Water transport between CNS compartments: contributions of aquaporins and cotransporters

Neuroscience 168 (2010) 941–956

REVIEW WATER TRANSPORT BETWEEN CNS COMPARTMENTS: CONTRIBUTIONS OF AQUAPORINS AND COTRANSPORTERS N. MACAULAY* AND T. ZEUTHEN

Monocarboxylate transporter Aquaporin 1 Other membrane proteins Water transport across the blood– brain barrier (BBB) Glucose transporter Na⫹/K⫹/Cl⫺ cotransporter Monocarboxylate transporter Aquaporin 4 Conclusion Acknowledgments References

Nordic Center of Excellence in Water Imbalance-related Disorders (WIRED), Department of Cellular and Molecular Medicine, University of Copenhagen, Blegdamsvej 3, 2200 Copenhagen, Denmark

Abstract—Large water fluxes continuously take place between the different compartments of the brain as well as between the brain parenchyma and the blood or cerebrospinal fluid. This water flux is tightly regulated but may be disturbed under pathological conditions that lead to brain edema formation or hydrocephalus. The molecular pathways by which water molecules cross the cell membranes of the brain are not well-understood, although the discovery of aquaporin 4 (AQP4) in the brain improved our understanding of some of these transport processes, particularly under pathological conditions. In the present review we introduce another family of transport proteins as water transporters, namely the cotransporters and the glucose uniport GLUT1. In direct contrast to the aquaporins, these proteins have an inherent ability to transport water against an osmotic gradient. Some of them may also function as water pores in analogy to the aquaporins. The putative role of cotransport proteins and uniports for the water flux into the glial cells, through the choroid plexus and across the endothelial cells of the blood– brain-barrier will be discussed and compared to the contribution of the aquaporins. © 2010 IBRO. Published by Elsevier Ltd. All rights reserved.

A molecular model of CSF production K⫹/Cl⫺ cotransporter Na⫹/K⫹/Cl⫺ cotransporter

951 952 952 952 952 953 953 953

Brain water is continuously shifted between different compartments and across the blood– brain and CFS-brain interface. Disturbances in this well-regulated water homeostasis may have deleterious effects on brain function and may be fatal in cases where water accumulates in the brain following pathologies such as ischemia, haemorrhage, or brain trauma. It is not well-understood via which molecular pathways the water enters and exits the brain, although the glial aquaporin 4 (AQP4) is clearly an important factor in brain edema formation (Zador et al., 2009). In this review, we describe the water transport properties of a different class of membrane transport proteins and discuss their putative role in the brain water homeostasis. It is now generally accepted that cotransporters and uniports can transport water (Agre et al., 2004; King et al., 2004), although the underlying molecular mechanism is debated. The fundamental physiological issue we want to address is the following: Transport of ions or substrate between two compartments will require an associated water transport in the order of around 165 water molecules per particle in order to achieve isotonicity. In an entirely osmotic model, water transport would take place through aquaporins or via the lipid phase of the membrane. This would require the build-up of an osmotic gradient that may become quite high and lead to unwarranted secondary movements of water. Such an osmotic mechanism is problematic, particularly in a closed system such as the brain where osmotic pressures have to be strictly controlled. We have previously presented extensive data that demonstrate that cotransporters and uniports cotransport significant amounts of water coupled to the substrate by a mechanism within the protein. We will discuss how this cotransport of water mitigates or even removes the requirement for a build-up of osmotic gradients within the brain and its cells.

Key words: cotransporters, cell swelling, aquaporins, water transport, glial cells, choroid plexus. Contents Water transport by cotransport proteins The role of cotransporters and aquaporins in glial cell water homeostasis Glial cell swelling Glutamate transporters Na⫹/K⫹/Cl⫺ cotransporter K⫹/Cl⫺ cotransporter AQP4 and Kir4.1 Pathological glial cell swelling Glutamate, K⫹ and AQP4 Monocarboxylate transporter Cerebrospinal fluid (CSF) production by the choroid plexus

950 950 951

942 944 944 945 946 946 946 947 947 947 948 949 949 949

*Corresponding author. Tel: ⫹45-35327566; fax: ⫹45-35327526. E-mail address: [email protected] (N. MacAulay). Abbreviations: AQP4, aquaporin 4; CSF, cerebrospinal fluid; EAAT, glutamate transporter; GAT, GABA transporter; KCC, K⫹/Cl⫺ cotransporter; MCT, monocarboxylate transporter; NKCC, Na⫹/K⫹/2Cl⫺ cotransporter.

0306-4522/10 $ - see front matter © 2010 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2009.09.016

941

942

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

WATER TRANSPORT BY COTRANSPORT PROTEINS Cotransporters are membrane-spanning transport proteins that couple ion and substrate transport. In mammalian cells, Na⫹ is usually employed as the principal cotransported ion, due to its large inwardly-directed electro-chemical gradient. Through this coupled transport, the Na⫹ can drive the uptake of a substrate against gradients as large as several thousand-fold in case of the glutamate transporters (Danbolt, 2001). We propose that this cotransport may carry a fixed amount of water molecules along with each transported solute, in a manner that will allow for water to be transported against an osmotic gradient. This feature is not found in the aquaporins but is nonetheless physiologically important as well as necessary in certain tissues where the osmotic gradients would favour water transport in the wrong direction, as in the small intestine after a meal or in the retinal pigment epithelium (Pappenheimer, 1998; Hamann et al., 2000, 2003; Zeuthen et al., 1996). Various cotransport proteins have been shown to possess this ability for transport against an osmotic gradient or “uphill water transport”. Examples are the K⫹/Cl⫺ cotransporter, KCC (Zeuthen, 1991a,b, 1994); the Na⫹/K⫹/2Cl⫺ cotransporter, NKCC1 (Hamann et al., 2005); the glucose cotransporter, SGLT1 (Loo et al., 1996; Zeuthen et al., 1997, 2001, 2002, 2006; Meinild et al., 1998; Zeuthen and Zeuthen, 2007); the glial glutamate transporter, EAAT1 (MacAulay et al., 2001); the GABA transporter, GAT-1 (MacAulay et al., 2002b); the monocarboxylate transporter 1, MCT-1 (Hamann et al., 2000, 2003; Zeuthen et al., 1996); the dicarboxylate transporter 1, NaDC-1 (Meinild et al., 2000), for recent reviews see (Loo et al., 2002; MacAulay et al., in press; Zeuthen and MacAulay, 2002). The phenomenon has also been observed for the glucose uniports GLUT1 and GLUT2 (Zeuthen et al., 2007; Zeuthen et al., unpublished observations). The amount of cotransported water is significant, ranging from 35 to 500 water molecules (Table 1) and the water transport proceeds along with the substrate in a manner that is independent of the osmotic gradient. This cotransport of water

has been studied in various preparations, such as native tissue, cultured mammalian cells, and by heterologous expression in Xenopus laevis oocytes. Various techniques have been employed: ion-selective micro-electrodes, fluorescence, and sensitive optical methods for volume measurements. It seems evident, therefore, that secondary active water transport is not due to an artefact arising from the expression system and/or the experimental technique. Although the ability for water transport in cotransport proteins and uniports is generally accepted (Agre et al., 2004; King et al., 2004), the underlying molecular mechanism is debated. Any conventional unstirred layer effect can been ruled out, mainly because the diffusion coefficient in the cytoplasm is too high to result in a significant build-up of osmolarity in the vicinity of the transport protein (Charron et al., 2006; Zeuthen et al., 2002, 2006, 2007; Zeuthen and Zeuthen, 2007; Naftalin, 2008; Zifarelli and Pusch, 2009). The coupling between water and substrate must therefore take place by a mechanism within the protein itself (Zeuthen, 1994; Zeuthen and Stein, 1994; Naftalin, 2008). Secondary active cotransport of water has several characteristics. First, it is initiated instantaneously after the cotransport begins without the lag that would be observed if the water transport was based on the osmotic build-up of solutes in the cytoplasm (Zeuthen et al., 2006). This point is illustrated by data from the GABA transporter expressed in Xenopus laevis oocytes in Fig. 1. Panel A shows that the strict proportionality between the accumulated current associated with the GABA transport and the instant influx of water determined from the swelling of the oocyte. Second, the observed water accumulation is not simply due to an intracellular build-up of osmotic particles but requires a functional turnover of the protein. The GABA transporter, in addition to the GABA translocation, possesses an Li⫹-leak current mode which in the absence of Na⫹ and GABA and in the presence of Li⫹ leads to a large influx of Li⫹ into the voltage-clamped GAT-1-expressing oocyte (MacAulay et al., 2002b; Mager et al., 1996). Importantly, this current, being equal or larger than the GABA-coupled current, does not lead to an instant water accumulation in the same GAT-1 expressing oocytes (Fig. 1B). Third, cotransport of water is independent of external osmotic gradients. This

Table 1. Water transport properties of brain aquaporins, cotransporters and uniporters Protein Substrates KCC4 NKCC1 MCT1 GAT-1 EAAT1 GLUT1 AQP4 AQP1

K⫹/Cl⫺ Na⫹/K⫹/Cl⫺ H⫹/lactate Na⫹/Cl⫺/GABA Na⫹/glutamate Glucose H2O H2O

Unit Lp (10⫺14 cm3 s⫺1) Water molecules transported References NA NA NA 0.7 0.2 0.2b 24 4

500 590a 500 330 425 40 NA NA

Zeuthen, 1991b, 1994 Hamann et al., 2005; Hamann et al., unpublished observations Zeuthen et al., 1996; Hamann et al., 2000, 2003 Loo et al., 1999; MacAulay et al., 2002b MacAulay et al., 2001, 2002a Zeuthen et al., unpublished observations Yang and Verkman, 1997 Zeidel et al., 1992

ND, not determined; NA, not applicable. The coupling ratio for the NKCC1 was determined by comparing two different situations which give the same rate of water transport. In the pigmented epithelium of the cilliary body of the eye, hyperosmolar addition of NaCl to the outside solution did not alter the transport rate for water by the NKCC1. b Calculated from the glucose uptake, the glucose permeability Ps (1.5⫻10⫺6 cm s⫺1), and the water permeability Lp (6.5⫻10⫺5 cm s⫺1) for human GLUT1 expressed in Xenopus oocytes at room temperature. The turnover rate was 151 s⫺1 at room temperature (Simpson et al., 2007). The measurements were performed in analogy to the study of GLUT2 (Zeuthen et al., 2007). a

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

Fig. 1. Water transport properties of GAT-1 expressed in Xenopus laevis oocytes. The GABA transporter was expressed in Xenopus oocytes. The current was recorded by two-electrode voltage clamp and the water flux from the rate of swelling of the oocyte. The GABAinduced current and the associated water transport are compared in panel A. The jagged line represents the volume of the voltage-clamped GAT-1-expressing oocyte when GABA was added to test solution (black box). The straight line in the same panel corresponds to the integrated GABA-induced current and is therefore representative of the accumulated charges and hence the associated GABA transport. The fact that these two lines superimpose is an indication that the water influx is strictly coupled to the GABA transport with a fixed ratio of water molecules per GABA molecule (MacAulay et al., 2002b). (B) Open circles show that the water flux into the oocyte is proportional to the magnitude of the GABA-induced current. The GAT-1 also supports a channel-mediated Li⫹ current, which is equal to or larger than the GABA-coupled current. The Li⫹ current, however, does not lead to any instant water fluxes in the same GAT-1 expressing oocytes. (C) The

943

could be ascertained in experiments where the extracellular osmolarity was increased at the same time as the cotransport was initiated by addition of GABA (Fig. 1C). The cotransport of water was independent of the osmotic gradient across the oocyte membrane. In the absence of GABA the water transport was a simple function of the osmotic gradient across the membrane whereas in the presence of GABA the total water transport was a sum of the cotransported component and the osmotic component. Some cotransporters, in particular the Na⫹-coupled transporters of neurotransmitters or carbohydrates (SGLT1, EAAT1, GAT-1, DCT-1) possess an additional mode of water transport analogous to that of aquaporins, namely passive water transport through a pore in the protein (Loo et al., 1999; Fischbarg et al., 1989, 1990; Leung et al., 2000; MacAulay et al., 2001, 2002a,b; Meinild et al., 2000). This passive water permeability is entirely independent of the secondary active water transport as illustrated for the GAT-1 in Fig 1C. In contrast, the KCC, the NKCC, and the MCT apparently do not have such passive water permeability. Water will be transported through these proteins in response to an osmotic gradient, but the flow of water will be accompanied by cotransport of the other substrates. The ability of KCC to transport water against an osmotic gradient was demonstrated directly in the choroid plexus with ion-selective micro-electrodes and is illustrated in Fig. 2A, B. When the apical surface of the choroid plexus was challenged with an osmotic gradient of 100 mOsm (obtained with either 100 mM mannitol or 50 mM NaCl) the tissue shrank. When the osmotic gradient was created by the addition of 50 mM KCl instead, swelling of the tissue was observed. This swelling was due to the inward transport of K⫹ and Cl⫺ by KCC since 50 mM KCl lead to cell shrinkage after inhibition of KCC with furosemide. It should be emphasized that channel-mediated transport of K⫹ did not lead to significant K⫹-fluxes or to changes in osmotic gradients. This was ascertained by the measurements in which the KCC was inhibited by furosemide (Zeuthen, 1991a,b, 1994). The ability of the NKCC1 to transport water against an osmotic gradient was demonstrated by fluorescence methods in primary cultures from the human corpus cilliare epithelium of the eye. The influx of Cl⫺ initiated an influx of water that could proceed despite an opposing osmotic gradient of more than 50 mOsm (Fig. 2C, D). Similar experiments demonstrated that osmotic gradients implemented by the addition of around 40 mM NaCl or KCl gave rise to no net water transport since the

cotransport of water is independent of any osmotic gradient across the oocyte membrane. The filled circles illustrate the passive water flux (in the complete absence of GABA) out of the oocyte as a function of the external osmotic challenge as indicated on the X-axis. When GABA was added to the same test solution, the water flux was shifted upwards with an amount equivalent to the water transported inward along with the GABA translocation. This component of water transport was constant and independent of the external osmotic gradient, indicating that water passively exits the GAT-1-expressing oocyte following an osmotic challenge, whereas water simultaneously enters the GAT-1-expressing oocyte by means of secondary active water transport (Adapted from MacAulay et al., 2002b).

944

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

Fig. 2. Water transport properties of the KCC and the NKCC1. (A) The KCC couples the flux of one K⫹ ion and one Cl⫺ ion to the flux of 500 water molecules. (B) Addition of 100 mM of mannitol or 50 mM of NaCl to the outer bathing solution caused cell shrinkage. Addition of 50 mM of KCl caused cell swelling. This is a clear example of uphill water transport: The external osmolarity is here higher than the intracellular osmolarity by 100 mOsm; the intracellular concentrations of K⫹ and Cl⫺ changed only a few millimolar during the exposure to KCl. When the KCC was blocked by furosemide (fur) the cell shrank osmotically in response to the addition of 50 mM of KCl. Data from the choroid plexus epithelium of Necturus Maculosus (Zeuthen, 1994). (C) The NKCC1 couples salt and water transport by a mechanism inside the protein. (D) The coupling leads to uphill water transport (i.e. against the direction of the water chemical potential difference). First, Cl⫺ is replaced isosmotically by an inert anion (Cl⫺ free), which causes the cell to shrink and subsequently swell rapidly when Cl⫺ is returned: water moves in. When the solution is made hyperosmolar by the addition of mannitol (⫹50 man) the cell shrinks due to osmosis. In the last episode we test the combined experiment: addition of Cl⫺ and of mannitol (Cl⫺⫹50 man), which causes a rapid swelling of the cell. Clearly the influx of water coupled to the influx of Cl⫺ can proceed in face of the adverse osmotic gradient created by the mannitol (Hamann et al., 2005; Hamann et al., unpublished observations).

osmotic efflux across the epithelium was more or less matched by the cotransported influx of water (Hamann et al., 2005; Hamann et al., unpublished observations).

THE ROLE OF COTRANSPORTERS AND AQUAPORINS IN GLIAL CELL WATER HOMEOSTASIS Glial cell swelling Glial cells are known to swell in response to neuronal activity but the molecular mechanisms underlying this swelling are not well-defined (Østby et al., 2009). Activitydependent glial cell swelling is most likely due to contribution of several mechanisms (Fig 3). In the following section we will distinguish between normo-physiological and

pathophysiological water transport. Under normal conditions, i.e. neuronal activity, one might envision that the osmolarity would initially increase in the extracellular space as K⫹ and glutamate are released. Immediately thereafter the transport processes would initiate the clearing of K⫹ and glutamate. If the water balance was entirely driven by osmotic water transport, one would expect that the initial outwardly-directed osmotic gradient would result in cell shrinkage. This should be followed by cell swelling as the osmolytes would gradually be accumulated in the glial cell. However, the glial cell swelling is instantaneous with no phase of initial shrinkage (Walz and Hinks, 1985). As we will present in the following section, we believe that part of the glial cell swelling is due to the secondary active water transport by different glial cotransporters, such as

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

945

Fig. 3. Water-transporting proteins in the glial cell. AQP4 and Kir4.1 are enriched in the perivascular glial endfeet. The EAATs are concentrated in the perisynaptic area where AQP4 and Kir4.1 are also located. NKCC1 and KCC are found in the glial cells but their area-specific localization has not been determined.

the glutamate cotransporter and the NKCC1. Finally, we will address the glial cell swelling observed during pathophysiological conditions. Glutamate transporters. Glutamate is the most abundant excitatory neurotransmitter in the CNS and is wellknown to be neurotoxic if not swiftly removed from the synaptic cleft after its release from the presynaptic structure (Sattler and Tymianski, 2001). The glutamate is reabsorbed by specialized transport proteins that transport glutamate along with Na⫹ and H⫹ and with countertransport of K⫹. That way, glutamate can be transported against its own gradient due to the electro-chemical driving forces for Na⫹ and K⫹ (Danbolt, 2001). Five isoforms of the Na⫹-dependent glutamate transporter (EAAT1-5) are expressed in the CNS: the glial transporters EAAT1 and EAAT2 are responsible for the majority of the glutamate uptake, whereas the neuronal EAAT3 and the Purkinje cell-specific EAAT4 participate to a slighter degree and the retinal EAAT5 is specific to this CNS structure (for review, see Danbolt, 2001). It is well established that addition of glutamate to glial cell cultures, primary glial cultures, or retinal segments induces glutamate–transporter-dependent glial cell swelling (Bender et al., 1998; Izumi et al., 1996, 1999, 2002; Koyama et al., 2000; Schneider et al., 1992). For example, glutamate concentrations ranging from 50 ␮M to 5 mM give rise to 5%–18% increase in glial cell volume (Schneider et al., 1992). The glial cell swelling is reversible (Izumi et al., 1999; Schneider et al., 1992) and is directly assigned to the transport function of EAATs; transported glutamate analogues induce a swelling similar to that of glutamate and non-transported analogues prevent the swelling (Bender et al., 1998; Izumi et al., 1996, 1999, 2002; Koyama et al., 2000). The swelling appears beneficial for

the cells since neuronal toxicity is prevented only under conditions in which glial cell swelling occurs and the glutamate transporters are active (Izumi et al., 1999). For the EAAT-dependent glial cell swelling to occur, a functional Na⫹/K⫹-ATPase and an intact Na⫹ electro-chemical gradient are required (Bender et al., 1998; Izumi et al., 1999; Koyama et al., 2000; Schneider et al., 1992). Activation of metabotropic glutamate receptors (mGluRs) also leads to cell swelling under physiological conditions (Hansson, 1994; Hansson et al., 1994). The mechanism, however, appears to be distinct from that of the glutamate transporters (Koyama et al., 2000) and not entirely understood (Bender et al., 1998). As described below, the role of mGluRs is better described during hyposmotic and pathophysiological conditions. The glutamate transporter EAAT1 translocates water along with its substrate in a manner that is independent of the osmotic gradient across the membrane (MacAulay et al., 2001). The amount of water translocated along with the glutamate molecule renders the glutamate transportate close to isotonic (Table 1, MacAulay et al., 2001) which ensures that the changes in the osmolarity of the glial cell remain minimal even when large and sustained glutamate loads are cleared from the synaptic space. The role of the glutamate transporter in activity-dependent glial cell swelling is underlined by its predominant expression in glial membranes facing the neuropil as opposed to the perivascular membranes (Chaudhry et al., 1995)—an expression pattern opposite to that of the AQP4 (Nagelhus et al., 1999; Nielsen et al., 1997). Besides the secondary active water transport, the glutamate transporter has in addition a parallel water pathway which is driven entirely by the osmotic gradient, in analogy to that of the aquaporins. The unit water permeability of EAAT1 is around 20-fold smaller than that of AQP1 but due to its

946

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

abundant expression in glial membranes facing the neuropil (Chaudhry et al., 1995; Danbolt, 2001), a large proportion of the collective water permeability of those specific membrane areas may be assigned to the glutamate transporters. Na⫹/K⫹/Cl⫺ cotransporter. The Na⫹/K⫹/Cl⫺ cotransporter carries unidirectional electroneutral transport of Na⫹, K⫹ and 2 Cl⫺. Two isoforms of this transporter have been described, NKCC1 and NKCC2, with NKCC1 being expressed in a variety of cell types and NKCC2 exclusively in the kidney (Chen and Sun, 2005). The NKCC1 isoform has been shown to transport water in a manner strictly coupled to its ion translocation (Fig. 2C, D), whereas the NKCC2 does not support water flux (Hamann et al., 2005; MacAulay et al., in press). In the present review, we will focus on the functional role of NKCC1 expressed in the endothelial cells comprising the blood– brain-barrier (see below) and in the glial cells (Kimelberg and Frangakis, 1985; MacVicar et al., 2002; Tas et al., 1986). In glial cells, NKCC1 has been shown to play an important role in K⫹reuptake following neuronal activity or addition of K⫹ to the extracellular medium of cultured glial cells (Walz and Hinks, 1986; Walz, 1987) with associated NKCC1-dependent glial cell swelling (MacVicar et al., 2002; Walz and Hinks, 1985; Walz and Mukerji, 1988). Interestingly, the NKCC1-dependent water accumulation of the glial cells exceeds that of the ionic accumulation, rendering the osmolarity of the fluid entering the glial cell hyposmolar compared to the medium (Walz and Hinks, 1985). This finding is a clear indication that it is not the initial ion accumulation that subsequently leads to osmotic water accumulation, since a simple osmotic mechanism could never support more water molecules than those accompanying the ions isotonically into the cell. This finding, taken together with the lack of delay between the ionic accumulation and the glial cell swelling (Walz and Hinks, 1985), suggests that the K⫹-dependent cell swelling cannot entirely be due to osmotic water accumulation but indicate that the secondary active water transport by the NKCC1 may be involved. K⫹/Cl⫺ cotransporter. The K⫹/Cl⫺ cotransporter belongs to the same family of cation-coupled cotransporters as NKCC1 (Lauf and Adragna, 2000) and couples the electroneutral cotransport of K⫹ and Cl⫺. The four isoforms of KCC (1– 4) show differential tissue expression patterns but relatively little is known about the cell-specific expression of the different isoforms (MacAulay et al., 2004; Ringel and Plesnila, 2008). The presence of mRNA encoding KCC1, KCC3, and KCC4 has been detected in a C6 glioblastoma cell line (Gagnon et al., 2007). KCC has long been known to be involved in regulatory volume decrease in most cells (Lauf and Adragna, 2000) and has now also been shown to be responsible for maintenance and regulation of the cell volume of cultured astrocytes (Ringel and Plesnila, 2008). This study emphasized the importance of a functional K⫹/Cl⫺ cotransporter for the glial volume to return to normal after astrocytic swelling following a hypotonic challenge. The KCC has been shown to cotransport water, even in the face of large and oppositely directed

osmotic challenges, Fig. 2 (Zeuthen, 1991a,b, 1994). The associated water transport makes this cotransport protein ideal for a quick return of cell volume following cell swelling (MacAulay et al., in press). AQP4 and Kir4.1. AQP4 is the predominant water channel of the CNS, with a characteristic polarized expression in glial cells: Glial endfeet membranes facing the pia and the small capillaries of the brain are enriched with AQP4 (Nagelhus et al., 2004; Nielsen et al., 1997), anchored in these structures by the ␣-syntrophin and dystrophin complex (Neely et al., 2001; Vajda et al., 2002), whereas the perisynaptic membranes show less AQP4expression (Nagelhus et al., 1998, 2004). Primary glial cell cultures appear to loose this polarized arrangement and express AQP4 uniformly along the membrane (Nicchia et al., 2000), providing the entire glial membrane with an increased water permeability. AQP4 has a characteristic co-localization with the inward rectifier K⫹-channel Kir4.1 at the glial endfeet (in contrast to AQP4, Kir4.1 is also abundantly expressed in the perisynaptic membranes), which has lead to speculations of a functional coupling between the two membrane proteins (Nagelhus et al., 1999, 2004). Following neuronal activity, the local concentration of K⫹ is increased in the area surrounding the synapse up to a ceiling level of 12 mM under normophysiological conditions (Heinemann and Lux, 1977; Walz and Hertz, 1983). The surrounding glial cells appear to be the major sinks for K⫹ which enters the glial cell and leads to concomitant cell swelling (MacVicar et al., 2002; Walz and Hinks, 1985; Walz and Mukerji, 1988). The molecular mechanisms behind this K⫹ and water entry, however, are debated. A phenomenon called spatial buffering has been described for glial cells, in which the increased extracellular K⫹ is cleared through the action of the glial perisynaptic Kir4.1 channels, diffusing through the extended glial cell and exiting at the endfeet through the Kir4.1 channels located therein (Gardner-Medwin and Nicholson, 1983; Orkand et al., 1966, reviewed by Kofuji and Newman, 2004). It has been proposed that this Kir4.1-mediated K⫹entry is followed by an osmotic flow of water into the glial cell (Nagelhus et al., 2004). Studies with mice either lacking (Binder et al., 2006) or mislocated (Amiry-Moghaddam et al., 2003b) AQP4 have shown impaired K⫹ clearance following neuronal stimulation, which, however, may be partly due to secondary effects of the enlarged extracellular space in the AQP4 knock-out mice (Yao et al., 2008). This AQP4-dependent effect on K⫹ clearance suggested a direct coupling between the Kir4.1 and AQP4. However, the functional properties of Kir4.1 were not altered in isolated glial and Müller cells of AQP4-null mice, which points to the independent function of these two proteins (RuizEderra et al., 2007; Zhang and Verkman, 2008). Studies on optical nerves, hippocampal slices, and primary cultures of glial cells have shown little or no contribution from Kir4.1 in K⫹-clearance from the extrasynaptic space upon neuronal activity (D’Ambrosio et al., 2002; Meeks and Mennerick, 2007; Ransom et al., 2000). An alternative interpretation has been arrived at in several studies

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

which have demonstrated a prominent role for the Na⫹/K⫹ATPase and the NKCC1 in K⫹ clearance from the extracellular space upon neuronal activity and increased extracellular K⫹-concentration (D’Ambrosio et al., 2002; MacVicar et al., 2002; Ransom et al., 2000; Walz and Hertz, 1982; Walz and Hinks, 1985, 1986). We would therefore question a wide-spread and dominating role of a K⫹ channel in glial reabsorption of K⫹ under normo-physiological conditions: the membrane potential of the glial cells is mainly determined by the equilibrium potential for K⫹ and the predominant presence of the Kir4.1 channels (Djukic et al., 2007; Kucheryavykh et al., 2007; Walz et al., 1984). The Kir4.1 would under basal conditions therefore conduct K⫹ efflux and for the Kir4.1 channel to conduct a significant K⫹ influx, the membrane potential has to be rendered quite a lot more negative than the equilibrium potential for K⫹. Although glial cells may allow for local changes in the equilibrium potential for K⫹ and spatial clamp (Kofuji and Newman, 2004), it is not trivial to see how this would lead to a large Kir4.1-dependent K⫹-accumulation under physiological conditions with the lack of an external voltage clamp to render the membrane potential more hyperpolarized than the equilibrium potential for K⫹. We suggest that the primary role for Kir4.1 is to maintain the membrane potential of the cell, i.e. to act as the voltage clamp of the cell: during nervous activity, glutamate is released to the extracellular space and subsequently transported primarily into the nearby glial cells by the glial glutamate transporters EAAT1 and EAAT2 (Danbolt, 2001). This Na⫹-dependent re-uptake of glutamate leads to a depolarization of the glial cell and therefore reduces the uptake of glutamate by these voltage-dependent transporters. Depolarization of the cells may increase the K⫹efflux through the Kir4.1 channels, which, in turn, will lead to re-polarization of the cell and thereby maintain the glutamate uptake. Reduced glutamate uptake in glial cells lacking the Kir4.1 channel illustrates the importance of these channels and their contribution to the membrane potential (Djukic et al., 2007; Kucheryavykh et al., 2007). Thus, during nervous activity, glutamate is taken up via the glutamate transporters, which would then lead to a membrane depolarization as well as glial cell swelling, as described above. We recently found new evidence for a swelling-induced increase in current flow through Kir4.1 (Soe et al., 2009). By this mechanism, the cell swellinginduced activation of the Kir4.1 channel would lead to a faster return to the more hyperpolarized glial membrane potential which is beneficial for the glutamate clearance via the EAATs. Pathological glial cell swelling Glutamate, K⫹ and AQP4. Glial cells swell in response to several pathophysiological conditions, such as brain trauma and ischemia, and may precede the brain edema formation associated with these pathologies (Kimelberg, 1995). During pathophysiological conditions, the extracellular K⫹ concentration may rise to as much as 60 mM, the extracellular glutamate concentration may increase, the cells may be severely depolarized, and the

947

intracellular osmolarity increased (Davalos et al., 2000; Hossmann et al., 1977; Kimelberg, 1995; Walz et al., 1984). The driving forces for ion-, neurotransmitter- and water transport are therefore altered and the molecular mechanisms behind the K⫹ reabsorption and the glial cell swelling are presumably different from those mentioned above. As the Na⫹/K⫹-ATPase activity is reduced during brain ischemia, the Na⫹ and K⫹ gradients diminish, leading to depolarization of the glial cell membrane and a reduced driving force for the voltage-sensitive Na⫹-coupled transporters (Walz and Mukerji, 1988). The transport activity of the glutamate transporters as well as the NKCC1 may therefore play a relatively minor role in the observed pathophysiological glial cell swelling, and NKCC1 and the Na⫹/K⫹-ATPase will not have the full capacity to re-absorb K⫹ from the extracellular space. It has been reported, however, that high extracellular K⫹ leads to NKCC1-dependent cell swelling in mouse cortical slices, indicating that the NKCC1 is capable of inwardly directed transport at 75 mM extracellular K⫹ (Su et al., 2002). In the absence of an active Na⫹/K⫹-ATPase, the Donnan forces will dictate a re-distribution of ions across the cell membrane, which— aided by the Cl⫺/HCO3⫺ and Na⫹/H⫹ exchangers—leads to an intracellular accumulation of osmolytes (Kimelberg, 1995; Walz and Mukerji, 1988). The resulting osmotic driving forces will lead to water transport into the cell via the passive water transport pathways present in the glial cell membrane. A major player in this type of water transport may be the AQP4 that is abundantly expressed in the glial endfeet but also in the reminder of the glial membranes, around one-tenth of the density observed in the endfeet membrane (Nagelhus et al., 1998; Nielsen et al., 1997). The water permeability of AQP4 is increased by activation of a glial metabotropic glutamate receptor (Gunnarson et al., 2008) in response to an increase in the extracellular glutamate concentration associated with brain ischemia (Davalos et al., 2000). This effect would contribute further to the glial cell swelling. The glutamate transporters possess an inherent capacity for passive water transport analogous to that of an aquaporin. Although the unit water permeability of EAAT1 is smaller than that of AQP4 (Table 1), its high expression level (Danbolt, 2001), may suggest an important role in the glial water accumulation under pathological conditions. The water permeability of EAAT1 is increased in the presence of glutamate and when the cell membrane is depolarized (MacAulay et al., 2002a)— conditions which are both present during the course of ischemia—and would therefore augment the osmotically dependent glial cell swelling. Monocarboxylate transporter. The MCTs mediate H⫹-coupled transport of lactate across glial, neuronal, and endothelial membranes. The MCT1 and MCT4 isoforms are expressed in glial cells whereas the high affinity MCT2 is the primary isoform in neurons (Pierre and Pellerin, 2005). MCT1 has been shown to cotransport water during lactate translocation, 500 molecules per lactate/H⫹, independently of the external osmotic gradient (Hamann et al., 2000, 2003; Zeuthen et al., 1996). During the initial isch-

948

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

emic insult to the brain, the extracellular concentration of H⫹ and lactate increases (Mabe et al., 1983; Paschen et al., 1987; Rehncrona et al., 1981) and is partly cleared from the extracellular space by the glial MCT. H⫹-coupled lactate transport into a human astrocyte-derived cell-line, a C6 glioma cell line, and astrocytes from primary cultures is associated with around 25% increase in glial cell volume (Bender et al., 1997; Lomneth et al., 1990; Staub et al., 1990). In addition, one study found that a depolarization (to ⫺15 mV) of the glial cell, induced by 60 mM K⫹ in the extracellular fluid, increased the cell swelling to almost 100% without affecting the rate of lactate transport and with high K⫹ by itself only contributing to 10%–15% of the cell swelling (Lomneth et al., 1990). Although the synergistic effect was not explained, these factors are all likely to be present in an ischemic setting, and further investigations into this synergy are required. In a study on rat primary astrocyte cultures, the lactate/H⫹-induced glial cell swelling was prevented by the addition of 200 mM mannitol in the extracellular fluid (Bender et al., 1997). A 200 mM osmotic challenge would, in the absence of a secondary active water transporter, lead to a steep and abrupt shrinkage of the glial cell. It appears therefore that MCT, under these experimental conditions, is able to cotransport water against an osmotic gradient of around 200 mOsm energized by the transport of H⫹ and lactate.

CEREBROSPINAL FLUID (CSF) PRODUCTION BY THE CHOROID PLEXUS CSF is continuously produced by the choroid plexus at a rate of about 500 ml per day. The CSF circulates through the ventricular system after which the majority of it drains out of the brain through the arachnoid villiae (for recent reviews, see Redzic and Segal, 2004; Millar et al., 2007; Abbott, 2004; Praetorius, 2007). It is important to emphasize that the rate of CSF formation is rather insensitive to osmotic and hydrostatic pressure changes in the CSF and therefore relatively independent of changes in intracranial pressure and plasma osmolarity (Heisey et al., 1962; Welch, 1966; Oshio et al., 2005). On the other hand, it also shows that the mechanism by which the choroid plexus transports water cannot result from simple osmosis. In order for the choroid plexus to be insensitive to changes in ambient osmotic and hydraulic driving forces, it must have a low passive water permeability, such as the Lp of the rabbit choroid plexus of 0.7⫻10⫺4 cm s⫺1(osm l⫺1)⫺1 (House, 1974). Given a CSF production rate of 60 ␮l cm⫺2 h⫺1 it can be calculated that it would take an osmotic difference of 250 mOsm to generate a water flux equivalent to the observed net secretion (Heisey et al., 1962; Welch, 1966; House, 1974). Under physiological conditions the CSF is only about 5 mOsm hyperosmolar relative to plasma and this driving force is clearly insufficient to explain physiological transport rates by simple osmosis (Cserr, 1971). In fact, the mammalian choroid plexus can transport water against adverse osmotic gradients as large as 60 mOsm (Heisey et al., 1962) which clearly rules out simple osmotic transport as the basic transport mecha-

nism. Under these conditions, the direction of osmotically driven water transport would proceed in the opposite direction. Indeed, if the epithelia had high passive water permeability, the osmotic back-flux would be prohibitively large. In the choroid plexus, the low overall passive water permeability is determined by the properties of the basolateral membrane as the apical membrane has a high water permeability due to its AQP1 expression (Oshio et al., 2005). It is generally accepted that epithelial cells are slightly hyperosmolar by about 5–10 mOsm relative to the surrounding fluids (Curran and Macintosh, 1962). This has been confirmed directly in gall bladder cells by means of ion-selective micro-electrodes and micro-osmometry (Zeuthen, 1981, 1982, 1983). If, by analogy, the epithelial cells of the choroid plexus were hyperosmolar relative to plasma, water transport could proceed from the plasma into the cell across the basolateral membrane by osmosis. The magnitude of the required osmotic gradient would be quite large, though, as the water permeability of the basolateral membrane is low. However, the hyperosmolarity of the cell would prevent water to proceed from the cell and into the CSF by osmosis. If the cell was hyperosmolar, the osmotic driving force would simply be oriented in the wrong direction (Fig. 4). It might be argued that since the CSF is around 5 mOsm hyperosmolar relative to plasma, water could be driven by two small (say 2.5 mOsm) osmotic gradients across the two membranes. But such a scheme would require the two epithelial cell membranes to have unrealistically large water permeabilities, which are contradicted by the experimental findings (for a review, see Zeuthen, 2002). Notably, when CSF is replaced by a hyposmolar test solution, the choroid plexus epithelium continues to transport water at an almost undisturbed rate, which supports the notion that the CSF production is not primarily based on osmotic driving forces (Heisey et al., 1962). To circumvent these issues, it has been suggested that the osmolarity in lateral intercellular spaces or in some subepithelial compartment is different from that of the cell or the external solutions. This is unlikely, however. Detailed studies with ion-selective microelectrodes in the epithelium from gall bladder have shown that the osmolarity in lateral intercellular spaces is isosmotic with the bathing solutions under a variety of experimental conditions (Zeuthen, 1983; Ikonomov et al., 1985), for reviews see (Zeuthen, 2002; Hill, 2008). The trans-cellular route via the leaky junctions (Zonulae occludence) has also been suggested as a major pathway for transepithelial water transport. Careful analysis has shown, however, that the required combination of sufficient water permeability and semipermeability (i.e. permeable to water but not to ions and other osmolytes) is difficult to achieve. First, the actual area occupied by the junctions is a very small percentage of the total epithelial area so it would require quite an open structure to achieve sufficient water permeability of this limited space. This in turn may make it difficult to restrict ion movements sufficiently to achieve semipermeability. Second, the difference in osmolarity across the junctions (in our case between the

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

949

expressed in choroid plexus and are capable of uphill water transport and discuss their role in the CSF production. We will also include a discussion of the possible role of the apical AQP1.

Fig. 4. Simple osmotic mechanisms cannot explain water transport across the choroid plexus epithelium. (A) It is unlikely that water crosses the epithelium via the cellular route by simple osmosis. (B) If the cell is hyperosmolar relative to plasma, water may enter the cell by osmosis but how does it get out again? The osmotic gradient has the wrong direction. For a detailed discussion see text. (C) The problems are even more pronounced in the situation where the choroid plexus epithelium transports against an osmotic gradient; the mammalian choroid plexus has been shown to transport water into the ventricles even when these are perfused with solutions that are hyposmolar relative to plasma (Heisey et al., 1962).

CSF and the apical end of the lateral spaces) might be reduced by the flow of water itself (Finkelstein, 1987; Shachar-Hill and Hill, 2002). If the junctional route is to provide significant rates of water transport, one would have to postulate some kind of non-osmotic transport mechanism (Shachar-Hill and Hill, 2002). We conclude that water transport across the choroid plexus epithelium cannot be explained simply by an osmotic mechanism across the cellular route or the paracellular route. A molecular model of CSF production In order to explain water transport in the choroid plexus, one has to explain how a high rate of water transport can be achieved across an epithelium with a rather low water permeability. The analysis above indicates a need to explain how water can move uphill across one of the cell membranes or across the whole epithelium. In the following we will present some cotransport proteins that are

K⫹/Cl⫺ cotransporter. The KCC is the key protein in our molecular model of water transport in choroid plexus: It can transport water uphill in rates compatible to those maintained by the whole epithelium (Fig. 2A, B). In analogy to other epithelia, KCC is co-localized with the Na⫹/K⫹ATPase (Fig. 5) in the membrane across which water leaves the cell; in the choroid plexus it is the KCC4 isoform (Zeuthen, 1994; Karadsheh et al., 2004; Adragna et al., 2004). As intracellular K⫹ and Cl⫺ ions are accumulated above electro-chemical equilibrium, the direction of the ion transport by the KCC is out of the cell and into the CSF, i.e. identical to that of the water transport of the whole epithelium (Zeuthen, 1996). There is a significant amount of water associated with the KCC-mediated ion transport (Fig 2A, B), 500 water molecules for each pair of K⫹ and Cl⫺ ions. The outwardly directed gradient of KCl will therefore maintain a water flux in the direction of the epithelial flux. Importantly, the energy contained in the K⫹ and Cl⫺ gradients can be converted into uphill transport of water. Given typical intracellular concentration of K⫹ and Cl⫺, the KCC can transport water against osmotic gradients of up to 200 mOsm, fully sufficient to overcome the osmotic gradients discussed above (Zeuthen, 1994). Importantly, the high rate of water transport associated with the KCC suggests that a major fraction of the total transepithelial water flux may be mediated by the KCC. The mechanism and route of water transport across the basolateral membrane are not well understood. The direction of the osmotic difference would allow water to enter by osmosis but the osmotic permeability appears to be quite low as discussed above. KCC3 is expressed in the basolateral membrane (Pearson et al., 2001). The direction of transport of the KCC3 is from the choroid plexus epithelial cell into the plasma, so the protein could play a role for K⫹ absorption across the choroid plexus. It should be noted that in other water-transporting epithelia, KCC is exclusively localized in the same membrane as the Na⫹/K⫹ ATPase. Na⫹/K⫹/Cl⫺ cotransporter. NKCC1 is located in the apical membrane of the choroid plexus (Karadsheh et al., 2004) which resembles the situation in other water-transporting epithelia such as the small intestine (MacLeod and Hamilton, 1990; O’Brien et al., 1993) and the kidney proximal tubule (Chen and Verkman, 1988) which also have NKCC1 in the membrane across which water leaves the cell. The role of this transporter in the choroid plexus is not clear. First, the direction of the driving force for the transport may be quite close to zero under resting conditions: the product of the concentrations [Na⫹]⫻[Cl⫺]2⫻[K⫹] is similar on the two sides of the membrane, so it is hard to predict in which direction the transport takes place. Second, the rate of transport is regulated by phosphorylation in such a way as to be activated in case of cell shrinkage

950

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

Fig. 5. Membrane proteins directly involved in water transport across the choroid plexus epithelium. At the apical membrane the KCC4 and the Na⫹/K⫹ ⫺ATPase collaborate in order to expel water from the cell. The apical membrane also contains AQP1, NKCC1, and MCT1. Less is known about the water-transporting membrane proteins at the basolateral membrane (plasma-facing membrane). There is a KCC3 which has been suggested to participate in K⫹ absorption, and a sporadic distribution of AQP1 (Praetorius, 2007). The epithelium has a low overall water permeability, probably due to the low water permeability of the basolateral membrane.

(Russell, 2000). This would suggest a direct and simple role for the NKCC1 in volume regulation of the epithelial cell but not necessarily in transepithelial water transport. In an epithelial context, the NKCC1 is usually considered to be important in secretory cells such as those encountered in gland acini or airway epithelia where they together with the CFTR Cl⫺ channel located in the opposite membrane ensure water movements into the lumen (Silva et al., 1977). By analogy to the secretory cells, it could be speculated that the NKCC1 is involved in regulation of the water secretion into CSF by being responsible for a component of water movement from the CSF into the blood, i.e. in the direction opposite to the normal CSF secretion. This would comply with the finding that KCC and the NKCC1 are oppositely regulated by phosphorylation (for a review, see Russell, 2000). In other words, a shift in the rate of water transport by the choroid plexus could be achieved by a concomitant change in the phosphorylation of the two proteins KCC and NKCC1. Monocarboxylate transporter. Monocarboxylates such as lactate, pyruvate, ␤-hydroxybutyrate, and acetoacetate are sources of metabolic energy during neonatal periods, fasting, and during high fat diets (Philp et al., 2001; Halestrap and Price, 1999). Lactate transport into and between brain compartments is maintained by various isoforms of the H⫹-coupled monocarboxylate cotransporters. Of the four main subtypes, at least the MCT1, 2, and 4 are found in brain whereas in the choroid plexus epithelium the MCT3 in the basolateral membrane and the MCT1 is found

in the apical membrane (Philp et al., 1998, 2001). The water transport properties of MCT1 have been investigated in the retinal pigment epithelium of the bullfrog eye (Zeuthen et al., 1996), in cultures of human fetal retinal epithelium (Hamann et al., 2000), and in cultures of porcine retinal pigment epithelium (Hamann et al., 2003). The major finding was that the MCT1 cotransports 500 water molecules with each lactate molecule (Table 1). These data suggest that in the presence of lactate or other monocarboxylates in the blood, there will be a transport of lactate across the choroid plexus into the CSF. Importantly, the lactate flux will be accompanied by a water flux of a magnitude that will ensure that the increase in osmolarity caused by the lactate in the CSF and the brain cells will be maintained within reasonable limits. The H⫹ ion will be buffered and will not contribute to an increase in osmolarity of the CSF. The lactate, however, will be subject to metabolic breakdown and ATP production which results in an increase in osmotically active particles. Under normal conditions, mammalian body fluids contain 165 water molecules per osmotic active ion or molecule. The MCT-driven transport of one molecule of lactate will supply 500 water molecules which are enough for about three particles. Aquaporin 1. The role of AQP1 in CSF production is ambiguous. AQP1 is mainly present in the apical membrane facing the CSF rendering this membrane rather water permeable (Oshio et al., 2005). However, the water permeability of the whole epithelium, which comprises the

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

apical and basolateral membranes in series, appears to be rather low (Cserr, 1971). It follows that it is the basolateral membrane that limits the rate of osmotic water movements. Studies of AQP1 ⫺/⫺phenotype have not clarified the role of AQP1 in CSF production. Osmotically induced water fluxes across the apical membrane of the choroid plexus was reduced by 80% in mice without AQP1, yet the rate of CSF production was only reduced by 25% (Oshio et al., 2005). A similar quantitative discrepancy between wildtype and the knock-out animal was observed in other water transporting systems such as the kidney proximal tubule (Hill et al., 2004). These results do not support models based on transepithelial or transmembrane osmotic differences in which water transport is driven entirely by osmosis. An alternative role for AQP1 has been suggested, namely that of an osmosensor (Hill et al., 2004). One might speculate that the AQP1s enable the choroidal cells to sense the osmolarity of the CSF and thereby initiate the appropriate adjustments of transport rates. A sporadic distribution of AQP1 in the basolateral membrane has been reported, albeit the expression level is much lower than at the apical membrane (Praetorius, 2007). Other membrane proteins. In addition to the proteins discussed above, the apical and the basolateral membranes of the choroid plexus contain a variety of electrogenic or electroneutral Na⫹-coupled acid base transporters Na⫹/HCO3⫺, 2Na⫹/HCO3⫺, etc. (Praetorius, 2007). It would be interesting to test if these proteins supported water transport, in analogy to other Na⫹-coupled transporters of carbohydrates or neurotransmitters (Zeuthen and MacAulay, 2002). Experiments with the carbonic anhydrase inhibitor acetazolamide supports this notion: Inhibition of the enzyme reduces CSF production by more than 40% (Vogh et al., 1987). A number of ion channels are located in the apical membrane. These channels participate in the ion homeostasis of the cells but have not been found to play a direct role in water transport.

951

WATER TRANSPORT ACROSS THE BLOOD– BRAIN BARRIER (BBB) The endothelium of the capillaries is the main pathway for nutrients and gasses into and out of the brain (for a recent review, see Abbott et al., 2006). To ensure an efficient supply of all cells, the capillaries have to lie relatively close, typically separated by 50 –100 ␮m. Consequently, the area of endothelial wall that supplies the brain is large; in each cubic centimetre of brain there is about 100 cm2 of endothelial membrane (Paulson et al., 1977). Given the high surface-to-volume ratio (capillary area/brain volume) of the brain, transport of water across the endothelial cell layer becomes significant even if the endothelial barrier has one of the lowest reported water permeabilities per square centimeter (MacAulay et al., in press). If, for example, the osmolarity of the plasma changes by 1%, equivalent to an osmotic gradient of 3 mOsm across the endothelial wall, the brain volume would alter by as much as 2% each hour. Which proteins contribute significantly to the water permeability of the endothelial wall? It is notable that AQP4 is absent from the endothelium although they are highly represented in the glial endfeet adjacent to the endothelium (Amiry-Moghaddam et al., 2003a; Nagelhus et al., 1999; Nielsen et al., 1997). A variety of water-transporting transport proteins are expressed in one or both of the endothelial membranes and in the following section we will discuss the possible role of some of these endothelial transport proteins that may be involved in the water homeostasis of the brain, i.e. the glucose transporter GLUT1, NKCC1, and MCT1. These transporters are present both in the luminal and the cerebral-facing membranes of the endothelium (Fig. 6). As outlined in the introduction, we suggest that these proteins cotransport water along with their substrate. The physiological advantage of the cotransport of water would be the reduction of the build-up of osmotic pressure in the brain in association with the influx of the different substrates. Each molecule transported into the brain will lead

Fig. 6. Water-transporting proteins in the endothelium of the blood– brain barrier. The NKCC1, the GLUT1, and the MCT1 are present in the membrane facing the blood plasma, as well as in the membrane facing the brain. The glial endfeet enwrapping the endothelium are enriched in AQP4 and Kir4.1.

952

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

to an increase in osmolarity. In case of no cotransport of water, an osmotic gradient would build up and water dragged in. This is problematic, because there will be no control of the route through which this osmotic transport will take place, the compartment from which the water will be taken from, or of the osmotic gradient required to move sufficient water. In case of cotransport, a large fraction of the required water transport is taken care of simultaneously and linked to the substrate transport. This would mitigate the need for the build-up of unwarranted and potentially dangerous osmotic gradients. Glucose transporter GLUT1 is located in the endothelium in both the membrane facing the plasma and in the membrane facing the brain parenchyma (Lund-Andersen, 1979). Accordingly, glucose transport into the brain is thought to take place via the endothelial cell, driven by the glucose concentration difference between blood and brain. The water transport properties of GLUT1 and GLUT2 have been studied in details after expression in Xenopus laevis oocytes (Zeuthen et al., 2007; Zeuthen et al., unpublished observations). It was found that water is transported through the GLUT1 by two mechanisms: First, it had a small but well-defined passive osmotic water permeability which was operative both in the absence and presence of glucose. Second, when glucose was added to the outer solution, an inwardly directed flux of glucose was initiated. Under this condition it was found that 40 water molecules were cotransported for each glucose molecule (Zeuthen et al., unpublished observations). The transport of these water molecules was coupled to the glucose transport by a mechanism inside the protein and can be seen as secondary active water transport driven by the glucose transport. The total passive osmotic water permeability provided by the GLUT1 to the BBB is significant. Although the water permeability per protein is small (Table 1), the number of transporters required to supply the brain with glucose is large. The ratio of water to glucose permeability of the GLUT1 expressed in Xenopus laevis oocytes can be estimated to about 6:1 at body temperature (Simpson et al., 2007; Zeuthen et al., unpublished observations). With a PGLUC of around 10⫺4⫻cm s⫺1 (Zeuthen et al., 2007; Zeuthen et al., unpublished observations), the Lp contributed by GLUT1 can be estimated to 0.5⫻10⫺3 cm s⫺1 at 37 °C, in good agreement with the overall Lp measured in humans in vivo of 1.0⫻10⫺3 cm s⫺1 (Paulson et al., 1977). Naⴙ/Kⴙ/Clⴚ cotransporter The NKCC1 is present in both membranes of the endothelium (O’Donnell et al., 2004) where it ensures transport of Na⫹, K⫹ and Cl⫺ between plasma and brain parenchyma. The water transport properties of NKCC1 have been tested in primary cultures from the ciliary epithelium of the human eye where the NKCC1 is present in the luminal membrane, see Fig. 2C, D (Hamann et al., 2005). The water transport by NKCC1 is strictly coupled to its ionic transport and although the density of the protein and its rate of transport are not known quantitatively, its contribution to the water

transport across the BBB may be significant. O’Donnell and coworkers have shown that (i) the activity of NKCC1 is increased during ischemic conditions and (ii) during an experimentally inflicted ischemic insult to the brain, inhibition of NKCC1 led to a reduction in brain edema formation (Foroutan et al., 2005; O’Donnell et al., 2004). Blood levels of vasopressin have been shown to increase during brain ischemia (Barreca et al., 2001) and the presence of this peptide hormone increases the activity of NKCC1 (Katay et al., 1998; O’Donnell et al., 2005), as well as the overall permeability of the BBB to Na⫹, K⫹, and H2O (Nagao, 1998) in a vasopressin receptor (V1aR)-dependent fashion (O’Donnell et al., 2005). Altogether, increased activity of NKCC1, whether vasopressin receptor-mediated or due to ischemia itself, appears to lead to increased water accumulation in the brain. Monocarboxylate transporter Under certain conditions such as prolonged fasting, high fat diets, but also in the neonatal period, monocarboxylates are important sources of energy. The isoform MCT1 is expressed in both membranes of the endothelial wall (Philp et al., 2001; Halestrap and Price, 1999). Given a high concentration of monocarboxylate (lactate) in the blood, influx into the brain is ensured. Due to the abovementioned water transport capabilities of MCT1, water will be cotransported along with the lactate. Unfortunately, it is not possible to estimate the contribution of the protein to the overall water permeability of the BBB because the level of expression of MCT1 in the endothelial wall is not known. Nevertheless, the 500 water molecules accompanying the monocarboxylate transported by the MCT1 will be sufficient to supply the water required for isosmolarity in regard to the monocarboxylate molecule alone (as the associated H⫹ ion will be buffered) and partly to the increased number of osmotic particles arising through its metabolic breakdown. Aquaporin 4 AQP4 appears not to be expressed by the endothelial cells of the blood– brain-barrier (Amiry-Moghaddam et al., 2003a; Nagelhus et al., 1999; Nielsen et al., 1997) and will therefore not contribute to the water transport across the endothelium per se. The lack of an apparent phenotype of the AQP4 knock-out mice also suggested that the role of AQP4 under normo-physiological conditions may be limited (Ma et al., 1997). However, during pathological conditions leading to the formation of brain edema, AQP4 is an important player. Knock-out of AQP4 or disruption of its polarized expression pattern mitigated the brain water accumulation associated with brain ischemia, water intoxication, and hyponatremia (Amiry-Moghaddam et al., 2003a, 2004; Manley et al., 2000; Yang et al., 2008), suggesting that during pathophysiological conditions, AQP4 is a main entrance route for water from the plasma and into the brain. In the basal state of ␣-Syntrophin -/- mice, in which AQP4 has lost its polarized expression at the endfeet, the glial endfeet appeared swollen (Amiry-Moghaddam et al., 2003a), which suggested that AQP4 may form an exit

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

route for water from brain to plasma under basal physiological conditions. In addition, osmotherapy of mice with experimentally inflicted cerebral-edema appeared to take effect through the action of the perivascular pool of AQP4, which allowed water to exit from the brain parenchyma (Zeynalov et al., 2008).

CONCLUSION The water homeostasis of the brain and the molecular mechanisms involved in the associated water transport are poorly understood. It appears that many different membrane transport proteins are involved in the intricate regulation of the water content of the entire brain as well as of the sub-compartments of the brain. We propose that the ability of the cotransporters to translocate water along with their substrates contributes significantly to the level of regulation of the water homeostasis as there will be no requirement for large osmotic gradients to drive water transport. The majority of the cotransporters transport a near-isotonic amount of water along with their substrates and in this way the osmolarity of the brain cells would not be disturbed every time ions, nutrients, and neurotransmitters were shifted between the different compartments. During pathophysiological conditions, where the osmotic gradients are disturbed, the passive water permeability of the cells (mainly in the form of AQP4 in the glial cells) will be an important pathway for water intrusion. Acknowledgments—We thank Nordic Center of Excellence in Water Imbalance-related Disorders (WIRED), The Danish Medical Research Council, the Lundbeck Foundation, The Novo Nordic Foundation and E. Danielsen’s Foundation. Thanks to Svend Christoffersen for the artwork.

REFERENCES Abbott NJ (2004) Evidence for bulk flow of brain interstitial fluid: significance for physiology and pathology. Neurochem Int 45:545– 552. Abbott NJ, Ronnback L, Hansson E (2006) Astrocyte-endothelial interactions at the blood-brain barrier. Nat Rev Neurosci 7:41–53. Adragna NC, Di Fulvio M, Lauf PK (2004) Regulation of K-Cl cotransport: from function to genes. J Membr Biol 201:109 –137. Agre P, Nielsen S, Ottersen OP (2004) Towards a molecular understanding of water homeostasis in the brain. Neuroscience 12:849 – 850. Amiry-Moghaddam M, Otsuka T, Hurn PD, Traystman RJ, Haug FM, Froehner SC, Adams ME, Neely JD, Agre P, Ottersen OP, Bhardwaj A (2003a) An alpha-syntrophin-dependent pool of AQP4 in astroglial end-feet confers bidirectional water flow between blood and brain. Proc Natl Acad Sci U S A 100:2106 –2111. Amiry-Moghaddam M, Williamson A, Palomba M, Eid T, de Lanerolle NC, Nagelhus EA, Adams ME, Froehner SC, Agre P, Ottersen OP (2003b) Delayed K⫹ clearance associated with aquaporin-4 mislocalization: phenotypic defects in brains of alpha-syntrophin-null mice. Proc Natl Acad Sci U S A 100:13615–13620. Amiry-Moghaddam M, Xue R, Haug FM, Neely JD, Bhardwaj A, Agre P, Adams ME, Froehner SC, Mori S, Ottersen OP (2004) Alphasyntrophin deletion removes the perivascular but not endothelial pool of aquaporin-4 at the blood-brain barrier and delays the development of brain edema in an experimental model of acute hyponatremia. FASEB J 18:542–544.

953

Barreca T, Gandolfo C, Corsini G, Del SM, Cataldi A, Rolandi E, Franceschini R (2001) Evaluation of the secretory pattern of plasma arginine vasopressin in stroke patients. Cerebrovasc Dis 11:113–118. Bender AS, Schousboe A, Reichelt W, Norenberg MD (1998) Ionic mechanisms in glutamate-induced astrocyte swelling: role of K⫹ influx. J Neurosci Res 52:307–321. Bender AS, Young LP, Norenberg MD (1997) Effect of lactic acid on L-glutamate uptake in cultured astrocytes: mechanistic considerations. Brain Res 750:59 – 66. Binder DK, Yao X, Zador Z, Sick TJ, Verkman AS, Manley GT (2006) Increased seizure duration and slowed potassium kinetics in mice lacking aquaporin-4 water channels. Glia 53:631– 636. Charron FM, Blanchard MG, Lapointe J-Y (2006) Intracellular hypertonicity is responsible for water flux associated with Na⫹/glucose cotransport. Biophys J 90:3546 –3554. Chaudhry FA, Lehre KP, van Lookeren CM, Ottersen OP, Danbolt NC, Storm-Mathisen J (1995) Glutamate transporters in glial plasma membranes: highly differentiated localizations revealed by quantitative ultrastructural immunocytochemistry. Neuron 15:711–720. Chen H, Sun D (2005) The role of Na-K-Cl cotransporter in cerebral ischemia. Neurol Res 27:280 –286. Chen P-Y, Verkman AS (1988) Sodium-dependent chloride transport in basolateral membrane vesicles isolated from rabbit proximal tubule. Biochemistry 27:655– 660. Cserr HF (1971) Physiology of the choroid plexus. Physiol Rev 51:273–311. Curran PF, Macintosh JR (1962) A model system for biological water transport. Nature 193:347–348. D’Ambrosio R, Gordon DS, Winn HR (2002) Differential role of KIR channel and Na⫹/K⫹-pump in the regulation of extracellular K⫹ in rat hippocampus. J Neurophysiol 87:87–102. Danbolt NC (2001) Glutamate uptake. Prog Neurobiol 65:1–105. Davalos A, Shuaib A, Wahlgren NG (2000) Neurotransmitters and pathophysiology of stroke: evidence for the release of glutamate and other transmitters/mediators in animals and humans. J Stroke Cerebrovasc Dis 9:2– 8. Djukic B, Casper KB, Philpot BD, Chin LS, McCarthy KD (2007) Conditional knock-out of Kir4.1 leads to glial membrane depolarization, inhibition of potassium and glutamate uptake, and enhanced shortterm synaptic potentiation. J Neurosci 27:11354 –11365. Finkelstein A (1987) Water movement through lipid bilayers, pores, and plasma membranes, Vol. 228. New York: John Wiley & Sons. Fischbarg J, Kuang K, Hirsch J, Lecuona S, Rogozinski L, Silverstein SC, Loike J (1989) Evidence that the glucose transporter serves as a water channel in J774 macrophages. Proc Natl Acad Sci U S A 86:8397– 8401. Fischbarg J, Kuang K, Vera JC, Arant S, Silverstein SC, Loike J, Rosen OM (1990) Glucose transporters serve as water channels. Proc Natl Acad Sci U S A 87:3244 –3247. Foroutan S, Brillault J, Forbush B, O’Donnell ME (2005) Moderate-tosevere ischemic conditions increase activity and phosphorylation of the cerebral microvascular endothelial cell Na⫹-K⫹-Cl⫺ cotransporter. Am J Physiol Cell Physiol 289:C1492–C1501. Gagnon KBE, Adragna NC, Fyffe REW, Lauf PK (2007) Characterization of glial cell K-Cl cotransport. Cell Physiol Biochem 20: 121–130. Gardner-Medwin AR, Nicholson C (1983) Changes of extracellular potassium activity induced by electric current through brain tissue in the rat. J Physiol 335:375–392. Gunnarson E, Zelenina M, Axehult G, Song Y, Bondar A, Krieger P, Brismar H, Zelenin S, Aperia A (2008) Identification of a molecular target for glutamate regulation of astrocyte water permeability. Glia 56:587–596. Halestrap AP, Price NT (1999) The proton-linked monocarboxylate transporter (MCT) family: structure, function and regulation. Biochem J 343:281–299.

954

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

Hamann S, la Cour M, Lui GM, Bundgaard M, Zeuthen T (2000) Transport of protons and lactate in cultured human fetal retinal pigment epithelial cells. Pflugers Arch 440:84 –92. Hamann S, Kiilgaard JF, La Cour M, Prause JU, Zeuthen T (2003) Cotransport of H⫹, lactate and H2O in porcine retinal pigment epithelial cells. Exp Eye Res 76:1–12. Hamann S, Herrera-Perez JJ, Bundgaard M, varez-Leefmans FJ, Zeuthen T (2005) Water permeability of Na⫹-K⫹-2Cl⫺ cotransporters in mammalian epithelial cells. J Physiol 568:123–135. Hansson E (1994) Metabotropic glutamate receptor activation induces astroglial swelling. J Biol Chem 269:21955–21961. Hansson E, Johansson BB, Westergren I, Ronnback L (1994) Glutamate-induced swelling of single astroglial cells in primary culture. Neuroscience 63:1057–1066. Heinemann U, Lux HD (1977) Ceiling of stimulus induced rises in extracellular potassium concentration in the cerebral cortex of cat. Brain Res 120:231–249. Heisey SR, Held D, Pappenheimer JR (1962) Bulk flow and diffusion in the cerebrospinal fluid system of the goat. Am J Physiol 203:775–781. Hill AE (2008) Fluid transport: a guide for the perplexed. J Membr Biol 223:1–11. Hill AE, Shachar-Hill B, Shachar-Hill Y (2004) What are aquaporins for? J Membr Biol 197:1–32. Hossmann KA, Sakaki S, Zimmerman V (1977) Cation activities in reversible ischemia of the cat brain. Stroke 8:77– 81. House CR (1974) Water transport in cells and tissues. London: Edward Arnold. Ikonomov O, Simon MX, Frömter E (1985) Electrophysiological studies on lateral intercellular spaces of necturus gallbladder epithelium. Pflugers Arch 403:301–307. Izumi Y, Kirby CO, Benz AM, Olney JW, Zorumski CF (1999) Muller cell swelling, glutamate uptake, and excitotoxic neurodegeneration in the isolated rat retina. Glia 25:379 –389. Izumi Y, Kirby-Sharkey CO, Benz AM, Mennerick S, Labruyere J, Price MT, Olney JW, Zorumski CF (1996) Swelling of Muller cells induced by AP3 and glutamate transport substrates in rat retina. Glia 17:285–293. Izumi Y, Shimamoto K, Benz AM, Hammerman SB, Olney JW, Zorumski CF (2002) Glutamate transporters and retinal excitotoxicity. Glia 39:58 – 68. Karadsheh MF, Byun N, Mount DB, Delpire E (2004) Localization of the KCC4 potassium-chloride cotransporter in the nervous system. Neuroscience 123:381–391. Katay L, Latzkovits L, Fonagy A, Janka Z, Lajtha A (1998) Effects of arginine vasopressin and atriopeptin on chloride uptake in cultured astroglia. Neurochem Res 23:831– 836. Kimelberg HK (1995) Current concepts of brain edema. Review of laboratory investigations. J Neurosurg 83:1051–1059. Kimelberg HK, Frangakis MV (1985) Furosemide- and bumetanidesensitive ion transport and volume control in primary astrocyte cultures from rat brain. Brain Res 361:125–134. King LS, Kozono D, Agre P (2004) From structure to disease: the evolving tale of aquaporin biology. Nat Rev Mol Cell Biol 5:687– 698. Kofuji P, Newman EA (2004) Potassium buffering in the central nervous system. Neuroscience 129:1045–1056. Koyama Y, Ishibashi T, Okamoto T, Matsuda T, Hashimoto H, Baba A (2000) Transient treatments with L-glutamate and threo-beta-hydroxyaspartate induce swelling of rat cultured astrocytes. Neurochem Int 36:167–173. Kucheryavykh YV, Kucheryavykh LY, Nichols CG, Maldonado HM, Baksi K, Reichenbach A, Skatchkov SN, Eaton MJ (2007) Downregulation of Kir4.1 inward rectifying potassium channel subunits by RNAi impairs potassium transfer and glutamate uptake by cultured cortical astrocytes. Glia 55:274 –281. Lauf PK, Adragna NC (2000) K-Cl cotransport: properties and molecular mechanism. Cell Physiol Biochem 10:341–354.

Leung DW, Loo DF, Hirayama BA, Zeuthen T, Wright EM (2000) Urea transport by cotransporters. J Physiol 528(2):251–257. Lomneth R, Medrano S, Gruenstein EI (1990) The role of transmembrane pH gradients in the lactic acid induced swelling of astrocytes. Brain Res 523:69 –77. Loo DDF, Wright EM, Zeuthen T (2002) Water pumps. J Physiol 542:53– 60. Loo DDF, Zeuthen T, Chandy G, Wright EM (1996) Cotransport of water by the Na⫹/glucose cotransporter. Proc Natl Acad Sci U S A 93:13367–13370. Loo DF, Hirayama BA, Meinild A-K, Chandy G, Zeuthen T, Wright E (1999) Passive water and ion transport by cotransporters. J Physiol 518:195–202. Lund-Andersen H (1979) Transport of glucose from blood to brain. Physiol Rev 57:305–352. Ma T, Yang B, Gillespie A, Carlson EJ, Epstein CJ, Verkman AS (1997) Generation and phenotype of a transgenic knockout mouse lacking the mercurial-insensitive water channel aquaporin-4. J Clin Invest 100:957–962. Mabe H, Blomqvist P, Siesjo BK (1983) Intracellular pH in the brain following transient ischemia. J Cereb Blood Flow Metab 3:109 – 114. MacAulay N, Gether U, Klaerke DA, Zeuthen T (2001) Water transport by the Na⫹-coupled glutamate cotransporter. J Physiol 530:367– 378. MacAulay N, Gether U, Klaerke DA, Zeuthen T (2002a) Passive water and urea permeability of a human Na⫹-glutamate cotransporter expressed in Xenopus oocytes. J Physiol 542:817– 828. MacAulay N, Hamann S, Zeuthen T (2004) Water transport in the brain: role of cotransporters. Neuroscience 129:1031–1044. MacAulay N, Hamann S, Zeuthen T (in press) Chloride transporters as water pumps: elements in a new model of epithelial water transport. In: Physiology and pathology of chloride transporters and channels in the nervous system: from molecules to diseases (Alvarez-Leefmans FJ, Delpire E, eds). San Diego: Elsevier. MacAulay N, Zeuthen T, Gether U (2002b) Conformational basis for the Li⫹-induced leak current in the rat gamma-aminobutyric acid (GABA) transporter-1. J Physiol 544:447– 458. MacLeod RJ, Hamilton JR (1990) Regulatory volume increase in mammalian jejunal villus cells is due to bumetanide-sensitive NaKCl2 cotransport. Am J Physiol 258:G665–G674. MacVicar BA, Feighan D, Brown A, Ransom B (2002) Intrinsic optical signals in the rat optic nerve: role for K⫹ uptake via NKCC1 and swelling of astrocytes. Glia 37:114 –123. Mager S, Kleinberger-Doron N, Keshet GI, Davidson N, Kanner BI, Lester HA (1996) Ion binding and permeation at the GABA transporter GAT1. J Neurosci 16:5405–5414. Manley GT, Fujimura M, Ma T, Noshita N, Filiz F, Bollen AW, Chan P, Verkman AS (2000) Aquaporin-4 deletion in mice reduces brain edema after acute water intoxication and ischemic stroke. Nat Med 6:159 –163. Meeks JP, Mennerick S (2007) Astrocyte membrane responses and potassium accumulation during neuronal activity. Hippocampus 17:1100 –1108. Meinild A-K, Klaerke DA, Loo DDF, Wright EM, Zeuthen T (1998) The human Na⫹/glucose cotransporter is a molecular water pump. J Physiol 508:15–21. Meinild A-K, Loo DFF, Pajor A, Zeuthen T, Wright EM (2000) Water transport by the renal Na⫹/dicarboxylate cotransporter. Am J Physiol 278:F777–F783. Millar ID, Bruce JI, Brown PD (2007) Ion channel diversity, channel expression and function in the choroid plexuses. Cerebrospinal Fluid Res 4:8. Naftalin RJ (2008) Osmotic water transport with glucose in GLUT2 and SGLT. Biophys J 94:3912–3923. Nagao S (1998) Vasopressin and blood-brain barrier. No To Shinkei 50:809 – 815.

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956 Nagelhus EA, Horio Y, Inanobe A, Fujita A, Haug FM, Nielsen S, Kurachi Y, Ottersen OP (1999) Immunogold evidence suggests that coupling of K⫹ siphoning and water transport in rat retinal Muller cells is mediated by a coenrichment of Kir4.1 and AQP4 in specific membrane domains. Glia 26:47–54. Nagelhus EA, Mathiisen TM, Ottersen OP (2004) Aquaporin-4 in the central nervous system: cellular and subcellular distribution and coexpression with Kir4.1. Neuroscience 129:905–913. Nagelhus EA, Veruki ML, Torp R, Haug FM, Laake JH, Nielsen S, Agre P, Ottersen OP (1998) Aquaporin-4 water channel protein in the rat retina and optic nerve: polarized expression in Muller cells and fibrous astrocytes. J Neurosci 18:2506 –2519. Neely JD, Amiry-Moghaddam M, Ottersen OP, Froehner SC, Agre P, Adams ME (2001) Syntrophin-dependent expression and localization of aquaporin-4 water channel protein. Proc Natl Acad Sci U S A 98:14108 –14113. Nicchia GP, Frigeri A, Liuzzi GM, Santacroce MP, Nico B, Procino G, Quondamatteo F, Herken R, Roncali L, Svelto M (2000) Aquaporin-4-containing astrocytes sustain a temperature- and mercuryinsensitive swelling in vitro. Glia 31:29 –38. Nielsen S, Nagelhus EA, Amiry-Moghaddam M, Bourque C, Agre P, Ottersen OP (1997) Specialized membrane domains for water transport in glial cells: high-resolution immunogold cytochemistry of aquaporin-4 in rat brain. J Neurosci 17:171–180. O’Brien JA, Walters RJ, Valverde MA, Sepúlveda FV (1993) Regulatory volume increase after hypertonicity- or vasoactive-intestinalpeptide-induced cell-volume decrease in small-intestinal crypts is dependent on Na⫹-K⫹-2Cl⫺ cotransport. Pflügers Arch 423:67–73. O’Donnell ME, Duong V, Suvatne J, Foroutan S, Johnson DM (2005) Arginine vasopressin stimulation of cerebral microvascular endothelial cell Na-K-Cl cotransporter activity is V1 receptor and [Ca] dependent. Am J Physiol Cell Physiol 289:C283–C292. O’Donnell ME, Tran L, Lam TI, Liu XB, Anderson SE (2004) Bumetanide inhibition of the blood-brain barrier Na-K-Cl cotransporter reduces edema formation in the rat middle cerebral artery occlusion model of stroke. J Cereb Blood Flow Metab 24:1046 –1056. Orkand RK, Nicholls JG, Kuffler SW (1966) Effect of nerve impulses on the membrane potential of glial cells in the central nervous system of amphibia. J Neurophysiol 29:788 – 806. Oshio K, Watanabe H, Song Y, Verkman AS, Manley GT (2005) Reduced cerebrospinal fluid production and intracranial pressure in mice lacking choroid plexus water channel Aquaporin-1. FASEB J 19:76 –78. Østby I, Øyehaug L, Einevoll GT, Nagelhus EA, Plahte E, Zeuthen T, Lloyd CM, Ottersen OP, Omholt SW (2009) Astrocytic mechanisms explaining neural-activity-induced shrinkage of extraneuronal space. PLoS Comput Biol 5:e1000272. Pappenheimer JR (1998) Scaling of dimensions of small intestines in non-ruminant eutherian mammals and its significance for absorptive mechanisms. Comp Biochem Physiol 121:45–58. Paschen W, Djuricic B, Mies G, Schmidt-Kastner R, Linn F (1987) Lactate and pH in the brain: association and dissociation in different pathophysiological states. J Neurochem 48:154 –159. Paulson OB, Hertz MM, Bolwig TG, Lassen NA (1977) Filtration and diffusion of water across the blood-brain barrier. Microvasc Res 13:113–124. Pearson MM, Lu J, Mount DB, Delpire E (2001) Localization of the K⫹-Cl⫺ cotransporter, KCC3, in the central and peripheral nervous systems: expression in the choroid plexus, large neurons and white matter tracts. Neuroscience 103:481– 491. Philp NJ, Yoon H, Grollman EF (1998) Monocarboxylate transporter MCT1 is located in the apical membrane and MCT3 in the basal membrane of rat RPE. Am J Physiol 274:R1824 –R1828. Philp NJ, Yoon H, Lombardi L (2001) Mouse MCT3 gene is expressed preferentially in retinal pigment and choroid plexus epithelia. Am J Physiol Cell Physiol 280:C1319 –C1326.

955

Pierre K, Pellerin L (2005) Monocarboxylate transporters in the central nervous system: distribution, regulation and function. J Neurochem 94:1–14. Praetorius J (2007) Water and solute secretion by the choroid plexus. Pflügers Arch 454:1–18. Ransom CB, Ransom BR, Sontheimer H (2000) Activity-dependent extracellular K⫹ accumulation in rat optic nerve: the role of glial and axonal Na⫹ pumps. J Physiol 522:427– 442. Redzic ZB, Segal MB (2004) The structure of the choroid plexus and the physiology of the choroid plexus epithelium. Adv Drug Deliv Rev 56:1695–1716. Rehncrona S, Rosen I, Siesjo BK (1981) Brain lactic acidosis and ischemic cell damage: 1. Biochemistry and neurophysiology. J Cereb Blood Flow Metab 1:297–311. Ringel F, Plesnila N (2008) Expression and functional role of potassiumchloride cotransporters (KCC) in astrocytes and C6 glioma cells. Neurosci Lett 442:219 –223. Ruiz-Ederra J, Zhang H, Verkman AS (2007) Evidence against functional interaction between aquaporin-4 water channels and Kir4.1 potassium channels in retinal Muller cells. J Biol Chem 282:21866 –21872. Russell JM (2000) Sodium-potassium-chloride cotransport. Physiol Rev 80:211–276. Sattler R, Tymianski M (2001) Molecular mechanisms of glutamate receptor-mediated excitotoxic neuronal cell death. Mol Neurobiol 24:107–129. Schneider GH, Baethmann A, Kempski O (1992) Mechanisms of glial swelling induced by glutamate. Can J Physiol Pharmacol 70:S334 – S343. Shachar-Hill B, Hill AE (2002) Paracellular fluid transport by epithelia. Int Rev Cytol 215:319 –350. Silva P, Stoff J, Field M, Fine L, Forrest JN, Epstein FH (1977) Mechanism of active chloride secretion by shark rectal gland: role of Na-K-ATPase in chloride transport. Am J Physiol 233:F298 – F306. Simpson IA, Carruthers A, Vannucci SJ (2007) Supply and demand in cerebral energy metabolism: the role of nutrient transporters. J Cereb Blood Flow Metab 27:1766 –1791. Soe R, MacAulay M, Klaerke D (2009) Modulation of Kir4.1 and Kir5.1 channels by small changes in cell volume. Neurosci Lett 457: 80 – 84. Staub F, Baethmann A, Peters J, Weigt H, Kempski O (1990) Effects of lactacidosis on glial cell volume and viability. J Cereb Blood Flow Metab 10:866 – 876. Su G, Kintner DB, Sun D (2002) Contribution of Na⫹-K⫹-Cl⫺ cotransporter to high-[K⫹]-induced swelling and EAA release in astrocytes. Am J Physiol Cell Physiol 282:C1136 –C1146. Tas PW, Massa PT, Koschel K (1986) Preliminary characterization of a Na⫹,K⫹,Cl⫺ cotransport activity in cultured human astrocytes. Neurosci Lett 70:369 –373. Vajda Z, Pedersen M, Fuchtbauer EM, Wertz K, Stodkilde-Jorgensen H, Sulyok E, Doczi T, Neely JD, Agre P, Frokiaer J, Nielsen S (2002) Delayed onset of brain edema and mislocalization of aquaporin-4 in dystrophin-null transgenic mice. Proc Natl Acad Sci U S A 99:13131–13136. Vogh BP, Godman DR, Maren TH (1987) Effect of AlCl3 and other acids on cerebrospinal fluid production: a correction. J Pharmacol Exp Ther 243:35–39. Walz W (1987) Swelling and potassium uptake in cultured astrocytes. Can J Physiol Pharmacol 65:1051–1057. Walz W, Hertz L (1982) Ouabain-sensitive and ouabain-resistant net uptake of potassium into astrocytes and neurons in primary cultures. J Neurochem 39:70 –77. Walz W, Hertz L (1983) Functional interactions between neurons and astrocytes. II. Potassium homeostasis at the cellular level. Prog Neurobiol 20:133–183. Walz W, Hinks EC (1985) Carrier-mediated KCl accumulation accompanied by water movements is involved in the control of physiological K⫹ levels by astrocytes. Brain Res 343:44 –51.

956

N. MacAulay and T. Zeuthen / Neuroscience 168 (2010) 941–956

Walz W, Hinks EC (1986) A transmembrane sodium cycle in astrocytes. Brain Res 368:226 –232. Walz W, Mukerji S (1988) KCl movements during potassium-induced cytotoxic swelling of cultured astrocytes. Exp Neurol 99:17–29. Walz W, Wuttke W, Hertz L (1984) Astrocytes in primary cultures: membrane potential characteristics reveal exclusive potassium conductance and potassium accumulator properties. Brain Res 292:367–374. Welch K (1966) Volume flow across choroidal ependyma of the rabbit. Am J Physiol 210:232–236. Yang B, Verkman AS (1997) Water and glycerol permeabilities of aquaporin 1-5 and MIP determined quantitatively by expression of epitope-tagged constructs in Xenopus oocytes. J Biol Chem 272: 16140 –16146. Yang B, Zador Z, Verkman AS (2008) Glial cell aquaporin-4 overexpression in transgenic mice accelerates cytotoxic brain swelling. J Biol Chem 283:15280 –15286. Yao X, Hrabetova S, Nicholson C, Manley GT (2008) Aquaporin-4deficient mice have increased extracellular space without tortuosity change. J Neurosci 28:5460 –5464. Zador Z, Stiver S, Wang V, Manley GT (2009) Role of aquaporin-4 in cerebral edema and stroke. Handb Exp Pharmacol 159 –170. Zeidel ML, Ambudkar SV, Smith BL, Agre P (1992) Reconstitution of functional water channels in liposomes containing purified red cell CHIP28 protein. Biochemistry 31:7436 –7440. Zeuthen T (1981) Isotonic transport and intracellular osmolarity in the Necturus gallbladder epithelium. In: Water transport across epithelia. Alfred Benzon symposium (HH, Ussing N, Bindslev NA, Lassen O, eds), pp Sten-Knudsen 313–331. Copenhagen: Munksgaard. Zeuthen T (1982) Relations between intracellular ion activities and extracellular osmolarity in Necturus gallbladder epithelium. J Membr Biol 66:109 –121. Zeuthen T (1983) Ion activities in the lateral intercellular spaces of gallbladder epithelium transporting at low external osmolarities. J Membr Biol 76:113–122. Zeuthen T (1991a) Secondary active transport of water across ventricular cell membrane of choroid plexus epithelium of necturus maculosus. J Physiol 444:153–173. Zeuthen T (1991b) Water permeability of ventricular cell membrane in choroid plexus epithelium from necturus maculosus. J Physiol 444:133–151.

Zeuthen T (1994) Cotransport of K⫹, Cl⫺ and H2O by membrane proteins from choroid plexus epithelium of Necturus maculosus. J Physiol 478:203–219. Zeuthen T (1996) Molecular mechanisms of water transport. New York: Chapman & Hall. Zeuthen T (2002) General models for water transport across leaky epithelia. Int Rev Cytol 215:285–317. Zeuthen T, Belhage B, Zeuthen E (2006) Water transport by Na⫹ -coupled cotransporters of glucose (SGLT1) and of iodide (NIS). The dependence of substrate size studied at high resolution. J Physiol 570:485– 499. Zeuthen T, Hamann S, La Cour M (1996) Cotransport of H⫹, lactate and H2O by membrane proteins in retinal pigment epithelium of bullfrog. J Physiol 497:3–17. Zeuthen T, MacAulay N (2002) Cotransporters as molecular water pumps. Int Rev Cytol 215:259 –284. Zeuthen T, Meinild A-K, Klaerke DA, Loo DDF, Wright EM, Belhage B, Litman T (1997) Water transport by the Na⫹/glucose cotransporter under isotonic conditions. Biol Cell 89:307–312. Zeuthen T, Meinild A-K, Loo DDF, Wright EM, Klaerke DA (2001) Isotonic transport by the Na⫹-glucose cotransporter SGLT1. J Physiol 53:631– 644. Zeuthen T, Stein WD (1994) Cotransport of salt and water in membrane proteins: membrane proteins as osmotic engines. J Membr Biol 137:179 –195. Zeuthen T, Zeuthen E (2007) The mechanism of water transport in Na⫹ -coupled glucose transporters expressed in Xenopus oocytes. Biophys J 93:1413–1416. Zeuthen T, Zeuthen E, Klaerke DA (2002) Mobility of ions, sugar, and water in the cytoplasm of Xenopus oocytes expressing Na⫹ coupled sugar transporters (SGLT1). J Physiol 542:71– 87. Zeuthen T, Zeuthen E, MacAulay N (2007) Water transport by GLUT2 expressed in Xenopus laevis oocytes. J Physiol 579:345–361. Zeynalov E, Chen CH, Froehner SC, Adams ME, Ottersen OP, AmiryMoghaddam M, Bhardwaj A (2008) The perivascular pool of aquaporin-4 mediates the effect of osmotherapy in postischemic cerebral edema. Crit Care Med 36:2634 –2640. Zhang H, Verkman AS (2008) Aquaporin-4 independent Kir4.1 K⫹ channel function in brain glial cells. Mol Cell Neurosci 37:1–10. Zifarelli G, Pusch M (2009) Conversion of the 2 Cl⫺/1 H⫹ antiporter ClC-5 in a NO3-/H⫹ antiporter by a single point mutation. EMBO J 28:175–182.

(Accepted 8 September 2009) (Available online 15 September 2009)