance.

Our experiments with cultured astrocytes reveal that both acute and repeated treatment with Aβ 25–35 cause a significant increase in intracellular levels of Na+ and K+ that is associated with reduced levels of the Na+/K+ ATPase, the Na+-dependent glutamate transporters, and as previously reported, the glutamate/cystine antiporter (15). However, the observed ion imbalance (i.e. the simultaneous increase in Na+ and K+ levels in cells) cannot be explained by a decrease in these transporter levels. Additionally, the Na/K/Cl co-transporter(s) and Na/H exchanger(s) do not appear to be involved in the molecular mechanism of Aβ-induced astrocytic Na+ and K+ imbalance.

Aβ is known to influence the permeability of biological membranes (42). A decrease in astrocytic Kir4.1 (an inward rectifying potassium channel) and the BK channel (a calcium sensitive large conductance potassium channel) have been reported in mouse AD models and in human tissue (43). While perturbations in the transport and levels of Ca2+ and some transition metal ions in AD brain tissue and Aβ treated cells have been studied (42, 44, 45), changes in Na+ and K+ have received little if any attention. To our knowledge, this is the first demonstration that Aβ can cause an imbalance in Na+ and K+ levels in astrocytes and the underlying molecular mechanism of this Aβ-induced imbalance warrants further investigation.

Interestingly, the increase in Na+ and K+ levels observed in cultured Aβ-treated astrocytes is comparable to the estimated increases in intracellular Na+ and K+ concentrations in AD brain tissue, suggesting the possibility that Aβ also causes Na+ and K+ imbalance in AD brain cells. Other cell types in brain (neurons, oligodendrocytes and microglia) might also play role in total Na+ and K+ imbalance in AD brain tissue. We note that while Na+ and K+ imbalance in AD brain appears to be related to Aβ, our study does not address whether this is a specific feature of AD pathology that is not associated with other diseases, such as schizophrenia, dementia or autism, and merits investigation.

An increase in intracellular ion concentrations is expected to have widespread ramifications on cellular functions ranging from transport to ion conductance and intercellular signaling. An increase in intracellular Na+ concentration is expected to inhibit the Na+/Ca2+ exchanger leading to accumulation of intracellular Ca2+ that, in turn, can stimulate multiple signaling pathways, including cell death initiation (46). Membrane depolarization caused by a significant ion imbalance in brain is associated with such pathologies as migraine, stroke, and traumatic brain injury (47). These disturbances are likely to contribute to neuronal dysfunctions in AD, including dysfunction of intrinsic neuronal excitability (48), and might trigger cell death pathways contributing to the degeneration and atrophy seen in AD brain. A decreased transmembrane Na+ gradient in astrocytes might be one factor leading to a decrease in glutamate clearance capacity in AD brain. The dramatic decrease in GLAST and GLT-1 protein expression observed in repeatedly Aβ-treated astrocytes if replicated in AD brain, would predict disturbances in synaptic signal transmission. Furthermore, since the transmembrane sodium gradient serves as a driving force for the active transport of many amino acids into cells, dissipation of this gradient would elicit pleiotropic metabolic and physiological dysfunction. On the other hand, the steeper transmembrane K+ gradient observed in Aβ treated astrocytes if replicated in AD brain, would predict inhibition of uptake of K+ released by neurons during the repolarization phase of action potentials (16), affecting neuronal electrophysiological activity.

5. Conclusions

Perturbations of Na+ and K+ ion pools, critical for electrophysiological activity, membrane transport, and other cellular processes, are largely unstudied in AD. Our demonstration that Aβ application elicits profound increases in both intracellular Na+ and K+ concentrations in cultured astrocytes similar to changes in the Na+ and K+ pools found in post-mortem tissues from AD brain, suggests aberrations in a fundamental homeostatic mechanism that might be important in the pathophysiology of this neurodegenerative disease. Furthermore, our results demonstrate that while the cerebellum is considered to be unaffected in AD brain, we see evidence for cation imbalance in this region.


  • Intracellular [Na+] and [K+] levels are increased in brain regions in Alzheimer’s disease patients.
  • Similar ion increases were observed ex vivo in murine astrocytes treated with amyloid β.
  • This ion imbalance might contribute to the pathophysiology of Alzheimer’s disease.

Supplementary Material



Support from the National Institutes of Health (5P50 AG008671 to the Michigan Alzheimer’s Disease Research Center) is gratefully acknowledged. The CSF samples and associated clinical data were generously provided by the Biomarker Core of the Washington University Alzheimer’s Disease Research Center supported by P50AG05681, and P01AG026276 and P01AG03991 from the National Institutes of Health. We gratefully acknowledge Dr. Anne Fagan (University of Washington) for her assistance with providing us information on the CSF samples. This work was supported in part by NIH grants (DK64959 to R.B and NS34709 to R.K).

Abbreviations used

Alzheimer’s disease
amyloid β peptide
levels of Aβ 1–42 peptide in CSF
clinical dementia rating
cerebrospinal fluid
frontal cortex
magnetic resonance imaging
parietal cortex
Results of [11C]-Pittsburgh B compound positron emission tomography imaging
CSF phospho-tau 181 levels
CSF tau levels


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


1. Mattson MP. Pathways towards and away from Alzheimer’s disease. Nature. 2004;430:631–639. [PMC free article] [PubMed]
2. Pike CJ, Cummings BJ, Monzavi R, Cotman CW. Beta-amyloid-induced changes in cultured astrocytes parallel reactive astrocytosis associated with senile plaques in Alzheimer’s disease. Neuroscience. 1994;63:517–531. [PubMed]
3. Butterfield DA, Drake J, Pocernich C, Castegna A. Evidence of oxidative damage in Alzheimer’s disease brain: central role for amyloid beta-peptide. Trends Mol Med. 2001;7:548–554. [PubMed]
4. Lauderback CM, Hackett JM, Huang FF, Keller JN, Szweda LI, Markesbery WR, Butterfield DA. The glial glutamate transporter, GLT-1, is oxidatively modified by 4-hydroxy-2-nonenal in the Alzheimer’s disease brain: the role of Abeta1–42. J Neurochem. 2001;78:413–416. [PubMed]
5. Hattori N, Kitagawa K, Higashida T, Yagyu K, Shimohama S, Wataya T, Perry G, Smith MA, Inagaki C. CI-ATPase and Na+/K(+)-ATPase activities in Alzheimer’s disease brains. Neurosci Lett. 1998;254:141–144. [PubMed]
6. Dickey CA, Gordon MN, Wilcock DM, Herber DL, Freeman MJ, Morgan D. Dysregulation of Na+/K+ ATPase by amyloid in APP+PS1 transgenic mice. BMC Neurosci. 2005;6:7. [PMC free article] [PubMed]
7. Fuller S, Steele M, Munch G. Activated astroglia during chronic inflammation in Alzheimer’s disease-Do they neglect their neurosupportive roles? Mutat Res 2009 [PubMed]
8. Chvatal A, Anderova M, Neprasova H, Prajerova I, Benesova J, Butenko O, Verkhratsky A. Pathological potential of astroglia. Physiol Res. 2008;57(Suppl 3):S101–110. [PubMed]
9. Meister A, Anderson ME. Glutathione. Annu Rev Biochem. 1983;52:711–760. [PubMed]
10. McBean GJ. Cerebral cystine uptake: A tale of two transporters. Trends Pharmacol Sci. 2002;23:299–302. [PubMed]
11. Banerjee R, Vitvitsky V, Garg SK. The undertow of sulfur metabolism on glutamatergic neurotransmission. Trends Biochem Sci. 2008;33:413–419. [PubMed]
12. Danbolt NC. Glutamate uptake. Prog Neurobiol. 2001;65:1–105. [PubMed]
13. Parpura-Gill A, Beitz D, Uemura E. The inhibitory effects of beta-amyloid on glutamate and glucose uptakes by cultured astrocytes. Brain Res. 1997;754:65–71. [PubMed]
14. Harris ME, Wang Y, Pedigo NWJ, Hensley K, Butterfield DA, Carney JM. Amyloid beta peptide (25–35) inhibits Na+-dependent glutamate uptake in rat hippocampal astrocyte cultures. J Neurochem. 1996;67:277–286. [PubMed]
15. Garg SK, Vitvitsky V, Albin R, Banerjee R. Astrocytic Redox Remodeling by Amyloid Beta Peptide. Antioxid Redox Signal. 2011;14:2385–2397. [PMC free article] [PubMed]
16. Amedee T, Robert A, Coles JA. Potassium homeostasis and glial energy metabolism. Glia. 1997;21:46–55. [PubMed]
17. Alvarez-Maubecin V, Garcia-Hernandez F, Williams JT, Van Bockstaele EJ. Functional coupling between neurons and glia. J Neurosci. 2000;20:4091–4098. [PubMed]
18. Behl C, Davis JB, Lesley R, Schubert D. Hydrogen peroxide mediates amyloid beta protein toxicity. Cell. 1994;77:817–827. [PubMed]
19. Consensus recommendations for the postmortem diagnosis of Alzheimer’s disease. The National Institute on Aging, and Reagan Institute Working Group on Diagnostic Criteria for the Neuropathological Assessment of Alzheimer’s Disease. Neurobiol Aging. 1997;18:S1–2. [PubMed]
20. Morris JC. The Clinical Dementia Rating (CDR): current version and scoring rules. Neurology. 1993;43:2412–2414. [PubMed]
21. Fagan AM, Mintun MA, Mach RH, Lee SY, Dence CS, Shah AR, LaRossa GN, Spinner ML, Klunk WE, Mathis CA, DeKosky ST, Morris JC, Holtzman DM. Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid Abeta42 in humans. Ann Neurol. 2006;59:512–519. [PubMed]
22. Garg SK, Banerjee R, Kipnis J. Neuroprotective immunity: T cell-derived glutamate endows astrocytes with a neuroprotective phenotype. J Immunol. 2008;180:3866–3873. [PubMed]
23. Garg SK, Kipnis J, Banerjee R. IFN-gamma and IL-4 differentially shape metabolic responses and neuroprotective phenotype of astrocytes. J Neurochem. 2009;108:1155–1166. [PubMed]
24. Pike CJ, Burdick D, Walencewicz AJ, Glabe CG, Cotman CW. Neurodegeneration induced by beta-amyloid peptides in vitro: the role of peptide assembly state. J Neurosci. 1993;13:1676–1687. [PubMed]
25. Garg SK, Yan Z, Vitvitsky V, Banerjee R. Analysis of sulfur-containing metabolites involved in redox and methionine metabolism. In: Das DK, editor. Methods in Redox Signaling. Mary Ann Liebert; New Rochelle: 2010. pp. 7–11.
26. Garg S, Vitvitsky V, Gendelman HE, Banerjee R. Monocyte differentiation, activation, and mycobacterial killing are linked to transsulfuration-dependent redox metabolism. J Biol Chem. 2006;281:38712–38720. [PubMed]
27. Sykova E, Vorisek I, Antonova T, Mazel T, Meyer-Luehmann M, Jucker M, Hajek M, Ort M, Bures J. Changes in extracellular space size and geometry in APP23 transgenic mice: a model of Alzheimer’s disease. Proc Natl Acad Sci U S A. 2005;102:479–484. [PubMed]
28. Tilleux S, Hermans E. Neuroinflammation and regulation of glial glutamate uptake in neurological disorders. J Neurosci Res. 2007;85:2059–2070. [PubMed]
29. Ernst T, Chang L, Melchor R, Mehringer CM. Frontotemporal dementia and early Alzheimer disease: differentiation with frontal lobe H-1 MR spectroscopy. Radiology. 1997;203:829–836. [PubMed]
30. Rupsingh R, Borrie M, Smith M, Wells JL, Bartha R. Reduced hippocampal glutamate in Alzheimer disease. Neurobiol Aging. 2011;32:802–810. [PubMed]
31. Mellon EA, Pilkinton DT, Clark CM, Elliott MA, Witschey WR, 2nd, Borthakur A, Reddy R. Sodium MR imaging detection of mild Alzheimer disease: preliminary study. AJNR Am J Neuroradiol. 2009;30:978–984. [PMC free article] [PubMed]
32. Perry TL, Hansen S, Berry K, Mok C, Lesk D. Free amino acids and related compounds in biopsies of human brain. J Neurochem. 1971;18:521–528. [PubMed]
33. Perry TL, Hansen S, Gandham SS. Postmortem changes of amino compounds in human and rat brain. J Neurochem. 1981;36:406–410. [PubMed]
34. Kaneko I, Morimoto K, Kubo T. Drastic neuronal loss in vivo by beta-amyloid racemized at Ser(26) residue: conversion of non-toxic [D-Ser(26)]beta-amyloid 1–40 to toxic and proteinase-resistant fragments. Neuroscience. 2001;104:1003–1011. [PubMed]
35. Kubo T, Nishimura S, Kumagae Y, Kaneko I. In vivo conversion of racemized beta-amyloid ([D-Ser 26]A beta 1–40) to truncated and toxic fragments ([D-Ser 26]A beta 25–35/40) and fragment presence in the brains of Alzheimer’s patients. J Neurosci Res. 2002;70:474–483. [PubMed]
36. Terranova JP, Kan JP, Storme JJ, Perreaut P, Le Fur G, Soubrie P. Administration of amyloid beta-peptides in the rat medial septum causes memory deficits: reversal by SR 57746A, a non-peptide neurotrophic compound. Neurosci Lett. 1996;213:79–82. [PubMed]
37. Allaman I, Gavillet M, Belanger M, Laroche T, Viertl D, Lashuel HA, Magistretti PJ. Amyloid-beta aggregates cause alterations of astrocytic metabolic phenotype: impact on neuronal viability. J Neurosci. 2010;30:3326–3338. [PubMed]
38. Kaminsky YG, Marlatt MW, Smith MA, Kosenko EA. Subcellular and metabolic examination of amyloid-beta peptides in Alzheimer disease pathogenesis: evidence for Abeta(25–35) Exp Neurol. 2010;221:26–37. [PubMed]
39. Chen X, Zhang J, Chen C. Endocannabinoid 2-arachidonoylglycerol protects neurons against beta-amyloid insults. Neuroscience. 2011;178:159–168. [PMC free article] [PubMed]
40. Chiu WT, Shen SC, Yang LY, Chow JM, Wu CY, Chen YC. Inhibition of HSP90-dependent telomerase activity in amyloid beta-induced apoptosis of cerebral endothelial cells. J Cell Physiol. 2011;226:2041–2051. [PubMed]
41. Zussy C, Brureau A, Delair B, Marchal S, Keller E, Ixart G, Naert G, Meunier J, Chevallier N, Maurice T, Givalois L. Time-Course and regional analyses of the physiopathological changes induced after cerebral injection of an amyloid beta fragment in rats. Am J Pathol. 2011;179:315–334. [PubMed]
42. Glabe CG, Kayed R. Common structure and toxic function of amyloid oligomers implies a common mechanism of pathogenesis. Neurology. 2006;66:S74–78. [PubMed]
43. Wilcock DM, Vitek MP, Colton CA. Vascular amyloid alters astrocytic water and potassium channels in mouse models and humans with Alzheimer’s disease. Neuroscience. 2009;159:1055–1069. [PMC free article] [PubMed]
44. Supnet C, Bezprozvanny I. The dysregulation of intracellular calcium in Alzheimer disease. Cell Calcium. 2010;47:183–189. [PMC free article] [PubMed]
45. Bonda DJ, Lee HG, Blair JA, Zhu X, Perry G, Smith MA. Role of metal dyshomeostasis in Alzheimer’s disease. Metallomics. 2011;3:267–270. [PMC free article] [PubMed]
46. Bano D, Nicotera P. Ca2+ signals and neuronal death in brain ischemia. Stroke. 2007;38:674–676. [PubMed]
47. Lauritzen M, Dreier JP, Fabricius M, Hartings JA, Graf R, Strong AJ. Clinical relevance of cortical spreading depression in neurological disorders: migraine, malignant stroke, subarachnoid and intracranial hemorrhage, and traumatic brain injury. J Cereb Blood Flow Metab. 2011;31:17–35. [PMC free article] [PubMed]
48. Santos SF, Pierrot N, Octave JN. Network excitability dysfunction in Alzheimer’s disease: insights from in vitro and in vivo models. Rev Neurosci. 2010;21:153–171. [PubMed]