Aquaporin Water Channels and Hydrocephalus

Abstract

The aquaporins (AQPs) are a family of water-transporting proteins that are broadly expressed in mammalian cells. Two AQPs in the central nervous system, AQP1 and AQP4, might play a role in hydrocephalus and are thus potential drug targets. AQP1 is expressed in the ventricular-facing membrane of choroid plexus epithelial cells, where it facilitates the secretion of cerebrospinal fluid (CSF). AQP4 is expressed in astrocyte foot processes and ependymal cells lining ventricles, where it appears to facilitate the transport of excess water out of the brain. Altered expression of these AQPs in experimental animal models of hydrocephalus and limited human specimens suggests their involvement in the pathophysiology of hydrocephalus, as do data in knockout mice demonstrating a protective effect of AQP1 deletion and a deleterious effect of AQP4 deletion in hydrocephalus. Though significant questions remain, including the precise contribution of AQP1 to CSF secretion in humans and the mechanisms by which AQP4 facilitates clearance of excess brain water, AQP1 and AQP4 have been proposed as potential drug targets to reduce ventricular enlargement in hydrocephalus.

Introduction

The aquaporins (AQPs) are a family of water channels whose primary function is to facilitate the transport of water across cell membranes in response to osmotic gradients created by active solute transport. The AQPs are small, integral membrane proteins with a monomer size of ∼30 kDa. AQP monomers assemble in membranes as tetramers, with each AQP monomer containing a narrow aqueous pore. Of the 13 mammalian AQPs, a subset called aquaglyceroporins also transport glycerol and possibly other small solutes such as urea. There is a large body of literature on the structure and function of AQPs, including high-resolution structures for several AQPs, and molecular-dynamics simulations of water movement through AQP aqueous pores [reviewed in [1], [2]].

AQPs are broadly expressed in mammalian tissues, including many types of epithelial cells, microvascular endothelia, astrocytes, epidermal cells, adipocytes, and other cells. Data from AQP-knockout mice have demonstrated the involvement of AQPs in the mechanism of urinary concentration, glandular fluid secretion, brain water balance, angiogenesis, cell migration, and other functions [reviewed in [3]]. The glycerol-transporting aquaglyceroporins are involved in cell proliferation, adipocyte metabolism, and epidermal water retention. The utility of AQP-modulating drugs in a wide variety of diseases, including edema, cancer, glaucoma, and obesity for example, has been proposed [4], though there has been limited progress in the identification of compounds that alter AQP function.

Hydrocephalus, from Greek hydro-, meaning “water” and kephalos, meaning “head,” is an abnormal accumulation of cerebrospinal fluid (CSF) in the brain, which is usually due to blockage of CSF outflow in the ventricles or in the subarachnoid space over the brain, impairing drainage into the circulatory system. The treatment of hydrocephalus is surgical, creating a shunt to drain excess fluid. Two AQPs, AQP1 and AQP4, are of potential relevance to hydrocephalus, as they are involved in the generation of CSF and the removal of excess water from the brain, respectively (Fig. 1). Because in hydrocephalus ventricular water is determined from rates of water entry into and removal out of the brain, it is logical to speculate that modulation of these processes, i.e., reducing water entry and/or accelerating water removal, could be beneficial in hydrocephalus. Here, we review evidence for the involvement of AQP1 and AQP4 in hydrocephalus and the possibility of AQP-targeted drugs for the nonsurgical therapy of hydrocephalus.

Fig. 1.

Routes of water movement into and out of the brain: edema formation (green) - water entry into the brain in the two types of brain edema: cytotoxic (through AQP4) and vasogenic edema (through interendothelial spaces) - and clearance of excess brain water (orange) - edema fluid is eliminated by AQP4 through the glia limitans into subarachnoid CSF, through ependyma and subependymal astroglia into ventricular CSF, and through astroglial pericapillary foot processes into the blood.

AQP1 - A Water Channel in Choroid Plexus Epithelium

AQP1, which was originally identified as responsible for high water permeability in erythrocytes, is expressed in many fluid-secreting and -absorbing epithelia in the kidney, gastrointestinal system, eye, brain, and other organs, as well in microvascular endothelia outside of the brain. Major functions of AQP1 include water transport in kidney tubules, endothelial cell migration, tumor angiogenesis, and nociception. Of relevance to CSF production in the brain, AQP1 is expressed in ciliary epithelium in the eye, where it facilitates the secretion of aqueous fluid and regulates intraocular pressure [5]. In the brain, AQP1 is expressed in the ventricular-facing membrane of epithelial cells in the choroid plexus (Fig. 2a), which are responsible for CSF secretion.

Fig. 2.

Role of AQP1 in CSF production and hydrocephalus. a AQP1 immunocytochemistry of mouse choroid plexus. Arrows point to the ventricular-facing surface of choroidal epithelial cells. b Intracranial pressure (ICP) in individual wild-type and AQP1-null (-/-) mice. * p < 0.01. c Ventricle area 5 days after intracisternal injection of kaolin. Adapted from Oshio et al. [6] and Wang et al. [8].

Experimental evidence in mice supports the involvement of AQP1 in CSF production and intracranial pressure (ICP) regulation. Measurements in ex vivo choroid plexus from mice showed that AQP1 gene deletion reduces osmotic water permeability of the ventricular- facing membrane approximately 5-fold [6]. However, whether this reduction in cell membrane water permeability translates into a comparable reduction in transepithelial water permeability is not known. ICP, measured using an intraventricular microneedle, was 9–10 cm H₂O in wild-type mice and approximately 4 cm H₂O in AQP1-null mice (Fig. 2b). CSF production, measured by a dye dilution method in which a second microneedle is introduced into the cisterna magna, was reduced by ∼25% in AQP1-null mice; pressure-dependent CSF outflow, measured from steady-state ICP at different ventricular infusion rates, was not affected by AQP1 deletion.

The relatively minor reduction in CSF production in AQP1-null mice is somewhat surprising and suggests that AQP1-facilitated transcellular water transport by choroid plexus epithelial cells accounts for only part of the total CSF production, with remaining water transport occurring through paracellular and non-AQP1-mediated transcellular routes. There may be extrachoroidal CSF production as well, as data from the older literature showed a relatively minor, approximately 30%, reduction in CSF production following choroid plexus removal. Another potential explanation is compensatory upregulation of prosecretory choroidal solute transporters in the AQP1-knockout mice. The more substantial approximately 2-fold reduction in ICP in AQP1-null mice may be due largely to reduced central venous pressure, and hence sagittal sinus pressure, as AQP1-null mice manifest mild volume depletion which reduced blood pressure because of impaired urinary-concentrating function [7]. Nevertheless, reduced CSF production in AQP1-null mice also contributes to the reduced ICP, as evidenced experimentally by only partial reduction in ICP in mice lacking AQP3, a different water channel expressed in the kidney but not in the choroid plexus, which like AQP1 produces polyuria and reduced central venous pressure.

The possible involvement of AQP1 in hydrocephalus was evaluated in the established kaolin-induced hydrocephalus model in mice [8]. Interestingly, in kaolin-treated wild-type mice, AQP1 cell surface expression was reduced by ∼50% by an endocytic retrieval mechanism, which may represent a compensatory response to reduce CSF secretion. In an earlier study in the congenital hydrocephalic Texas rat, choroidal AQP1 expression was mildly (∼30%) reduced early in life but normalized by postnatal day 26 just before death [9]. In AQP1-deficient mice, ventricular size was reduced compared to wild-type mice both at baseline and following kaolin-induced hydrocephalus [8] (Fig. 2c).

Together, the mouse data suggest but do not prove a significant role for AQP1 in hydrocephalus. Though AQP1 deletion reduced CSF production and ICP, the reduction in CSF production was modest, and the reduction in ICP may be largely explicable by reduced central venous pressure. The reduced ventricular size in the mouse model of hydrocephalus is interesting, though the reduction was seen both at baseline and after kaolin, and may be explicable largely by reduced central venous pressure rather than reduced CSF production. Whether these data in knockout mice can be translated into human hydrocephalus is unclear. There are no reported data on choroidal AQP1 expression in human hydrocephalus. Regarding possible drug therapy of hydrocephalus targeting AQP1, the reduced AQP1 cell surface expression in kaolin-induced hydrocephalus would limit the effectiveness of AQP1 inhibition therapy, as would the relatively minor reduction in CSF production following AQP1 deletion. Nevertheless, though the role of AQP1 in hydrocephalus remains uncertain, there is sufficient suggestive evidence to warrant testing of AQP1 inhibitors, if and when they become available, in human hydrocephalus.

AQP4 - An Astrocyte Water Channel

AQP4 is expressed in glia at the borders between major water compartments and brain parenchyma, including astroglial foot processes at the blood-brain barrier, the glia-limiting membrane at the subarachnoid CSF-brain interface, and ependymal and subependymal astroglia at the ventricular CSF-brain interface. This expression pattern suggests the involvement of AQP4 in water movement into and out of the brain. Interesting and unique features of AQP4 are the presence of two AQP4 isoforms produced by alternative splicing and the supramolecular assembly of AQP4 into orthogonal arrays [10, 11], which may be important in astrocyte migration and in AQP4 polarization to foot processes [12]. In addition to its role in brain water movement, AQP4 is involved in astrocyte migration, by a mechanism that may involve facilitated lamellipodial extensions [13], and in neuroexcitation, by a mechanism that may involve coupling of potassium and water movement [14]. Interestingly, AQP4 is the target in the autoimmune demyelinating disease neuromyelitis optica, in which anti-AQP4 antibodies cause complement- and cell-mediated astrocyte cytotoxicity, resulting in inflammation, demyelination, and neuron loss [15]. Outside of the brain, AQP4 is expressed in astrocytes in the spinal cord and optic nerve, as well as in a subset of epithelial cells in the stomach, kidney, secretory glands, and airways, and in skeletal muscle. The reader is referred to a recent review on the rich and complex biology of AQP4 [16].

There is compelling evidence from experimental animal models for a role of AQP4 in brain water movement. Manley et al. [17] showed reduced brain swelling and improved survival in AQP4-null versus wild-type mice after water intoxication, and reduced hemispheric swelling after focal cerebral ischemia; AQP4-null mice also showed improved survival in a model of bacterial meningitis [18]. Reduced brain swelling was also reported in α-syntrophin-null mice, which model partial AQP4 deletion because of secondary cellular redistribution of AQP4 [19]. According to the Klatzo classification of brain edema, these are primarily models of cytotoxic (cell swelling) edema in which excess water moves from the vasculature into the brain parenchyma through an intact blood-brain barrier. Inhibitors of AQP4 have been proposed to have potential utility in reducing early brain swelling following insults that produce cytotoxic brain edema. In the opposite direction, AQP4 overexpression in transgenic mice worsened brain swelling and outcome in a water intoxication model of cytotoxic brain edema [20], demonstrating that increasing AQP4 expression over its normal level can have functional consequences.

AQP4 may also be involved in the elimination of excess brain water, though, as discussed below, the mechanism is less clear. Following blood-brain barrier disruption, as occurs in brain tumors or abscesses, or in experimental animals by focal freeze injury, water is driven from the vasculature into the brain extracellular space (ECS) in an AQP4-independent manner by hydrostatic forces, which has been called “vasogenic” edema. Excess water in vasogenic edema is cleared primarily through the glial-limiting membrane into the CSF. AQP4-null mice showed remarkably increased brain water gain and ICP compared to wild-type mice with brain tumor, brain abscess, focal cortical freeze injury, and after saline infusion directly into brain parenchyma [21, 22], suggesting that excess brain water in vasogenic edema is eliminated by an AQP4-dependent route. It was proposed that AQP4 activators or upregulators may reduce brain swelling in pathologies associated with vasogenic brain edema.

AQP4 and Hydrocephalus

AQP4 expression and subcellular distribution appear to be regulated in a variety of pathological conditions [16]. Several descriptive expression studies in hydrocephalus provide indirect evidence for a role of AQP4 in hydrocephalus. Though no change in brain AQP4 expression was reported in a rat model of mild hydrocephalus produced by kaolin injection [23], other rat kaolin model studies reported greatly increased AQP4 expression [24] or complex time- and region-dependent alterations in AQP4 expression [25]. Increased AQP4 expression was found in congenital hydrocephalic Texas rats [9], in dogs with idiopathic communicating internal hydrocephalus [26], and in a rat model of inflammatory communicating hydrocephalus in which AQP4 expression correlated with the severity of hydrocephalus [27]. In humans, cortical brain biopsies from patients with chronic hydrocephalus showed increased AQP4 immunoreactivity compared with controls, with some loss of AQP4 polarization in astrocyte end feet [28]. One report found mildly (∼40%) increased AQP4 protein in CSF of congenital communicating hydrocephalus, suggesting ependymal denudation [29]. Most studies therefore report increased AQP4 expression associated with hydrocephalus, which could represent a compensatory response to increase brain water clearance.

A potentially important role of AQP4 in hydrocephalus was demonstrated using the kaolin model in which kaolin injection into the cisterna magna obstructs CSF outflow from the 4th ventricle [30]. Figure 3a, b shows significantly greater ventricular volume 3 and 5 days after kaolin injection in AQP4-null mice compared to wild-type mice. ICP was remarkably greater in AQP4-null mice at 3 days (∼22 vs. 14 mm Hg; Fig. 3c). Brain parenchymal water content increased to a greater extent in the AQP4-null mice at 3 days (Fig. 3d), suggesting backflow of CSF from ventricles into the parenchymal ECS. A survival study showed 66% survival of AQP4-null mice at 5 days compared to 84% of wild-type mice (Fig. 3e). The worse outcome in AQP4-null mice in the kaolin model was proposed to be the consequence of reduced transependymal water clearance, which was quantified using a multicompartment model based on experimental data in wild-type mice. The model predicted the greater severity of hydrocephalus in AQP4-null mice as well as much reduced severity of hydrocephalus if AQP4 expression/function was increased, for example, by drug therapy. These results implicated the involvement of AQP4-mediated transparenchymal CSF absorption in hydrocephalus and suggested the possibility of increasing AQP4 function to reduce ventricular enlargement.

Fig. 3.

Accelerated progression of hydrocephalus in AQP4-null (-/-) mice. a Coronal sections of fixed brains from wild-type and AQP4-null mice 3 and 5 days after intracisternal kaolin injection. b Lateral ventricle volume following kaolin injection. * p < 0.05, ** p < 0.005. c ICP after kaolin injection. ** p < 0.005. d Percent brain water content measured from periventricular tissue (differences not significant). e Survival of kaolin-injected mice over 5 days (33-35 mice/group). Adapted from Bloch et al. [30].

Mechanisms of AQP4-Dependent Brain Fluid Clearance

Though AQP4-dependent water movement into the brain through an intact blood-brain barrier is understandable in terms of osmotic gradient-driven water transport into the brain, the mechanisms responsible for AQP4-dependent fluid (water + solutes) clearance from brain parenchyma remain largely unresolved. Because AQP4 is a water-only channel, solute transport must follow an alternative route. Pressure-driven water transport through brain parenchyma and out across an ependymal interface would generate an opposing osmotic gradient that would oppose water transport, preventing effective AQP4-dependent fluid clearance. In AQP4-null mice, the ECS is mildly (∼20%) expanded compared to wild-type mice [31, 32], which would predict greater fluid clearance in AQP4 deficiency, opposite to the conclusions from the various models of vasogenic brain edema mentioned above. One might speculate that the fine structure of the ECS is modulated by AQP4 expression in astrocytes, though there is no experimental evidence to support this possibility.

Recently, Iliff et al. [33] proposed a “glymphatic” mechanism to clear solutes and macromolecules from the ECS, such as β-amyloid in neurodegenerative diseases and metabolic waste during sleep, as well as to facilitate the distribution of nutrients to brain cells [34]. In contrast to the conventional idea that solute movement in brain ECS is largely diffusive, the glymphatic mechanism postulates vectorial, convective flow of CSF from para-arterial to venous spaces through brain parenchyma in which convective flow entraps interstitial solutes in ECS fluid and enhances their clearance. It was reported that fluorescently labeled albumin movement into brain parenchyma is greatly reduced in AQP4-null mice [33], which was taken as evidence in support of convective fluid transport in brain parenchyma. However, the glymphatic hypothesis has been questioned in terms of its uncertain plausibility and failure to reconcile the predictions of the hypothesis with experimental evidence supporting nonvectorial, diffusive solute transport in brain ECS [35, 36]. In our laboratory, we found slightly increased rather than greatly reduced solute movement in brain ECS in AQP4-null mice, as well as experimental [37] and theoretical [38] evidence against convective parenchymal fluid transport.

Therefore, the mechanisms of impaired brain water clearance in AQP4-deficient mice remain unresolved, and so it is not possible to rule out some structural or developmental alterations in brains of AQP4 mice that could explain the experimental data. AQP4 inhibitors, if and when they become available, would be informative in elucidating mechanisms of AQP4-dependent fluid clearance in the brain, as water and solute clearance can be measured following acute inhibition.

Progress in the Identification of AQP Modulators for Potential Testing in Hydrocephalus

As discussed above, there is a rationale, albeit limited, for application of AQP1 inhibitors to reduce CSF secretion in hydrocephalus. There has been considerable motivation in the identification of AQP1 inhibitors for more compelling indications as diuretics, anti-angiogenic agents, and for the reduction of intraocular pressure in glaucoma. Heavy-metal, sulfhydryl-reactive small molecules such as HgCl₂ inhibit AQP1 by interaction with a cysteine residue near the extracellular surface of the AQP1 aqueous pore, though heavy-metal cysteine-reactive compounds are not candidates for drug development because of their toxicity and lack of selectivity. There have been several potential AQP1-inhibiting compounds reported in the literature, including carbonic anhydrase inhibitors such as acetazolamide, K⁺ channels blockers such as tetraethylammonium, and loop diuretic analogs, as well as compounds identified in small-molecule screens [4]. However, rigorous retesting of proposed AQP1 inhibitors by multiple laboratories did not confirm bona fide inhibition activity of the proposed compounds [39, 40, 41, 42], perhaps because of technical challenges and artifacts in measurement methods and expression systems, including swelling assays in Xenopus oocytes, as discussed [4]. Several putative AQP4 inhibitors have been reported as well, including several chemically unrelated antiepileptics, antimigraine drugs, and carbonic anhydrase inhibitors, though subsequent retesting did not confirm bona fide inhibition [43]. For potential therapy of hydrocephalus, notwithstanding the limited understanding of brain water clearance mechanisms, agents that increase AQP4 expression or function would be needed. The identification of AQP4 water transport activators would seem unlikely, as AQP4 is constitutively active; however, it may be possible to identify selective upregulators of AQP4 expression that might enhance brain water clearance, though effects on neuroexcitation, glial scarring, and cytotoxic brain swelling may preclude their safe use in hydrocephalus.

Summary and Perspective

Evidence in experimental animal models supports a role for AQP1 in CSF production and AQP4 in the clearance of excess brain water, including excess ventricular fluid in hydrocephalus. Whether the data in mouse models translate into human hydrocephalus is unclear. If AQP1 is an important contributor to choroidal CSF production in humans, then it would be appropriate to test small-molecule AQP1 inhibitors in hydrocephalus, if and when they become available. Notwithstanding the caveats mentioned above, the available data also support the possibility of identifying and testing transcriptional upregulators of AQP4 to accelerate removal of excess brain water in hydrocephalus. It may be interesting to study whether AQP1 or AQP4 polymorphisms might predispose to hydrocephalus or modulate its severity; AQP4 polymorphisms have been reported with potential functional significance [44], and possible associations with edema following ischemic stroke [45], intracranial hemorrhage [46], and traumatic brain injury [47]. In terms of basic biology, there remain major questions about the magnitude of AQP1-facilitated CSF production in humans and the mechanisms of AQP4-dependent clearance of excess ventricular and parenchymal brain water.