A calcein fluorescence quenching method was applied to measure osmotic water permeability in highly differentiated primary cultures of brain astrocytes from wild-type and aquaporin-4 (AQP-4)-deficient mice. Cells grown on coverglasses were loaded with calcein for measurement of volume changes after osmotic challenge. Hypotonic shock producing twofold cell swelling resulted in a reversible ∼12% increase in calcein fluorescence, which was independent of cytosolic calcein concentration at levels well below where calcein self-quenching occurs. Calcein fluorescence was quenched in <200 ms in response to addition of cytosol in vitro, indicating that the fluorescence signal arises from changes in cytosol concentration. In astrocytes from wild-type CD1 mice, calcein fluorescence increased reversibly in response to hypotonic challenge with a half-time of 0.92 ± 0.05 s at 23°C, corresponding to an osmotic water permeability (Pf) of ∼0.05 cm/s. Pf was reduced 7.1-fold in astrocytes from AQP-4-deficient mice. Temperature dependence studies indicated an increased Arrhenius activation energy for water transport in AQP-4-deficient astrocytes (11.3 ± 0.5 vs. 5.5 ± 0.4 kcal/mol). Our studies establish a calcein quenching method for measurement of cell membrane water permeability and indicate that AQP-4 provides the principal route for water transport in astrocytes.
- water transport
- fluorescence quenching
- brain swelling
- cerebral edema
osmotically driven water transport across cell membranes is an important process in cell and organ physiology for fluid absorption and secretion, cell volume regulation, and other processes. Water permeability is high or subject to regulation in some cell types that contain aquaporin water channels. Astrocytes in the central nervous system express aquaporin-4 (AQP-4) (7, 26), a water-selective transporter that may be important in the pathophysiology of brain edema after various types of brain injury (27, 34). Brain edema was reduced and clinical outcome improved in mice lacking AQP-4 after water intoxication and ischemic stroke (20), and AQP-4 expression in brain astrocytes in vivo is sensitive to many factors, including traumatic brain injury (12, 13, 30, 37), focal ischemia (32), infarction (1), tumor (28), and hyponatremia (33). Primary astrocyte cultures have been used as a model to study AQP-4 regulation. AQP-4 expression in astrocyte cultures is reduced by phorbol esters (23, 39) and hypoxia (40), increased by testosterone (8), and is cell cycle dependent (42). Although cultured astrocytes have been shown to undergo osmotically induced swelling (25), as do all cells, it has been assumed without direct evidence that AQP-4 is functional as a water transport in the plasma membrane.
Several strategies have been devised to measure osmotic water permeability in intact cells, all based on changes in cell volume in response to an osmotic challenge (reviewed in Ref. 35). Light scattering provides a semiquantitative index of cell volume changes in relatively large adherent cells such as macrophages (6). Other methods that have been used to measure water permeability in unlabeled cells include laser interferometry (5) and spatial filtering transmission microscopy (3), both of which rely on cell volume-dependent changes in intracellular refractive index and optical path length. Osmotically induced swelling and shrinking of fluorescently labeled cells has been followed by total internal reflection (4) and confocal microscopy (43), which are based on changes in the concentration of a fluorescent probe dissolved in the cytoplasm. Other approaches to follow osmotically induced changes in cell volume include tracking of fluorescent beads immobilized at the cell surface (11) and laser reflection microscopy (21).
The goal of this study was to measure osmotic water permeability in primary cultures of murine brain astrocytes to define quantitatively the role of AQP-4 in water transport across the astrocyte plasma membrane. The best-established methods for measurement of rapid osmosis in monolayers of cultured cells, confocal and spatial filtering microscopies, did not produce acceptable signal changes in astrocytes, probably because astrocytes have a low profile and complex shape. We then evaluated a calcein self-quenching method reported recently for water transport measurements by Hamann et al. (9). In this method, high concentrations of calcein are loaded into cells to produce volume-dependent changes in total cell fluorescence in response to changes in cytoplasmic calcein concentration. We confirmed volume-dependent changes in cell calcein fluorescence in astrocytes; however, analysis of the origin of the signal changes revealed that calcein quenching by cytosol rather than self-quenching was responsible for the volume-dependent fluorescence signal. We report here a characterization of the calcein quenching method for measurement of cellular osmotic water permeability and demonstrate remarkably reduced water permeability in AQP-4-deficient astrocytes with increased Arrhenius activation energy.
Cell culture. Wild-type and AQP-4 null mice in a CD1 genetic background, the latter generated by targeted gene deletion (18), were bred to produce neonatal mice. Primary astrocyte cultures were prepared from neocortical tissues of 1-day-old mice by a modification of the procedure of Swanson et al. (31). Briefly, the cerebral hemispheres were isolated, the meninges were carefully dissected away, and the hippocampus, basal ganglia, and olfactory bulb were removed. Neocortical tissues were then minced with forceps and incubated in minimal essential medium (MEM) containing 0.25% trypsin and 0.01% DNase for 15 min at 37°C. Dissociated cells were centrifuged at 1,500 rpm for 5 min and resuspended in MEM containing 10% fetal bovine serum and 1% penicillin/streptomycin. Cells were seeded on poly-l-lysine-coated 18-mm-diameter round coverglasses and grown at 37°C in a 5% CO2 incubator with a change of medium twice a week. At confluence (days 12–15), cultures were treated with 10 μM cytosine arabinoside for 48 h to prevent proliferation of other cell types, and the medium was replaced with MEM containing 3% fetal bovine serum and 0.15 mM dibutyryl cAMP to induce differentiation. Cultures were continued for 2 more weeks with a change of medium twice a week. Immunocytochemistry showed that >95% of the cells stained positively for the astrocytic marker, glial fibrillary acidic protein (GFAP). For some experiments, Chinese hamster ovary (CHO) cells stably expressing AQP-1 were generated and cultured as described previously (17).
Water permeability measurements. For fluorescence measurements, cells on round coverglasses were loaded with calcein by incubation for 15 min with calcein-AM (Molecular Probes; 0.1–10 μM) at 23°C. After being rinsed in PBS (pH 7.4), the coverglasses were mounted in a custom perfusion chamber (see Fig. 2A) designed for rapid solution exchange without causing cell detachment. The chamber consisted of a stainless steel base with an acrylic insert containing perfusion channels and a 16-mm-diameter, 2-mm-wide rubber O-ring as a spacer. The chamber volume was 50 μl. Solution exchange time was <200 ms at a 50 ml/min perfusion rate as used in water transport measurements. For temperature dependence studies, an in-line heater/cooler was constructed in which perfusate passed through 80 cm of tubing enclosed by a water jacket with circulating fluid. Effluent temperature was monitored via an in-line thermistor. For osmotic water permeability measurements, solutions were exchanged between PBS and hypotonic saline (diluted with distilled water) or hypertonic saline (with added 100 mM d-mannitol). Solution osmolalities were measured using a freezing point-depression osmometer.
Microscopy. Calcein fluorescence was measured continuously using a Nikon Diaphot inverted epifluorescence microscope equipped with halogen light source, calcein filter set (480-nm excitation, 490-nm dichroic mirror, 535-nm emission), photomultiplier detector, and 14-bit analog-to-digital converter. Most experiments were done with the use of a ×40 objective lens [oil immersion, numerical aperture (NA) 1.3]. For some studies, ×16 (NA 0.25), ×50 (NA 0.55), ×100 (oil immersion, NA 1.4) objectives were used. In measurements involving spatial filtering microscopy, monochromatic transmitted light was measured in unlabeled cells by using phase-contrast optics as described previously (3).
Fluorescence measurements. Steady-state fluorescence quenching titrations were performed on a SPEX fluorimeter (494-nm excitation, 512-nm emission). Fluorescence quenching kinetics were measured on a stopped-flow apparatus with <1-ms dead/mixing times. Calcein fluorescence was measured in response to rapid mixing of equal volumes of saline solutions containing calcein and cytosol. Rat liver cytosol was isolated by homogenization and high-speed centrifugation to remove membranes and debris. Calcein concentration in cytoplasm was deduced by fluorescence correlation spectroscopy (FCS) from the zero-time correlation amplitude, G(0), which is related to reciprocal fluorophore concentration (10, 19). The FCS apparatus consisted of an inverted fluorescence microscope containing a 488-nm diode laser light source, 490-nm dichroic, ×100 objective, 535-nm emission filter, avalanche photodiode, and hardware correlator.
Measurements were made on primary cultures of astrocytes at 26–30 days after isolation from neonatal brains of wild-type and AQP-4 null mice. Phase-contrast microscopy showed highly differentiated astrocytes grown on glass coverslips (Fig. 1A). There was no difference in appearance of astrocyte cultures from wild-type vs. AQP-4 null mice, and there was no difference in cell growth rate as judged by cell number at 1–4 wk after plating. By GFAP staining, astrocyte cultures were judged to be >95% pure, with little contamination by neurons, microglia, or other cell types (Fig. 1B). Immunocytochemistry confirmed a plasma membrane pattern for AQP-4 expression in astrocytes from wild-type mice as found previously (24, 37, 39), with no detectable staining in astrocytes from AQP-4 null mice (not shown).
Initial studies indicated a very rapid osmotic equilibration time of ∼1 s or less in astrocytes from wild-type mice. A custom perfusion chamber was thus designed to permit fluid exchange in 100–200 ms without causing detachment of astrocytes from their glass support (Fig. 2A). Figure 2A, right, shows the rapid solution exchange as measured from changes in light transmittance upon exchange between colorless and dye-containing aqueous solutions.
Several optical methods were evaluated for their suitability in following changes in astrocyte volume after osmotic challenge. Experiments were done on astrocyte cultures and cultures of stably transfected, AQP-1-expressing CHO cells. The CHO cells are large and highly water permeable, and they have been used to measure osmotically induced cell volume changes by a variety of optical methods. Figure 2B, top, shows very small signals in astrocytes as determined by spatial filtering transmitted light microscopy, probably because of their complex shape and low profile. Good signals were found in CHO cells, as found previously for several types of cultured cells (3). The signal from astrocytes was improved little by changing wavelength or optical configuration. Also, poor signals were found in astrocytes by using light scattering and fluorescence confocal microscopy, whereas good signals were seen in the CHO cells (not shown). On the basis of a recent report suggesting calcein self-quenching as a method to measure water transport in cultured cells (9), we measured fluorescence in astrocytes and CHO cells after loading with calcein-AM, a cell-permeable calcein derivative that is cleaved and trapped in the cytoplasm. Robust changes in calcein fluorescence in response to osmotic gradients were found in astrocytes and CHO cells (Fig. 2B, bottom).
The calcein quenching method was characterized by using the CHO cell cultures to validate its application to cell water permeability measurements and to establish the quenching mechanism. Figure 3A, left, shows that the magnitude of the change in calcein fluorescence is sensitive to the size and direction of the osmotic gradient, with an approximately linear dependence of fluorescence signal (ΔF/F) on the relative change in cell volume (computed from the ratio of solution osmolalities). To determine whether partial confocality of the optical system was responsible for the calcein signal change (due to high numerical aperture detection of changes in cytosolic calcein concentration), measurements were made by using objective lenses of different magnifications and numerical apertures. Figure 3B shows similar ΔF/F values for very different objectives, which rules out confocality effects and indicates that cell volume alters intrinsic calcein fluorescence. To investigate whether calcein self-quenching was responsible for the fluorescence signal change, as suggested by Hamann et al. (9), measurements were made at different intracellular calcein concentrations, reasoning that signals should be near-zero for low calcein concentration and increase as calcein concentration enters the range at which self-quenching occurs. Figure 3C, top, shows that ΔF/F was not sensitive to cytosolic calcein concentration even at concentrations orders of magnitude below that at which self-quenching can occur (14). Cytosolic calcein concentration (for very low calcein) was measured directly by fluorescence correlation spectroscopy and extrapolated to higher calcein concentrations. The similar G(0) for calcein-loaded cells and 5 nM calcein in saline (Fig. 3C, bottom) indicates that cytosolic calcein concentration was ∼5 nM; the approximately fourfold-increased correlation time for calcein in cytosol (right-shifted inflection point) is due to molecular crowding effects that slow solute diffusion (36).
We postulated that cytosolic calcein was quenched by intracellular components (proteins or salts) whose concentrations change upon cell swelling and shrinking. Figure 4A shows a solution fluorescence quenching study of calcein by cytosol isolated from mouse liver homogenate after removal of membranes by high-speed centrifugation. Cytosol was an effective quencher of calcein fluorescence in a concentration-dependent manner. Figure 4A, inset, shows rapid kinetics (<200 ms) of calcein fluorescence quenching by cytosol as measured in a stopped-flow apparatus. To determine whether proteins or salts were responsible for calcein quenching, calcein fluorescence was measured in the presence of different concentrations of albumin and NaCl. Figure 4B shows effective calcein quenching by albumin, but not by NaCl, in the cellular concentration range. These results suggest that calcein fluorescence signal in cells is sensitive to osmotic gradients because of changes in cytosolic protein concentration that alter calcein quenching, explaining the increased fluorescence with cell expansion in response to reduced solution osmolality.
The calcein fluorescence quenching method was applied to compare osmotic water permeability in wild-type vs. AQP-4-deficient astrocytes. Figure 5, left, shows representative data for the kinetics of astrocyte cell swelling in response to solution exchange between saline and hypotonic saline. The kinetics of cell swelling and shrinking were remarkably slowed in AQP-4-deficient astrocytes. The confocal micrograph of calcein-loaded wild-type astrocytes showed fairly uniform labeling of cytosol (Fig. 5, middle). Figure 5, right, summarizes swelling rates for different cell cultures, showing an average 7.1-fold reduction in osmotic water permeability in AQP-4-deficient astrocytes at 23°C (P < 0.0001). With the use of an astrocyte surface-to-volume ratio of 6,000 cm—1 based on the diameter of the astrocyte body, osmotic water permeability coefficients computed from the time course data (35) of control and AQP-4-deficient astrocytes were 0.05 and 0.007 cm/s, respectively.
Temperature dependence measurements were done to determine the Arrhenius activation energy for astrocyte water transport, a biophysical indicator of water transporting mechanisms. Figure 6A, left, shows that the rate of astrocyte swelling in response to an osmotic gradient was temperature dependent, as expected, with greater temperature sensitivity for the AQP-4-deficient astrocytes. Figure 6A, right, shows an Arrhenius plot of logarithm of swelling rate vs. reciprocal temperature. Plots were approximately linear, with computed activation energies (from slopes) of 5.5 ± 0.4 and 11.3 ± 0.5 kcal/mol for control and AQP-4-deficient astrocytes, respectively. The relatively low activation energy for control astrocytes suggests a channel/pore pathway for water movement, whereas the high activation energy for AQP-4-deficient astrocytes is consistent with diffusive water transport through the lipid portion of the cell membrane.
Additional characterization of the water pathway in AQP-4-deficient astrocytes was done by examination of inhibitor effects and solute reflection coefficients. The inhibitors HgCl2 and phloretin were studied at concentrations of 100 and 200 μM, respectively, where toxic effects were not seen during the brief time course of the experiments. As shown in Fig. 6B, neither HgCl2 nor phloretin inhibited water permeability (relative water permeabilities were 0.96 ± 0.06 and 0.92 ± 0.1, respectively, n = 4), in agreement with the conclusion above that the lipid bilayer provides the AQP-4-independent water transport pathway in astrocytes. Figure 6C shows that mannitol and NaCl produced comparable reductions in calcein fluorescence (cell shrinking), indicating similar reflection coefficients. Assuming a mannitol reflection coefficient of unity, computed NaCl reflection coefficients (SE, n = 5) were 0.96 ± 0.07 and 0.99 ± 0.05 for wild-type and AQP-4 null astrocytes, respectively. Interestingly, urea permeated astrocytes rapidly, precluding accurate determination of the urea reflection coefficient.
The goal of this study was to test whether AQP-4 is functional as a water channel in cultured astrocytes and, if so, to determine quantitatively its role in plasma membrane water permeability. Osmotic water permeability was measured in highly differentiated primary cultures (without passage) of astrocytes from changes in cell volume in response to osmotic gradients. After testing several experimental approaches to measure astrocyte water permeability, we validated a calcein fluorescence quenching method in which cell volume changes were deduced from total fluorescence of calcein-loaded cells. Because inhibitors of AQP-4 water permeability are not yet available, we compared water permeability in astrocytes from wild-type and AQP-4-deficient mice. AQP-4-deficient mice manifest a mild defect in urinary concentrating ability (18) because of reduced water permeability in inner medullary collecting duct (2). These mice were found subsequently to have reduced cerebral edema after water intoxication and ischemic stroke (20), reduced auditory brainstem evoked potentials (16, 22), and mildly impaired light-evoked retinal potentials (15). These phenotypic differences were thought to arise from reduced water permeability in cells normally expressing AQP-4, including astrocytes in central nervous system, supportive epithelial cells in cochlea, and Muller cells in retina. However, cell-specific AQP-4 expression does not prove physiological significance, because no abnormalities were found in skeletal muscle function (41) or gastric acid secretion (38) in AQP-4-deficient mice, despite the normal expression of AQP-4 in skeletal muscle plasmalemma and basolateral membrane in gastric parietal cells (7). We report here a 7.1 ± 0.3-fold reduced water permeability in AQP-4-deficient compared with control astrocytes, providing direct evidence that AQP-4 is the principal pathway for water movement across the astrocyte plasma membrane and suggesting a molecular basis for reduced cerebral edema in the in vivo studies.
The continuous readout of steady-state cellular calcein fluorescence provided a simple approach to follow cell volume changes in response to osmotic gradients in astrocytes. Although changes in calcein fluorescence were modest (13% change for doubling of cell volume), the signal-to-noise ratio for calcein intensity measurement was adequate to permit water permeability measurements with better than 15% reproducibility (based on standard deviations). Systematic investigation of the mechanism of calcein sensitivity to cell volume showed that neither quasi-confocal optics nor calcein self-quenching could account for the cell volume-dependent calcein fluorescence. Calcein fluorescence changes were insensitive to lens numerical aperture and cytoplasmic calcein concentration. In vitro experiments indicated that calcein quenching by cytoplasm results from quenching by cytosolic proteins rather than ions. Our findings do not agree with the conclusions of Hamann et al. (9) implicating calcein self-quenching in volume-dependent fluorescence changes. Hamann et al. assumed that cytoplasmic calcein concentration was very high and proposed that reduced calcein fluorescence with cell shrinkage was due to enhanced self-quenching. Our present data indicate that this is not the case, because the volume-dependent calcein signal was present at low cytosolic calcein concentrations, independent of cytosolic calcein concentration, and quenching in vitro was rapid but not instantaneous. Near-instantaneous changes in fluorescence would be expected for self-quenching, whereas quenching by cytosol could be rapid or slow, depending on the interaction kinetics. Notwithstanding the relatively small volume-dependent changes in calcein fluorescence, our data support the general utility of the calcein method to measure water permeability even in cells of unusual size and shape where better optical methods may not be suitable.
Our finding that AQP-4 is the principal astrocyte water channel is in agreement with prior studies utilizing indirect methods. In our original report on AQP-4 null mice, water permeability was studied in fractionated whole brain vesicles by a stopped-flow light scattering method (18). Highly water permeable vesicles were absent in specific fractions from a sucrose gradient, suggesting functional AQP-4 water permeability in a subset of membranes from brain. Measurements of spinal cord and brain slice swelling by an optical transmission method localized AQP-4-dependent water permeability to gray matter but not to specific cell type(s) (29). Nicchia et al. (24) measured water permeability in mouse astrocyte cultures by a total internal reflection method and showed that water permeability was mercurial and temperature insensitive. However, it was not possible to draw conclusions regarding the role of AQP-4 because absolute permeabilities were not computed and comparison was not made with cells lacking functional AQP-4. In addition, the remarkably slower kinetics of cell volume changes in Nicchia et al. (osmotic equilibration time ∼9 s) compared with present data (equilibration time ∼1 s) suggest that water permeability was underestimated because of slow solution exchange and/or lower AQP-4 expression in their cell cultures. In addition, water permeability measurement by total internal reflection, which was developed originally by our laboratory (4), is probably not appropriate for use in very flat cells, such as astrocytes, where the height of the cell projections is comparable to the light penetration depth. We speculate that the apparent temperature-independence of water permeability reported by Nicchia et al. was because of slow solution exchange such that the signal changes were limited by solution mixing rather than intrinsic membrane water permeability. The same group recently used RNA interference to reduce AQP-4 expression in astrocyte cultures (25), reporting impaired cell growth, altered morphology, and ∼50% reduction in apparent water permeability. Their conclusion that AQP-4 is important in astrocyte growth and plasticity is not supported by our present findings of unimpaired growth and normal morphology in AQP-4-deficient astrocytes.
In summary, our present data establish a calcein fluorescence quenching method to measure rapid osmotically induced water transport in primary cultures of mouse brain astrocytes. Calcein fluorescence is quenched rapidly by changes in cytosolic protein concentration that accompany changes in cell volume. The calcein quenching method was able to reliably measure osmotic equilibration times of ∼1 s in AQP-4-containing astrocytes from wild-type mice. The 7.1-fold reduction in osmotic water permeability and increased Arrhenius activation in astrocytes from AQP-4 null mice indicate that AQP-4 functions as the principal water channel in cultured astrocytes.
This work was supported by National Institutes of Health Grants DK-35124, HL-73856, EB-00415, HL-59198, and EY-13574, a grant from the Cystic Fibrosis Foundation (to A. S. Verkman), and Russian Foundation Basic Research Grant RFBR 02-04-48071 (to E. Solenov).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society