Aquaporins and the central nervous system 1

Acta Neurochir (Wien) (2004) 146: 955–960 DOI 10.1007/s00701-004-0319-z

Review Article Aquaporins and the central nervous system E. Sulyok1 , Z. Vajda2;3 , T. Do´czi2 , and S. Nielsen3 1

Institute of Health Promotion and Family Care, Faculty of Health Sciences, University of P
ecs, Hungary Department of Neurosurgery, Faculty of Medicine, University of P
ecs, Hungary 3 The Water and Salt Research Center, University of Aarhus, Denmark 2

Published online July 19, 2004 # Springer-Verlag 2004

Structural characteristics and distribution of AQPs in the brain Rapid water transport across biological membranes is ensured by a family of membrane channel proteins, termed ‘‘aquaporins’’ [1]. The basic structure of aquaporins is described by the hourglass model. The functional unit of AQPs is a tetramer consisting of monomers that provide independent water pores. Each monomer comprises of two tandem repeats of three tilted membranespanning helices with amino- and carboxy-terminals located in the cytoplasma. The two tandem repeats contain short loops (hemipores) connecting the second and third helices. Hemipore loops are inserted into the membrane from opposite sides, they overlap each others in the middle of the membrane and form the aqueous pore of the channel protein which is surrounded by the six transmembrane helices [2]. Recent studies on AQP1 to reveal the structural determinants of specific water permeation through water channels have shown that AQP1 includes three topological elements: an extra- and intracellular vestibule that are connected with an extended narrow pore, ˚ -long selectivity filter. The channel wall along the 20 A the selectivity filter is mostly hydrophobic but contains several hydrophilic residues to displace hydration water. The diameter of the narrowest part of the pore, the ˚ which corresponds to constriction region is about 3 A the size of water molecule. Water selectivity and exceptionally high water permeability of the channels is assumed to be due to hydrophobicity of the inner pore

surface, size restriction, electrostatic repulsion and reorientation of water molecule by specific channel residues [3]. Up to now ten individual AQPs, designated as AQP0–AQP9 have been identified and cloned from mammals. Considerable progress has been accomplished in our understanding of the regulation, tissue distribution, cellular=subcellular localization, the structure=function relationship and the clinical significance of each distinct AQP. Two AQPs have been identified in the central nervous system: AQP1 and AQP4. AQP1 expression is confined to choroid plexus and it has been claimed to play a role in cerebrospinal fluid formation. AQP4 is the predominant water channel of the brain and spinal cord, it is widely distributed in cells at the blood-brain and braincerebrospinal fluid barriers. In fact, AQP4 protein is particularly abundant in astrogial cells lining the subarachnoid space and ventricules and in subpopulations of ependymal cells [4]. High-resolution immunogold electron microscopy revealed its polarized distribution in the plasma membrane of astrogial cells, AQP4 is most strongly expressed in membrane domains that are in direct contact with capillaries and pia mater [5]. AQP4 is also present in osmosensory brain regions including the supraoptic nucleus and subfornical organs where its distribution is non-neuronal, glial-specific with less pronounced polarization. This pattern of distribution implies that osmoreception and the control of AVP secretion and dipsogenesis may be related to glial rather than magnocellular neurosecretary cells [6, 7]. Furthermore, AQP4 expression has been detected in the

956

supporting cells in the inner ear, in the retina and optic nerves and also in the Purkinje cell layer in the cerebellum [8, 9]. AQP4 is characterized by mercurial insensivity, high water permeability and intramembranous clustering into orthogonal arrays of particles. This latter is assumed to contribute to the stability of this membrane protein and to provide higher water flux [5].

Regulation of brain-specific AQP4 The regulation of brain-specific AQP4 is not yet fully explored. It has been demonstrated, however, that treatment of cultured rat astrocytes with protein kinase C (PKC) activator phorbol ester caused a rapid, time- and dose-dependent decrease in AQP4 mRNA expression. The inhibition of mRNA levels was not related to their stability or to de novo protein synthesis, consequently the regulation of AQP4 mRNA via PKC activation could be at transcriptional level [10]. Water channel activity of AQP4 has also be been shown to be regulated by phorbol esterdependent protein phosphorylation via PCK pathway as evidenced by the presence of typical consensus sites for phosphorylation in the AQP4 protein and by the marked reduction of AQP4 protein by phorbol diesters [11]. It is of interest that AQP4 expression could be seen during chemically induced (retinoic acid) differentiation of neural stem cell into astrocytes and during the quiescent cell cycle phase of astrocytoma cells [12]. The observation by Zelenina et al. appears to be of particular importance by demonstrating that in cultured cells with features of renal medullary collecting duct AQP4 expression can be downregulated not only by PKC but also by dopamine. Dopaminergic regulation of brain-specific AQP4, therefore, should be considered as a possible mechanisms in the control of brain water metabolism [13].

E. Sulyok et al.

communication [7]. AQP4 knockout mice have been recently generated by targeted gene disruption to analyse the genotype=phenotype relationship and to get better insight into the physiological role of AQP4 [14, 15]. Using this approach it has been demonstrated that AQP4-null mice have normal growth, development and survival rate without apparent neurological abnormalities or defects in osmoregulation. The maximum urinary osmolality induced by water deprivation, however, was slightly decreased and mildly abnormal retinal function and impaired hearing without ultrastructural abnormalities were noted in transgenic mice lacking AQP4 water channel [16, 17]. These subtle, albeit welldefined alterations are consistent with the notion that AQP4 in the basolateral membrane of medullary collecting duct, in the retina and inner ear has a clinically relevant role in transcellular water transport in these particular tissues.

The role of AQP4 in the pathomechanisms of brain edema formation The significant contribution of AQP4 in maintaining brain water balance became more evident when brain edema was induced in AQP4-null mice and the response

Physiological significance of brain AQPs A growing body of evidence suggests that channelmediated water transport is intimately involved in multiple physiological processes operating in the central nervous system. These include brain volume regulation, water flow between various water compartments, cerebrospinal fluid formation, central osmoreception and body fluid homeostasis. Furthermore, AQP4 has been implicated in potassium channel activation and potassium buffering during neuronal activation. Based on circumstantial evidences AQP4 has also been assumed to play a role in volume transmission or parasynaptic

Fig. 1. Immunoblot of membrane enriched fractions from rat brain. The blot was reacted with affinity purified anti-AQP4 antibody against the C-terminal of the protein and revealed bands corresponding to splice variants (32 and 34 kDa, bar). Densitometric analysis revealed an increase of AQP4 immunoreactivity in brain to 164  12% of control levels (*P, 0.05) after 4 h of systemic hyponatremia. (Reprinted from ref. 24 # (2000) Academic Press)

Aquaporins and the central nervous system

pattern of AQP4 was analyzed in wild-type, normal animals subjected to pathologies known to be associated with vasogenic or cytotoxic edema. Manley et al. were the first to demonstrate that AQP4 deficiency protected the brain and reduced edema formation in mice exposed to acute water intoxication and focal ischemic stroke. When compared to their wildtype counterparts the AQP4 knockout mice had less brain water content, better neurological outcome and improved survival. These observations can be regarded as indicating that functioning AQP4 favors development of brain edema when animals are challenged by pathological conditions known to cause edemagenesis in the brain [18]. In support of these findings decreased osmotic swelling was noted in spinal cord slices taken from AQP4 knockout mice as compared to that from wildtype animals [19].

957

Almost simultaneously our group using different experimental model came essentially to the same conclusion. Namely, we found that in response to severe systemic hyponatremia induced by combined administration of dDAVP and hypotonic dextrose solution a rapid increase occurred in the immunoreactivity of astroglial AQP4 protein without significant changes in AQP4 mRNA levels or subcellular distribution of AQP4 protein (Figs. 1, 2). According to our interpretation the hypoosmotic stress-related post-transcriptional AQP4 protein changes may potentially be accounted for by enhanced phosphorylation with subsequent altered conformation and immunogenity of the channel protein [20]. With this contention in line phosphorylated AQP4 has been shown to have reduced water conductivity [11], it appeared relevant to suggest, therefore, that the observed rapid regulation of AQP4 in hyponatremic conditions may provide

Fig. 2. Immunocytochemical and immuno-electronmicroscopical localization of AQP4 in rat cerebellum. (A, B) Cryosections of cerebellar cortex from a control (A) and 4 h hyponatremic (B) animal, showing essentially the same labeling pattern for AQP4. AQP4 labeling is concentrated in glial processes surrounding intracerebral capillaries (arrowheads) but is not associated with neuronal elements. Glial processes in contact with granule cells (G) and Purkinje cells (P) are also prominently labeled (small arrows). Corresponding to the immunoblotting results, AQP4 labeling density in hyponatremic animals is stronger compared with that seen in cerebella from control animals. Magnification: 630. (C, D) Immunogold labeling of AQP4 in Lowicryl HK20 sections of rat cerebellum in control (C) and hyponatremic (D) animals. Immunogold particles are present along perivascular astroglial processes (GP) surrounding the capillary endothelial cells (E). Cytoplasm and intracellular compartments as well as membrane domains apposed to neighbouring glial processes and neuropil (asterisks) are immunonegative. The labeling pattern of control and hyponatremic animals are similar, with no evidence of AQP4 redistribution. (CL) capillary lumen, (N) endothelial cell nucleus, (BM) basement membrane. The bar indicates 0.15 mm. (Reprinted from ref. 24 # (2000) Academic Press)

958

a first-line defense mechanisms to maintain cerebral water balance and to protect brain volume. Focal traumatic brain injury with impaired bloodbrain barrier and vasogenic edema has also been demonstrated to be associated with reduction of AQP4 expression, whereas in diffuse brain injury with intact blood-brain barrier and cytotoxic edema AQP4 expression remained unaltered [21]. In contrast, brain edema related to chemically induced brain injury or brain tumor with disrupted blood-brain barrier and reactive, hypertrophic astrocytes resulted in upregulation of AQP4 mRNA to clear edema fluid and to reestablish the brain osmotic equilibrium [22, 23]. AQP4 and the dystrophin-associated protein complex (DAP) in the brain Dystrophin is a submembranous cytoskeletal protein which is associated with a membrane-spanning glycoprotein complex of DAP. This complex has been pro-

E. Sulyok et al.

posed to connect extracellular matrix components to cytoskeleton. Several members of the DAP complex have been identified including the cytoplasmic dystrobrevin and the syntrophin family, the transmembranous
-dystroglycan and the extracellular, laminin-binding -dystroglycan [24]. Neuronal dystrophin isoform and the related proteins are localized in the astrocyte endfeet [25]. 1-syntrophin component of the DAP complex has PDZ domain which functions to target membrane proteins such as channel-, receptor- and signal proteins to the specific domains of plasma membrane by interacting with their carboxy terminals [26]. On the basis of co-localization of DAPs and brain AQP4 and marked reduction of AQP4 in dystrophin deficient states several studies have been conducted in an attempt to explore the role of DAPs in polarized trafficking, regulated surface expression and membrane anchoring of AQP4 water channel [27]. Frigeri et al. have recently reported decreased AQP4 protein expression in the membrane fraction of brain

Fig. 3. (a–e) Immunofluorescence localization of AQP4 and -syn in brain. (a) In brain from control mice, AQP4 labeling is concentrated in glial processes, forming the glia limitans (arrowheads) and surrounding intracerebral capillaries (arrows), but AQP4 labeling is not associated with neuronal elements or meninges (asterisk). (b) In brain from the dystrophin-null mice, AQP4 labeling is markedly decreased. (c) In brain from G178 control mice, -syn is localized to the pericapillary glial end-feet (open arrows). (d) In brain from dystrophin-null mice, -syn labeling of pericapillary astroglia is markedly reduced. Asterisks in c and d indicate labeling of -syn in neuronal cells. (3250.) (e, f) Time course of hyponatremia-induced brain edema in control and dystrophin-null mice. (e) Average Apparent Diffusion Coefficient (ADC) values in brain were normalized to the mean of the baseline values. The arrow denotes intraperitoneal injection of distilled water and 8-deamino-arginine vasopressin (water intoxication). Note the gradual decline in ADC occurred in both experimental groups, indicating an increase in the size of the intracellular compartment. After 35 minutes, an abrupt, rapid decline in ADC occurred in the control mice (open circles, broken line). After 52.5 minutes, a similar decline was occurred in the dystrophin-null mice (closed squares, solid line). *P < 0.05 (Student’s t-test). Data are presented as mean  SE. (f) Survival profile for control and dystrophin-null mice. (Reprinted from ref. 23 # (2002) National Academy of Sciences)

959

Aquaporins and the central nervous system

homogenate, particularly in the perivascular astrocyte endfoot processes in dystrophin-null (mdx) mice. AQP4 mRNA content remained unaltered and no signs of intracellular compartmentalization was seen. The decrease in AQP4 protein, therefore, has been attributed to reduced stability of DAP complex and to the related premature protein degradation [26]. 1-syntrophin of the DAP complex has been proposed to be a key molecule in polarized distribution of AQP4 as it has been found to co-localize with AQP4 in astrocyte membrane and AQP4 has been proved to be absent in the brain of 1syntrophin knockout mice [27, 28]. By contrast, Neely et al. provided convincing evidence of normal level of total AQP4 expression with inversely polarized subcellular localization in astrocytes in the brain of -syntrophin null-mice. In addition, they observed increased degradation rate and decreased stability of AQP4 when the PDZ-binding motif at the AQP4 C-terminal was deleted. These observations lend further support to the notion that expression, localization and stability of AQP4 protein in astrocyte membrane is syntrophin-dependent [29]. To define more clearly the complex interrelationship between dystrophin, -syntrophin, AQP4 and brain water metabolism we investigated protein levels and subcellular localization of AQP4 and -syntrophin in dystrophin-null mice and their response pattern to water intoxication. Marked reduction was seen in the abundance of both AQP4 and -syntrophin protein expression in the perivascular astrocyte endfeet in animals missing dystrophin confirming that polarized membrane targeting of these proteins dependents on the presence of dystrophin. When animals were subjected to water intoxication lethal brain edema developed in all, intracellular water accumulation as measured with apparent diffusion coefficient of MRI, however, was delayed and the survival prolonged in dystrophin-null mice (Fig. 3) [30]. These findings are in good agreement with previous contention that complete lack, low expression, dysfunction or mislocalization of brain AQP4 appears to be protective against brain edema formation.

Clinical implications Enthusiasm was sparked by the recent progress of brain AQP4 research that seemed to provide new approaches to prevent or treat brain edema. Selective, non-toxic inhibition of brain AQP4 is likely to meet the clinical requirements of brain edema treatment. This therapeutic approach, however, has limitations which

are to be considered as follows: 1. inhibition of AQP4 may induce and activate other, still unidentified water channel proteins to ensure transcellular water flux 2. in the absence of transcellular water transport intercellular routes may be opened 3. water flow across AQP4 is symmetric and bidirectional, therefore, in case of established brain edema water elimination may be delayed 4. in the brain tissue motionally distinct water fractions have been established and only the free, bulky water is available for immediate transport. Water bound to intraand extracellular macromolecules creates microcompartments and serves as a reservoir from which water can be released or stored in it in a regulated manner to meet the actual need of brain volume regulation [31]. The restructuring of brain water is dictated by the macromolecules as A. Szent-Gy€orgyi, the Nobel prize laureate wrote: ‘‘Life is water dancing to the tune of solids’’. As highlighted by the limitations of AQP4-oriented therapeutic approach, the translation of brain AQP4 research into clinical practice is not apparent, rather it remains an ongoing process.

Acknowledgement The work was supported by the Hungarian National Research Foundation (OTKA No T 042956).

References 1. Preston GM, Carrol TP, Guggino WB, Agre P (1992) Appearance of water channels in Xenopus oocytes expressing red cell CHIP28 protein. Science 256: 385–387 2. Engel A, Walz T, Agre P (1994) Structure of the aquaporin water channels. Curr Opin Struct Biol 4: 545–553 3. Kozono D, Yasui M, King LS, Agre P (2002) Aquaporin water channels: atomic structure and molecular dynamics meet clinical medicine. J Clin Invest 109: 1395–1399 4. 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 5. Rash JE, Yı´sumura T, Hudson CS, Agre P, Nielsen S (1998) Direct immunogold labeling of aquaporin-4 in square arrays of astrocyte and ependymocyte plasma membranes in rat brain and spinal cord. Proc Natl Acad Sci USA 95: 11981–11986 6. Wells T (1998) Vesicular osmometers, vasopressin secretion and aquaporin-4: A new mechanism for osmoreception. Mol Cell Endocrinol 136: 103–107 7. Venero JL, Vizuete ML, Machado A, Cano J (2001) Aquaporins in the central nervous system. Progr Neurobiol 63: 321–336 8. Nagelhus EA, Veruki ML, Torp R et al (1998) Aquaporin-4 water channel protein in the rat retina and optic nerve: Polarized expression in M€uller cells and fibrous astrocytes. J Neurosci 18: 2506–2519 9. Takumi Y, Nagelhus EA, Eidet J et al (1998) Select types of supporting cell in the inner ear express aquaporin-4 water channel protein. Eur J Neurosci 10: 3584–3595

960 10. Nakahama K-I, Nagano M, Fujioka A, Shinoda K, Sasaki H (1999) Effect of TPA on aquaporin-4 mRNA expression in cultured rat astrocytes. Glia 25: 240–246 11. Han Z, Wax MB, Patil RV (1998) Regulation of aquaporin-4 water channels by phorbol ester-dependent protein phosphorylation. J Biol Chem 273: 6001–6004 12. Yoneda K, Yamamoto N, Asaik et al (2001) Regulation of aquaporin-4 expression in astrocytes. Mol Brain Res 89: 94–102 13. Zelenina M, Zelenin S, Bondar AA, Bondar BB, Brismar H, Aperia A (2002) Water permeability of aquaporin-4 is decreased by protein kinase C and dopamine. Am J Physiol Renal Physiol 283: F309–F318 14. 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 15. Verkman AS, Yang B, Song Y, Manley GT, Ma T (2000) Role of water channels in fluid transport studied by phenotype analysis of aquaporin knockout mice. Exp Physiol 85S: S233–S241 16. Li J, Verkman AS (2001) Impaired hearing in mice lacking aquaporin-4 water channel. J Biol Chem 276: 31233–31237 17. Li J, Patil RV, Verkman AS (2002) Mildly abnormal retinal function in transgenic mice without M€uller cell aquaporin-4 water channels. Invest Ophthalm Vis Sci 43: 573–579 18. Manley GT, Fujimura M, Ma T et al (2000) Aquaporin-4 deletion in mice reduces brain edema after acute water intoxication and ischemic stroke. Nature Med 6: 159–163 19. Solenon EI, Vetrivel L, Oshio K, Manley GT, Verkman AS (2002) Optical measurement of swelling and water transport in spinal cord slices from aquaporin null mice. J Neurosci Meth 113: 85–90 20. Vajda Zs, Promeneur D, D
oczi T et al (2000) Increased aquaporin-4 immunoreactivity in rat brain in response to systemic hyponatraemia. Biochem Biophys Res Comm 270: 495–503 21. Ke C, Poon WS, Ng HK, Pang JC-S, Chan Y (2001) Heterogeneous responses of aquaporin-4 in oedema formation in a replicated severe traumatic brain injury model in rats. Neurosci Lett 301: 21–24

E. Sulyok et al.: Aquaporins and the central nervous system 22. Vizuete ML, Venero JL, Vargas C, Ilundain AA, Echevarria M, Machado A, Cano J (1999) Differential upregulation of aquaporin-4 mRNA expression in reactive astrocytes after brain injury: potential role in brain edema. Neurobiol Dis 6: 245–258 23. Papadopoulos MC, Saadoun S, Davies DC, Bell BA (2001) Emerging molecular mechanisms of brain tumour oedema. Br J Neurosurg 15: 101–108 24. Frigeri A, Nicchea GP, Repetto S, Bado M, Minetti C, Svelto M (2002) Altered aquaporin-4 expression in human muscular dystrophies: a common feature? FASEB J 16: 1120–1122 25. Khurana TS, Watkins SC, Kunkel LM (1992) The subcellular distribution of chromosome 6-encoded dystrophin-related protein in the brain. J Cell Biol 119: 357–366 26. Frigeri A, Nicchia GP, Nico B et al (2001) Aquaporin-4 deficiency in skeletal muscle and brain of dystrophic mdx mice. FASEB J 15: 90–98 27. Yokota T, Miyagoe Y, Hosaka Y et al (2000) Aquaporin-4 is absent at the sarcolemma and perivascular astrocyte endfeet in 1-syntrophin knockout mice. Proc Japan Acad 76(B): 22–27 28. Inoue M, Wakayama Y, Liu JW, Murahashi M, Shibuya S, Oniki H (2002) Ultrastructural localization of aquaporin 4 and 1-syntrophin in the vascular feet of brain astrocytes. Tohoku J Exp Med 197: 87–93 29. 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 USA 98: 14108–14113 30. Vajda Zs, Pedersen M, F€uchtbauer E-M et al (2002) Delayed onset of brain edema and mislocalization of aquaporin-4 in dystrophin-null transgenic mice. Proc Natl Acad Sci USA 99: 13131–13136 31. Vajda Zs, Ber
enyi E, Bogner P, Repa I, D
oczi T, Sulyok E (1999) Brain adaptation to water loading in rabbits as assessed by NMR relaxometry. Pediatr Res 46: 450–454 Correspondence: Endre Sulyok, County Children’s Hospital, H-7624 P
ecs, POB 76, Hungary. e-mail: [email protected]