- Email: [email protected]
Altered cellular localization of aquaporin-1 in experimental hydrocephalus in mice and reduced ventriculomegaly in aquaporin-1 deficiency
Molecular and Cellular Neuroscience 46 (2011) 318–324
Contents lists available at ScienceDirect
Molecular and Cellular Neuroscience j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y m c n e
Altered cellular localization of aquaporin-1 in experimental hydrocephalus in mice and reduced ventriculomegaly in aquaporin-1 deficiency☆ Dongwei Wang a, Marko Nykanen a, Nan Yang a, David Winlaw a, Kathryn North a,b, A.S. Verkman c, Brian Kenneth Owler a,b,⁎ a b c
Institute for Neuroscience and Muscle Research Children's Hospital at Westmead, Sydney, Australia Discipline of Child Health and Paediatrics, Children's Hospital at Westmead, University of Sydney, Sydney, Australia Departments of Medicine and Physiology, University of California, San Francisco, San Francisco, California, USA
a r t i c l e
i n f o
Article history: Received 21 July 2010 Revised 29 September 2010 Accepted 21 October 2010 Available online 30 October 2010 Keywords: Hydrocephalus Aquaporins Choroid plexus Cerebrospinal fluid
a b s t r a c t Hydrocephalus is a pathological accumulation of cerebrospinal fluid (CSF) in the cerebral ventricles that constitutes a significant cause of neurological morbidity and mortality. Surgical treatment involving shunt placement is associated with a high failure rate and complications due to infection, motivating the development of alternative, non-surgical therapies. Here, we investigated the role in hydrocephalus of water channel aquaporin-1 (AQP1), which is expressed at the apical membrane of choroid plexus epithelium and is believed to facilitate CSF production. AQP1 expression and subcellular localization were studied in a kaolininduced hydrocephalus model in mice and the effect AQP1 deficiency on the severity of hydrocephalus was determined. While total choroidal AQP1 protein was not significantly altered in hydrocephalus, ~ 50% of AQP1 protein was redistributed from the apical membrane to intracellular vesicles. We found that the ventricular size in AQP1-deficient mice was smaller than in wild-type mice, both at baseline and following hydrocephalus. The reduced plasma membrane AQP1 localization following kaolin-induced hydrocephalus, which involves endocytosis, may be a compensatory mechanism to reduce CSF secretion. The reduced ventricular size in AQP1-deficient mice following kaolin-induced hydrocephalus suggests AQP1 inhibition or down-regulation as a potential adjunctive treatment for hydrocephalus. © 2010 Elsevier Inc. All rights reserved.
Introduction Hydrocephalus, a condition in which an excessive volume of cerebrospinal fluid (CSF) accumulates in the cerebral ventricles, remains an important cause of neurological deficit, delayed development and death. Treatment is generally surgical, requiring for most patients the insertion of a CSF shunt. However, shunts have a failure rate of N30% within the first 12 months, requiring revision. Improved or alternative treatments for hydrocephalus would be desirable. The water channel aquaporin-1 (AQP1) provides one molecular mechanism for CSF secretion and is thus a potential target for nonsurgical management of hydrocephalus. AQP1 is an integral membrane protein expressed in the apical membrane of choroid plexus epithelium. Studies in mice lacking AQP1 (AQP1KO mice) suggest involvement of AQP1 in CSF secretion. AQP1KO mice have ~ 2-fold reduced baseline intracranial pressure compared to wild-type (WT) mice, ~25% lower CSF production, and 5-fold lower osmotically driven water permeability (Oshio et al., 2005). In that study AQP1KO mice ☆ Grant information: (NIH DK35124). ⁎ Corresponding author. Suite 19, Children's Hospital Medical Centre, Hainsworth Street, Westmead NSW 2145, Australia. Fax: + 61 2 98060060. E-mail address: [email protected] (B.K. Owler). 1044-7431/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.mcn.2010.10.003
had improved clinical outcome compared to wild-type mice in a brain injury model; however, it was not possible to determine whether the improved outcome was the result of reduced CSF secretion and/or reduced central venous pressure (and consequent reduced intracranial pressure) related to renal effects of AQP1 deletion. Little is known about the role of AQP1 in the development of hydrocephalus. Hydrocephalus has not been studied in AQP1KO mice, and there is conflicting information about AQP1 expression in hydrocephalus. AQP1 expression was unchanged in a rat model of hydrocephalus (Mao et al., 2006), reduced in a single human case of choroid plexus hyperplasia (Smith et al., 2007), and elevated in choroid plexus papilloma associated with communicative hydrocephalus (Longatti et al., 2006). We report here an analysis of AQP1 expression and cellular localization in a kaolin-induced mouse model of hydrocephalus, and evaluation, using AQP1KO mice, of whether AQP1 deficiency alters the severity of experimental hydrocephalus. Results Kaolin-injected WT mice develop hydrocephalus By 5 days post injection, compared to saline-injected mice (Fig. 1A), dilated ventricles were observed in mice injected with kaolin (Fig. 1B
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
319
Fig. 1. Characterization of kaolin-induced hydrocephalus mouse model in wild-type (WT) mice. (A, B) H&E staining of coronal sections of mouse brain from saline (A) and kaolin (B)-injected mice at 5 dpi. (C, D) H&E staining of sagittal sections of mouse brain from kaolin-injected mice. Panel D is the high magnification image of the boxed region in panel C. Note that the fourth ventricle is filled with kaolin suspension (arrowhead). (E) Ventricular area in untreated, saline-injected and kaolin-injected mice at 3 and 5 dpi (*pb 0.05, **pb 0.001, Kruskal–Wallis Test). The error bar represents SEM. The number in the bar indicates the number of mice used per group. Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice.
and C), as well as inflammatory cells in the subarachnoid space in mice injected with kaolin (Fig. 1D), as expected in this inflammatory model of hydrocephalus. We measured the size of the area of the lateral ventricles in untreated, saline injected control mice and hydrocephalic WT mice at 3 and 5 dpi. There was no statistical significance between untreated and saline injected mice at 3 or 5 days. The mean ventricular area in the kaolin-injected mice was significantly larger compared to saline injected control mice at both time points (Fig. 1E). The degree of the increase in ventricular size is similar to mice reported with the mild hydrocephalus phenotypes in adult mouse models (Bloch et al., 2006) as opposed to the more severe neonatal models of hydrocephalus.
Total AQP1 expression is not altered in hydrocephalic mice We investigated whether AQP1 expression in WT mice was altered following experimental hydrocephalus. Western blot showed one band at 28 kDa, corresponding to non-glycosylated AQP1, and a diffuse band at ~ 38 kDa corresponding to glycosylated AQP1 (Fig. 2A). Quantitative analysis of total protein from choroid plexus showed no difference in AQP1 protein expression between hydrocephalic and control WT mice at 5 dpi (Fig. 2B) as well as at 3 dpi (data not shown).
Fig. 2. Quantification of AQP1 in saline-injected controls and kaolin-injected hydrocephalic WT mice at 5 days post-injection by Western blotting. (A) Representative Western blot of AQP1 and GAPDH of experimental and control mice. (B) The normalized ratio of the densitometry readings of AQP1 to GAPDH in saline-injected mice (n = 7) and kaolin-injected mice (n = 10). The error bar represents SEM. Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice.
320
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
Altered subcellular location of AQP1 in hydrocephalus mice We hypothesized that the pathological accumulation of CSF in hydrocephalic WT mice may be associated with reduced AQP1 protein in apical membrane of choroid plexus epithelium. Immunohistochemistry (IHC) and immunofluorescence (IF) were done on the mice randomly selected from the cohort of mice used for ventricle area measurement. As seen in Fig. 3A–E, AQP1 immunoreactivity was largely absent in the cytoplasm but present on the apical membrane of the epithelial cells in control mice. However, AQP1 immunoreactivity was seen both in the cytoplasm and on the apical membrane of epithelial cells in 3 out of 6 of randomly picked hydrocephalic mice at 3 dpi and 7 out of 10 hydrocephalic mice at 5 dpi (Fig. 3F–K) A measurement of mean gray value (MGV) of AQP1 immunofluorescence signals was performed on the above 6 saline-injected control mice and 10 hydrocephalic mice at 5 dpi. There was an increase in the fluorescence signals in the cytoplasm and basolateral membrane in the cells of hydrocephalic mice compared to salineinjected control mice (p b 0.05) (Fig. 3L). In total MGV of the choroidal cells, 91% was from the apical membrane in the saline-injected control mice, while 78% was from the apical membrane in the kaolin-injected hydrocephalic mice (Fig. 3M). Fig. 3O and P show AQP1 immunofluorescence in cells in an endosomal pattern, with occasional coloca-
lization with lysosome-associated membrane marker Lamp2, particularly in the region adjacent to the basolateral membrane. To further investigate the apparent mislocalization of AQP1 in experimental hydrocephalus, immuno-EM was done on 3 hydrocephalic WT mice at 3 dpi, 3 hydrocephalic WT mice at 5 dpi and 4 control mice at 3 dpi. AQP1-positive gold particles were numerous along the microvilli of the apical membrane on the choroidal epithelial cells in both control and hydrocephalic mice (Fig. 4A–D). However, in control mice, apart from the occasional presence of gold particles in mitochondria, no positive staining was found within the cell cytoplasm or along the basolateral membrane. In hydrocephalic mice, AQP1-positive gold particles were observed within vacuoles and vesicles (Fig. 4E(i)), within the enlarged intercellular space (Fig. 4E (ii)) and in the basolateral membrane (Fig. 4E(iii)). The number of gold particles along microvilli and inside the cytoplasm of choroidal cells in the electron-micrograph images from hydrocephalic and control mice was measured (Table 1). Data from hydrocephalic mice at 3 and 5 dpi were combined for analysis. Gold labeling within the cytoplasm was significantly increased in hydrocephalic mice compared to control mice (p = 0.03). There was a trend that the number of gold particles along the microvilli and the microvillus area was reduced in hydrocephalic mice but the difference between hydrocephalic mice and control mice did not reach significance
Fig. 3. AQP1 localization in choroid plexus epithelial cell of saline-injected control and kaolin-injected hydrocephalic WT mice at 5 dpi. (A, F) IHC images from frozen sections (7 μm) were stained with rabbit AQP1 as primary antibody followed by incubation of secondary antibody and standard ABC technique. (B–E, G–K) Confocal images of choroid plexus epithelial cell of control (B–E) and hydrocephalic (G–K) mice. Panel K is the high magnification image of the boxed region in panel J. AQP1 signals were detected by anti-rabbit IgG conjugated to Alexa 488. Positive staining for AQP1 is on the apical surface of the epithelial cells of the saline-injected mice (A–E). In contrast, in hydrocephalic mice, positive staining is also found inside the cytoplasm (F–K). (L) Intracellular AQP1 immunoreactivity demonstrating by mean gray values by Image J analysis. (M) Percentage of apical AQP1 immunoreactivity relative to total AQP1 immunoreactivity. The error bar represents SEM. The number in the bar indicates the number of mice used per group (*p b 0.05. Mann– Whitney Test U test). Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice. (N–P) Double immunofluorescence staining with anti-AQP1 (green) and anti-Lamp2 (red) on choroidal cells of control (N(i) and N(ii)) and hydrocephalic (O(i)–(iii) and P) mice. Panels N (ii), O (iii) and P are overlaid images. Images showing Lamp2 red fluorescence corresponding to panels N(ii) and P and AQP1 green fluorescence corresponding to panel P are not shown. Panel P is the high magnification image to boxed region in panel N. Colocalization of positive signals of AQP1 and Lamp2 is shown in yellow (arrowheads).
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
321
Fig. 4. Anti-AQP1 labeling of choroidal epithelium in saline-injected control and kaolin-injected hydrocephalic WT mice. (A–D) Micrograph of microvilli from control and hydrocephalic mice. Gold particles are on the surface of the microvilli in control mice (A and C) and hydrocephalic mice (B and D). The microvilli in hydrocephalic mice are swollen. (E) Low magnification image of choroidal epithelium in hydrocephalic mice. Note that the tight junction (white arrow) is still intact. (E(i)–(iii)) Higher magnification electron micrographs corresponding to the boxes in panel G show gold particles (black arrows) associated with membrane-like structure in a vacuole in the cytoplasm (i), in the intercellular space (ii) and in the nearby area of the basal membrane of the epithelia (iii). Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice. Scale bar in (i)–(iii) = 100 nm. Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice.
(p = 0.055). There was no difference in the density of AQP1 positive signals along microvilli of the apical membrane in control and kaolininjected mice. The data suggest that the reduction of the total amount of
AQP1 signals on the microvilli may result from a reduction of area of microvilli in kaolin-injected mice. Reduced ventricular size in kaolin-injected AQP1 KO mice
Table 1 Quantification of gold particles on apical membranes and microvilli and in the cytoplasm. Gold particle in Cytoplasm Mean Control (n = 4) HC (n = 6) p value
SEM
4.6 1.0 14.6 3.3 0.033
Mv area
Gold particles on Mv Mean
Density of gold particle on Mv
SEM
Mean
SEM
Mean
473 133 228 45 0.055
8.2 4.7 0.055
1.2 0.8
78.76 3.52 68.25 22.26 0.088
Mv: microvilli; Con: control mice; HC: hydrocephalic mice.
SEM
The reduction in apical membrane-associated choroidal AQP1 in hydrocephalic mice suggests the possibility of a compensatory mechanism to reduce CSF production in hydrocephalus. Therefore, the potential of AQP1 as a therapeutic target for the treatment hydrocephalus may be limited. In order to clarify the potential utility of AQP1 we studied whether the absence of AQP1 in the AQP1 KO mice influenced ventricular size and the formation of hydrocephalus. As shown in Fig. 5A, saline injected AQP1 KO mice had significantly smaller ventricular areas compared to saline injected WT mice (p = 0.001). After cisternal saline injection AQP1 KO mice demonstrated
322
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
Fig. 5. Characterization of kaolin-induced hydrocephalus mouse model in WT and AQP1 KO mice. (A) Ventricular area in untreated, saline-injected and kaolin-injected WT and AQP1 KO mice at 3 and 5 dpi. (B) Body weight loss percentage over 5 days after cisternal kaolin injection in WT and AQP1 KO mice. WT-Con: saline-injected WT mice (n = 15); WT-HC, kaolin-injected hydrocephalic WT mice (n = 24); KO-Con: saline-injected KO mice (n = 12); KO-HC: kaolin-injected KO mice (n = 17).
no change in ventricular size at either 3 or 5 dpi compared to untreated AQP1 KO mice. AQP1 KO mice that had undergone cisternal kaolin injection had significantly larger lateral ventricles compared to saline-injected AQP1 KO controls at both 3 and 5 dpi (p b 0.05). However, when compared with the ventricular sizes of kaolin-injected WT mice, we found that the ventricular sizes in hydrocephalic AQP1 KO mice were much smaller than that of hydrocephalic WT mice (30% of hydrocephalic WT mice at 3 dpi and 50% at 5 dpi) (p b 0.01). During post-operative monitoring of the experimental mice a daily weight was recorded. Initially both WT and AQP1 KO hydrocephalic mice lost a similar degree of body weight. However, after 3 dpi, there was a trend that hydrocephalic AQP1 KO mice regained body weight more quickly compared to hydrocephalic WT mice, although the difference in body weight loss percentage between groups was not significant (Fig. 5B). The grooming ability of AQP1 KO mice injected with kaolin was not affected when compared to WT mice administered with kaolin. This is consistent with the finding of lower ventricular dilation after cisternal kaolin injection seen in AQP1 KO mice compared to wild-type mice. Discussion Our study showed that total choroidal AQP1 protein expression, as measured by Western blot analysis, was not altered in mice with kaolininduced experimental hydrocephalus. However, these mice had reduced AQP1 labeling along the apical membrane and microvilli using immunohistochemistry, immunofluorescence and immuno-gold EM, associated with an increase in AQP1 labeling within the cytoplasm of choroid plexus epithelial cells. In addition, AQP1 KO mice had a smaller ventricular size compared to control WT mice, and ventricular dilation was less in AQP1 KO mice after kaolin injection compared to WT mice. Our study suggests that, despite a change in localization of choroidal AQP1 in response to hydrocephalus, choroidal AQP1 is a potential therapeutic target for management of hydrocephalus. We chose to study choroidal AQP1 using the cisternal kaolin model of hydrocephalus in adult mice. AQP1 was studied because of its role in choroidal CSF production. The other main aquaporin in the brain, AQP4, is expressed in astrocytes and has been shown to be upregulated in experimental hydrocephalus (Mao et al., 2006). As experimental hydrocephalus in AQP4 KO is exacerbated, it appears that AQP4 facilitates compensatory transependymal CSF absorption (Bloch et al., 2006). Strategies aimed at inhibiting or downregulating AQP1 have the advantage of reducing CSF production in the face of CSF outflow obstruction, which is the fundamental problem in hydrocephalus. The experimental model is well established in rat (Del Bigio and Zhang, 1998; Ding et al., 2001; Ishizaki et al., 2000; Tashiro et al., 1997a) and has been described in mice by other groups (Bloch et al., 2006; Hatta et al., 2006; Lopes Lda et al., 2009). We utilised a relatively
low dose of kaolin (5%) to induce hydrocephalus, which gave a high survival rate (97%) and reproducible ventricular enlargement. However, the degree of dilation was not as severe as in neonatal models in mice or chronic hydrocephalus models in rats (Cosan et al., 2000; Mao et al., 2006; Tashiro et al., 1997b). Repeating this study with a more severe hydrocephalus model may be informative. The altered AQP1 localization in response to hydrocephalus was unexpected, though there is precedent for AQP1 redistribution from the plasma membrane in some conditions. Following stimulation with choleretic agonists such as secretin, AQP1 was reported to translocate from intracellular vesicles to the apical membrane in cholangiocytes to increase ductal bile secretion (Tietz et al., 2006). In our study, AQP1 gold-labeling found in the cytoplasm was often seen in vesicles and colocalized with lysosome-associated membrane protein Lamp2 suggesting that endocytosis of AQP1 had occurred in choroidal cells in response to hydrocephalus. Whether our observations reflect a compensatory mechanism for hydrocephalus or simply reflect damage to the choroidal epithelial cells requires further study. In order to examine the potential role of AQP1 in the development of hydrocephalus we applied the cisternal kaolin model to AQP1 KO mice. Use of AQP1 KO mice was necessary as there are no non-toxic pharmacological agents for AQP1 blockade. The lower baseline ventricular size in AQP1 KO mice is consistent with the role of AQP1 in CSF formation. Although there was some ventricular enlargement after cisternal kaolin injection in AQP1 KO mice, it was much less than that observed in WT mice. The implication of these findings supports a role for AQP1 in CSF secretion and in the pathogenesis of hydrocephalus. While AQP1 is not essential for CSF formation it appears to contribute to a significant degree. Even though AQP1 KO mice have only a 30% reduction in CSF production (Oshio et al., 2005), this may be significant enough to influence hydrocephalus formation and the degree of symptoms particularly when the pressure volume characteristics of the intracranial compartment are considered. Acetazolamide, a carbonic anhydrase inhibitor, also reduces CSF production to a similar degree but remains clinically useful in conditions such as pseudotumor cerebri, although its benefits are limited by side effects such as paraesthesias and renal calculi. An AQP1 blocking agent may be a useful alternative or have additive effects to acetazolamide. We suggest that, while there may be a compensatory mechanism in hydrocephalus in the form of choroidal AQP1 endocytosis, this mechanism appears to be overwhelmed. The results in the AQP1 KO mice suggest that the ability to block the contribution of the remaining apical AQP1 may have clinical benefit. Conclusion We found a significant redistribution of AQP1 protein to intracellular vesicles following kaolin-induced hydrocephalus in
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
mice that appears to involve AQP1 endocytosis. The reduced plasma membrane AQP1 expression, and consequential reduced cell osmotic water permeability may be a compensatory mechanism resulting in reduced CSF secretion. AQP1 deficiency results in reduced baseline ventricular size and less severe ventriculomegaly than in WT mice. We conclude that AQP1 is a potential therapeutic target for the treatment of hydrocephalus. Experimental methods Wild-type (WT) Quackenbush Swiss (QS) mice were purchased from the Australian Animal Resources Centre (ARC). AQP1 KO mice were generated by targeted gene disruption was described previously (Ma et al., 1998). A total of 123 WT and 31 AQP1 KO mice aged 6 to 8 weeks (weight N 21 g) were used in this study. All procedures were performed in accordance with National Health and Medical Research Council (NHMRC) of Australia guidelines. All experiments were approved by the local Animal Care and Ethics Committee (Ethics No. K265/07 and K277/08). Cisternal kaolin hydrocephalus mouse model Hydrocephalic WT and AQP1 KO mice were generated as described below, and phenotypes were compared with saline injected control WT and control AQP1 KO mice. Under ketamine/xylazine anaesthesia, the mice were positioned with the head gently flexed and a 1 cm midline vertical incision was made over the posterior neck. The posterior cervical muscles were split in the midline and the occipitocervical membrane visualized. The cisterna magna, which is immediately located underneath the membrane, was injected with 10 μL of 5% kaolin suspension (Sigma-Aldrich, Cat#: K7375) in saline through a 30 G needle. The skin was sutured using a 4-0 silk suture and the mice recovered. Control mice received a 10 μL cisternal injection of sterile saline. Histological analysis Mice were sacrificed by CO2 asphyxia 3 or 5 days post injection (dpi). Mice were perfused with 4% paraformaldehyde and the brains were removed and processed for paraffin embedding. Serial coronal sections (7 μm) were cut. In order to compare the ventricular sizes, the coronal sections with the anatomical landmarks of anterior commissure and optic nerve were stained with haemoxylin and eosin (H&E). Images of H&E stained sections were taken at 12.5× magnification by using a Leica MZ8 Dissecting Microscope (Leica Microsystems Ltd, Heerbrugg, Switzerland) equipped with a DC500 camera and Image Manager (IM50) image capture software (Leica Microsystems Ltd , Heerbrugg, Switzerland). Images were analyzed with Image J software (http://rsb.info.nih.gov/ij) and resized by using Photoshop (Version 7.0). Immunohistochemical (IHC) analysis Coronal paraffin sections of mice were stained with anti-AQP1 antibodies (rabbit anti-rat AQP1, Cat#: AQP11-A, from AlphaDiagnostics, at 1: 200 dilution) and standard IHC procedures (Au et al., 2008, 2004) were followed. IHC images were analyzed with an Olympus BX50 epifluorescence microscope (Olympus Optical CO., Ltd. Japan), and a Jenoptik ProgRes CF Scan digital imaging camera (Jenoptik Laser Optik Systeme GmbH, Jena) was used to capture images at standard exposure and gain. Western blot analysis The lateral choroid plexuses from hydrocephalic and control mice at 3 and 5 dpi were collected and stored in a freezer at −80 °C. The
323
whole choroid plexus was homogenized in ice-cold whole cell lysis buffer [50 mM Tris (pH 8.0), 150 mM NaCl, 0.1% sodium n-dodecyl sulfate (SDS), 1% NP-40, 1 mM phenylmethyl sulfonyl fluoride (PMSF)] containing protease inhibitor cocktail (Sigma-Aldrich, St Louis, MO). The homogenate was briefly spun at 1000 ×g to remove cell debris. Protein concentrations in lysates were determined with Bio-Rad protein assay kit (Bio-Rad, Hercules, CA). Ten micrograms of proteins were used following the standard Western-Blot procedure for AQP1 protein as described (Au et al., 2008). Protein bands were scanned and analyzed with the GS-800 Calibrated Densitometer (Bio-Rad) and Quantity One program v4.2.2. Arbitrary densitometry reading of AQP1 protein band was normalized to the control protein glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and normalized to the average of the control group.
Immunofluorescence staining and confocal microscopy Frozen sections of 6-μm thickness were incubated with the primary AQP1 antibody (1:200 dilution) as described above and Lamp2 antibody (ABL-93, Developmental Studies Hybridoma Bank, Iowa City, IA) at 1:100 dilution. Immunoreactivity was detected with Alexa Fluor 488-conjugated goat anti-rabbit IgG (A-21426, Molecular Probes, Eugene, OR) and Cy3-conjugated goat anti-rat IgG (Jackson ImmunoResearch Laboratories, Westgrove, PA). Confocal microscopy was performed using a Leica TCS SP2 Scanning Confocal Microscope equipped with HCX Plan Apo (PH3) 63×/1. 4 and 100×/1.32 oil immersion objective lenses. Alexa 488 and Cy3 were excited at 488 and 543 nm, respectively. Images were merged using Leica LCS software and figures were assembled using Adobe Photoshop (Version 7). Confocal images were taken at the same photomultiplier tube (PMT) gain and threshold, and same frame and line average. Regions of interest (ROI) were selected manually in the intracellular region (including nucleus and basolateral membrane) and the apical membrane region of 30–50 choroidal cells from each control (n = 6) and hydrocephalic mice (n = 10) at 5 dpi. Mean gray values from each ROI were measured with Image J software.
Immuno-electron microscopy (Immuno-EM) At 3 and 5 dpi, choroid plexus from WT mice were collected and fixed with 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M phosphate buffer and then dehydrated in ethanol followed by embedded in LR White resin. Polymerization was performed for 2.5 h at 55 °C (Bowling and Vaughn, 2008), and ultrathin sections (70 nm) were cut. Immunostaining was carried out in a Leica IGL Immunostainer (Leica Microsystems, Vienna, Austria). After blocking, sections were incubated with rabbit anti-rat AQP1 (1:200) at room temperature for 2 h, followed by goat-anti-rabbit IgG conjugated to 10-nm gold particles (BBI, UK) (1:40). Omission of primary antibody served as the negative control. After staining with uranyl acetate and Reynold's lead citrate, sections were examined using a transmission electron microscope (Philips CM120 Biotwin TEM, FEI Worldwide Corporate Headquarters, Oregon, USA). Sections were visualized by using SIS iTEM TEM imaging software and SIViewer program (Soft Imaging Systems, analySIS, GmBh, Germany). For quantification of gold particles and microvilli area, 20 randomly selected fields (at ×24500) from apical membranes of microvilli from each section were taken. Numbers of immunogold particles were counted by using ImageTool software (http://ddsdx.uthscsa.edu/dig/itdesc.html). Total microvilli areas were measured with ImageJ by using the “threshold” and “find particles” command. Numbers of gold particles in cytoplasm were counted from 25 to 45 choroidal cells in each section from each mouse. Assessment was performed by two observers blinded to the genotype and treatment history of the specimen.
324
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
Statistical analysis All data were expressed as mean± SEM. SPSS for Windows (version 15.0) was used for analysis. The statistical significance of differences was determined by non-parametric Mann–Whitney test (for 2 independent samples) and Kruskal–Wallis Test (for 3 independent samples). p b 0.05 was considered statistically significant. Acknowledgments This work was funded by the Madeline Foundation for Neurosurgical Research and the Brain Foundation, Australia. We thank Mr. Ross Boadle for his assistance with electron microscopy and Ms. Rebecca Reilly, Manager, Transgenic Animal Facility. References Au, C.G., Cooper, S.T., Lo, H.P., Compton, A.G., Yang, N., Wintour, E.M., North, K.N., Winlaw, D.S., Au, C.G., Cooper, S.T., Lo, H.P., Compton, A.G., Yang, N., Wintour, E.M., North, K.N., Winlaw, D.S., 2004. Expression of aquaporin 1 in human cardiac and skeletal muscle. J. Mol. Cell. Cardiol. 36, 655–662. Au, C.G., Butler, T.L., Egan, J.R., Cooper, S.T., Lo, H.P., Compton, A.G., North, K.N., Winlaw, D.S., Au, C.G., Butler, T.L., Egan, J.R., Cooper, S.T., Lo, H.P., Compton, A.G., North, K.N., Winlaw, D.S., 2008. Changes in skeletal muscle expression of AQP1 and AQP4 in dystrophinopathy and dysferlinopathy patients. Acta Neuropathol. 116, 235–246. Bloch, O., Auguste, K.I., Manley, G.T., Verkman, A.S., 2006. Accelerated progression of kaolin-induced hydrocephalus in aquaporin-4-deficient mice. J. Cereb. Blood Flow Metab. 26, 1527–1537. Bowling, A.J., Vaughn, K.C., 2008. A simple technique to minimize heat damage to specimens during thermal polymerization of LR White in plastic and gelatin capsules. J. Microsc. 231, 186–189. Cosan, T.E., Tel, E., Durmaz, R., Gulec, S., Baycu, C., 2000. Non-hindbrain-related syringomyelia. Obstruction of the subarachnoid space and the central canal in rats. An experimental study. J. Neurosurg. Sci. 44, 123–127.
Del Bigio, M.R., Zhang, Y.W., 1998. Cell death, axonal damage, and cell birth in the immature rat brain following induction of hydrocephalus. Exp. Neurol. 154, 157–169. Ding, Y., McAllister II, J.P., Yao, B., Yan, N., Canady, A.I., 2001. Neuron tolerance during hydrocephalus. Neuroscience 106, 659–667. Hatta, J., Hatta, T., Moritake, K., Otani, H., 2006. Heavy water inhibiting the expression of transforming growth factor-beta1 and the development of kaolin-induced hydrocephalus in mice. J. Neurosurg. 104, 251–258. Ishizaki, R., Tashiro, Y., Inomoto, T., Hashimoto, N., 2000. Acute and subacute hydrocephalus in a rat neonatal model: correlation with functional injury of neurotransmitter systems. Pediatr. Neurosurg. 33, 298–305. Longatti, P., Basaldella, L., Orvieto, E., Dei Tos, A., Martinuzzi, A., Longatti, P., Basaldella, L., Orvieto, E., Dei Tos, A., Martinuzzi, A., 2006. Aquaporin(s) expression in choroid plexus tumours. Pediatr. Neurosurg. 42, 228–233. Lopes Lda, S., Slobodian, I., Del Bigio, M.R., 2009. Characterization of juvenile and young adult mice following induction of hydrocephalus with kaolin. Exp. Neurol. 219, 187–196. Ma, T., Yang, B., Gillespie, A., Carlson, E.J., Epstein, C.J., Verkman, A.S., 1998. Severely impaired urinary concentrating ability in transgenic mice lacking aquaporin-1 water channels. J. Biol. Chem. 273, 4296–4299. Mao, X., Enno, T.L., Del Bigio, M.R., Mao, X., Enno, T.L., Del Bigio, M.R., 2006. Aquaporin 4 changes in rat brain with severe hydrocephalus. Eur. J. Neurosci. 23, 2929–2936. Oshio, K., Watanabe, H., Song, Y., Verkman, A.S., Manley, G.T., Oshio, K., Watanabe, H., Song, Y., Manley, G.T., 2005. Reduced cerebrospinal fluid production and intracranial pressure in mice lacking choroid plexus water channel Aquaporin-1. FASEB J. 19, 76–78. Smith, Z.A., Moftakhar, P., Malkasian, D., Xiong, Z., Vinters, H.V., Lazareff, J.A., Smith, Z.A., Moftakhar, P., Malkasian, D., Xiong, Z., Vinters, H.V., Lazareff, J.A., 2007. Choroid plexus hyperplasia: surgical treatment and immunohistochemical results. Case report. J. Neurosurg. 107, 255–262. Tashiro, Y., Chakrabortty, S., Drake, J.M., Hattori, T., 1997a. Progressive loss of glutamic acid decarboxylase, parvalbumin, and calbindin D28K immunoreactive neurons in the cerebral cortex and hippocampus of adult rat with experimental hydrocephalus. J. Neurosurg. 86, 263–271. Tashiro, Y., Drake, J.M., Chakrabortty, S., Hattori, T., 1997b. Functional injury of cholinergic, GABAergic and dopaminergic systems in the basal ganglia of adult rat with kaolin-induced hydrocephalus. Brain Res. 770, 45–52. Tietz, P.S., McNiven, M.A., Splinter, P.L., Huang, B.Q., Larusso, N.F., Tietz, P.S., McNiven, M.A., Splinter, P.L., Huang, B.Q., Larusso, N.F., 2006. Cytoskeletal and motor proteins facilitate trafficking of AQP1-containing vesicles in cholangiocytes. Biol. Cell 98, 43–52.
Contents lists available at ScienceDirect
Molecular and Cellular Neuroscience j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y m c n e
Altered cellular localization of aquaporin-1 in experimental hydrocephalus in mice and reduced ventriculomegaly in aquaporin-1 deficiency☆ Dongwei Wang a, Marko Nykanen a, Nan Yang a, David Winlaw a, Kathryn North a,b, A.S. Verkman c, Brian Kenneth Owler a,b,⁎ a b c
Institute for Neuroscience and Muscle Research Children's Hospital at Westmead, Sydney, Australia Discipline of Child Health and Paediatrics, Children's Hospital at Westmead, University of Sydney, Sydney, Australia Departments of Medicine and Physiology, University of California, San Francisco, San Francisco, California, USA
a r t i c l e
i n f o
Article history: Received 21 July 2010 Revised 29 September 2010 Accepted 21 October 2010 Available online 30 October 2010 Keywords: Hydrocephalus Aquaporins Choroid plexus Cerebrospinal fluid
a b s t r a c t Hydrocephalus is a pathological accumulation of cerebrospinal fluid (CSF) in the cerebral ventricles that constitutes a significant cause of neurological morbidity and mortality. Surgical treatment involving shunt placement is associated with a high failure rate and complications due to infection, motivating the development of alternative, non-surgical therapies. Here, we investigated the role in hydrocephalus of water channel aquaporin-1 (AQP1), which is expressed at the apical membrane of choroid plexus epithelium and is believed to facilitate CSF production. AQP1 expression and subcellular localization were studied in a kaolininduced hydrocephalus model in mice and the effect AQP1 deficiency on the severity of hydrocephalus was determined. While total choroidal AQP1 protein was not significantly altered in hydrocephalus, ~ 50% of AQP1 protein was redistributed from the apical membrane to intracellular vesicles. We found that the ventricular size in AQP1-deficient mice was smaller than in wild-type mice, both at baseline and following hydrocephalus. The reduced plasma membrane AQP1 localization following kaolin-induced hydrocephalus, which involves endocytosis, may be a compensatory mechanism to reduce CSF secretion. The reduced ventricular size in AQP1-deficient mice following kaolin-induced hydrocephalus suggests AQP1 inhibition or down-regulation as a potential adjunctive treatment for hydrocephalus. © 2010 Elsevier Inc. All rights reserved.
Introduction Hydrocephalus, a condition in which an excessive volume of cerebrospinal fluid (CSF) accumulates in the cerebral ventricles, remains an important cause of neurological deficit, delayed development and death. Treatment is generally surgical, requiring for most patients the insertion of a CSF shunt. However, shunts have a failure rate of N30% within the first 12 months, requiring revision. Improved or alternative treatments for hydrocephalus would be desirable. The water channel aquaporin-1 (AQP1) provides one molecular mechanism for CSF secretion and is thus a potential target for nonsurgical management of hydrocephalus. AQP1 is an integral membrane protein expressed in the apical membrane of choroid plexus epithelium. Studies in mice lacking AQP1 (AQP1KO mice) suggest involvement of AQP1 in CSF secretion. AQP1KO mice have ~ 2-fold reduced baseline intracranial pressure compared to wild-type (WT) mice, ~25% lower CSF production, and 5-fold lower osmotically driven water permeability (Oshio et al., 2005). In that study AQP1KO mice ☆ Grant information: (NIH DK35124). ⁎ Corresponding author. Suite 19, Children's Hospital Medical Centre, Hainsworth Street, Westmead NSW 2145, Australia. Fax: + 61 2 98060060. E-mail address: [email protected] (B.K. Owler). 1044-7431/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.mcn.2010.10.003
had improved clinical outcome compared to wild-type mice in a brain injury model; however, it was not possible to determine whether the improved outcome was the result of reduced CSF secretion and/or reduced central venous pressure (and consequent reduced intracranial pressure) related to renal effects of AQP1 deletion. Little is known about the role of AQP1 in the development of hydrocephalus. Hydrocephalus has not been studied in AQP1KO mice, and there is conflicting information about AQP1 expression in hydrocephalus. AQP1 expression was unchanged in a rat model of hydrocephalus (Mao et al., 2006), reduced in a single human case of choroid plexus hyperplasia (Smith et al., 2007), and elevated in choroid plexus papilloma associated with communicative hydrocephalus (Longatti et al., 2006). We report here an analysis of AQP1 expression and cellular localization in a kaolin-induced mouse model of hydrocephalus, and evaluation, using AQP1KO mice, of whether AQP1 deficiency alters the severity of experimental hydrocephalus. Results Kaolin-injected WT mice develop hydrocephalus By 5 days post injection, compared to saline-injected mice (Fig. 1A), dilated ventricles were observed in mice injected with kaolin (Fig. 1B
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
319
Fig. 1. Characterization of kaolin-induced hydrocephalus mouse model in wild-type (WT) mice. (A, B) H&E staining of coronal sections of mouse brain from saline (A) and kaolin (B)-injected mice at 5 dpi. (C, D) H&E staining of sagittal sections of mouse brain from kaolin-injected mice. Panel D is the high magnification image of the boxed region in panel C. Note that the fourth ventricle is filled with kaolin suspension (arrowhead). (E) Ventricular area in untreated, saline-injected and kaolin-injected mice at 3 and 5 dpi (*pb 0.05, **pb 0.001, Kruskal–Wallis Test). The error bar represents SEM. The number in the bar indicates the number of mice used per group. Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice.
and C), as well as inflammatory cells in the subarachnoid space in mice injected with kaolin (Fig. 1D), as expected in this inflammatory model of hydrocephalus. We measured the size of the area of the lateral ventricles in untreated, saline injected control mice and hydrocephalic WT mice at 3 and 5 dpi. There was no statistical significance between untreated and saline injected mice at 3 or 5 days. The mean ventricular area in the kaolin-injected mice was significantly larger compared to saline injected control mice at both time points (Fig. 1E). The degree of the increase in ventricular size is similar to mice reported with the mild hydrocephalus phenotypes in adult mouse models (Bloch et al., 2006) as opposed to the more severe neonatal models of hydrocephalus.
Total AQP1 expression is not altered in hydrocephalic mice We investigated whether AQP1 expression in WT mice was altered following experimental hydrocephalus. Western blot showed one band at 28 kDa, corresponding to non-glycosylated AQP1, and a diffuse band at ~ 38 kDa corresponding to glycosylated AQP1 (Fig. 2A). Quantitative analysis of total protein from choroid plexus showed no difference in AQP1 protein expression between hydrocephalic and control WT mice at 5 dpi (Fig. 2B) as well as at 3 dpi (data not shown).
Fig. 2. Quantification of AQP1 in saline-injected controls and kaolin-injected hydrocephalic WT mice at 5 days post-injection by Western blotting. (A) Representative Western blot of AQP1 and GAPDH of experimental and control mice. (B) The normalized ratio of the densitometry readings of AQP1 to GAPDH in saline-injected mice (n = 7) and kaolin-injected mice (n = 10). The error bar represents SEM. Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice.
320
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
Altered subcellular location of AQP1 in hydrocephalus mice We hypothesized that the pathological accumulation of CSF in hydrocephalic WT mice may be associated with reduced AQP1 protein in apical membrane of choroid plexus epithelium. Immunohistochemistry (IHC) and immunofluorescence (IF) were done on the mice randomly selected from the cohort of mice used for ventricle area measurement. As seen in Fig. 3A–E, AQP1 immunoreactivity was largely absent in the cytoplasm but present on the apical membrane of the epithelial cells in control mice. However, AQP1 immunoreactivity was seen both in the cytoplasm and on the apical membrane of epithelial cells in 3 out of 6 of randomly picked hydrocephalic mice at 3 dpi and 7 out of 10 hydrocephalic mice at 5 dpi (Fig. 3F–K) A measurement of mean gray value (MGV) of AQP1 immunofluorescence signals was performed on the above 6 saline-injected control mice and 10 hydrocephalic mice at 5 dpi. There was an increase in the fluorescence signals in the cytoplasm and basolateral membrane in the cells of hydrocephalic mice compared to salineinjected control mice (p b 0.05) (Fig. 3L). In total MGV of the choroidal cells, 91% was from the apical membrane in the saline-injected control mice, while 78% was from the apical membrane in the kaolin-injected hydrocephalic mice (Fig. 3M). Fig. 3O and P show AQP1 immunofluorescence in cells in an endosomal pattern, with occasional coloca-
lization with lysosome-associated membrane marker Lamp2, particularly in the region adjacent to the basolateral membrane. To further investigate the apparent mislocalization of AQP1 in experimental hydrocephalus, immuno-EM was done on 3 hydrocephalic WT mice at 3 dpi, 3 hydrocephalic WT mice at 5 dpi and 4 control mice at 3 dpi. AQP1-positive gold particles were numerous along the microvilli of the apical membrane on the choroidal epithelial cells in both control and hydrocephalic mice (Fig. 4A–D). However, in control mice, apart from the occasional presence of gold particles in mitochondria, no positive staining was found within the cell cytoplasm or along the basolateral membrane. In hydrocephalic mice, AQP1-positive gold particles were observed within vacuoles and vesicles (Fig. 4E(i)), within the enlarged intercellular space (Fig. 4E (ii)) and in the basolateral membrane (Fig. 4E(iii)). The number of gold particles along microvilli and inside the cytoplasm of choroidal cells in the electron-micrograph images from hydrocephalic and control mice was measured (Table 1). Data from hydrocephalic mice at 3 and 5 dpi were combined for analysis. Gold labeling within the cytoplasm was significantly increased in hydrocephalic mice compared to control mice (p = 0.03). There was a trend that the number of gold particles along the microvilli and the microvillus area was reduced in hydrocephalic mice but the difference between hydrocephalic mice and control mice did not reach significance
Fig. 3. AQP1 localization in choroid plexus epithelial cell of saline-injected control and kaolin-injected hydrocephalic WT mice at 5 dpi. (A, F) IHC images from frozen sections (7 μm) were stained with rabbit AQP1 as primary antibody followed by incubation of secondary antibody and standard ABC technique. (B–E, G–K) Confocal images of choroid plexus epithelial cell of control (B–E) and hydrocephalic (G–K) mice. Panel K is the high magnification image of the boxed region in panel J. AQP1 signals were detected by anti-rabbit IgG conjugated to Alexa 488. Positive staining for AQP1 is on the apical surface of the epithelial cells of the saline-injected mice (A–E). In contrast, in hydrocephalic mice, positive staining is also found inside the cytoplasm (F–K). (L) Intracellular AQP1 immunoreactivity demonstrating by mean gray values by Image J analysis. (M) Percentage of apical AQP1 immunoreactivity relative to total AQP1 immunoreactivity. The error bar represents SEM. The number in the bar indicates the number of mice used per group (*p b 0.05. Mann– Whitney Test U test). Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice. (N–P) Double immunofluorescence staining with anti-AQP1 (green) and anti-Lamp2 (red) on choroidal cells of control (N(i) and N(ii)) and hydrocephalic (O(i)–(iii) and P) mice. Panels N (ii), O (iii) and P are overlaid images. Images showing Lamp2 red fluorescence corresponding to panels N(ii) and P and AQP1 green fluorescence corresponding to panel P are not shown. Panel P is the high magnification image to boxed region in panel N. Colocalization of positive signals of AQP1 and Lamp2 is shown in yellow (arrowheads).
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
321
Fig. 4. Anti-AQP1 labeling of choroidal epithelium in saline-injected control and kaolin-injected hydrocephalic WT mice. (A–D) Micrograph of microvilli from control and hydrocephalic mice. Gold particles are on the surface of the microvilli in control mice (A and C) and hydrocephalic mice (B and D). The microvilli in hydrocephalic mice are swollen. (E) Low magnification image of choroidal epithelium in hydrocephalic mice. Note that the tight junction (white arrow) is still intact. (E(i)–(iii)) Higher magnification electron micrographs corresponding to the boxes in panel G show gold particles (black arrows) associated with membrane-like structure in a vacuole in the cytoplasm (i), in the intercellular space (ii) and in the nearby area of the basal membrane of the epithelia (iii). Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice. Scale bar in (i)–(iii) = 100 nm. Con: saline-injected mice; HC: kaolin-injected hydrocephalic mice.
(p = 0.055). There was no difference in the density of AQP1 positive signals along microvilli of the apical membrane in control and kaolininjected mice. The data suggest that the reduction of the total amount of
AQP1 signals on the microvilli may result from a reduction of area of microvilli in kaolin-injected mice. Reduced ventricular size in kaolin-injected AQP1 KO mice
Table 1 Quantification of gold particles on apical membranes and microvilli and in the cytoplasm. Gold particle in Cytoplasm Mean Control (n = 4) HC (n = 6) p value
SEM
4.6 1.0 14.6 3.3 0.033
Mv area
Gold particles on Mv Mean
Density of gold particle on Mv
SEM
Mean
SEM
Mean
473 133 228 45 0.055
8.2 4.7 0.055
1.2 0.8
78.76 3.52 68.25 22.26 0.088
Mv: microvilli; Con: control mice; HC: hydrocephalic mice.
SEM
The reduction in apical membrane-associated choroidal AQP1 in hydrocephalic mice suggests the possibility of a compensatory mechanism to reduce CSF production in hydrocephalus. Therefore, the potential of AQP1 as a therapeutic target for the treatment hydrocephalus may be limited. In order to clarify the potential utility of AQP1 we studied whether the absence of AQP1 in the AQP1 KO mice influenced ventricular size and the formation of hydrocephalus. As shown in Fig. 5A, saline injected AQP1 KO mice had significantly smaller ventricular areas compared to saline injected WT mice (p = 0.001). After cisternal saline injection AQP1 KO mice demonstrated
322
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
Fig. 5. Characterization of kaolin-induced hydrocephalus mouse model in WT and AQP1 KO mice. (A) Ventricular area in untreated, saline-injected and kaolin-injected WT and AQP1 KO mice at 3 and 5 dpi. (B) Body weight loss percentage over 5 days after cisternal kaolin injection in WT and AQP1 KO mice. WT-Con: saline-injected WT mice (n = 15); WT-HC, kaolin-injected hydrocephalic WT mice (n = 24); KO-Con: saline-injected KO mice (n = 12); KO-HC: kaolin-injected KO mice (n = 17).
no change in ventricular size at either 3 or 5 dpi compared to untreated AQP1 KO mice. AQP1 KO mice that had undergone cisternal kaolin injection had significantly larger lateral ventricles compared to saline-injected AQP1 KO controls at both 3 and 5 dpi (p b 0.05). However, when compared with the ventricular sizes of kaolin-injected WT mice, we found that the ventricular sizes in hydrocephalic AQP1 KO mice were much smaller than that of hydrocephalic WT mice (30% of hydrocephalic WT mice at 3 dpi and 50% at 5 dpi) (p b 0.01). During post-operative monitoring of the experimental mice a daily weight was recorded. Initially both WT and AQP1 KO hydrocephalic mice lost a similar degree of body weight. However, after 3 dpi, there was a trend that hydrocephalic AQP1 KO mice regained body weight more quickly compared to hydrocephalic WT mice, although the difference in body weight loss percentage between groups was not significant (Fig. 5B). The grooming ability of AQP1 KO mice injected with kaolin was not affected when compared to WT mice administered with kaolin. This is consistent with the finding of lower ventricular dilation after cisternal kaolin injection seen in AQP1 KO mice compared to wild-type mice. Discussion Our study showed that total choroidal AQP1 protein expression, as measured by Western blot analysis, was not altered in mice with kaolininduced experimental hydrocephalus. However, these mice had reduced AQP1 labeling along the apical membrane and microvilli using immunohistochemistry, immunofluorescence and immuno-gold EM, associated with an increase in AQP1 labeling within the cytoplasm of choroid plexus epithelial cells. In addition, AQP1 KO mice had a smaller ventricular size compared to control WT mice, and ventricular dilation was less in AQP1 KO mice after kaolin injection compared to WT mice. Our study suggests that, despite a change in localization of choroidal AQP1 in response to hydrocephalus, choroidal AQP1 is a potential therapeutic target for management of hydrocephalus. We chose to study choroidal AQP1 using the cisternal kaolin model of hydrocephalus in adult mice. AQP1 was studied because of its role in choroidal CSF production. The other main aquaporin in the brain, AQP4, is expressed in astrocytes and has been shown to be upregulated in experimental hydrocephalus (Mao et al., 2006). As experimental hydrocephalus in AQP4 KO is exacerbated, it appears that AQP4 facilitates compensatory transependymal CSF absorption (Bloch et al., 2006). Strategies aimed at inhibiting or downregulating AQP1 have the advantage of reducing CSF production in the face of CSF outflow obstruction, which is the fundamental problem in hydrocephalus. The experimental model is well established in rat (Del Bigio and Zhang, 1998; Ding et al., 2001; Ishizaki et al., 2000; Tashiro et al., 1997a) and has been described in mice by other groups (Bloch et al., 2006; Hatta et al., 2006; Lopes Lda et al., 2009). We utilised a relatively
low dose of kaolin (5%) to induce hydrocephalus, which gave a high survival rate (97%) and reproducible ventricular enlargement. However, the degree of dilation was not as severe as in neonatal models in mice or chronic hydrocephalus models in rats (Cosan et al., 2000; Mao et al., 2006; Tashiro et al., 1997b). Repeating this study with a more severe hydrocephalus model may be informative. The altered AQP1 localization in response to hydrocephalus was unexpected, though there is precedent for AQP1 redistribution from the plasma membrane in some conditions. Following stimulation with choleretic agonists such as secretin, AQP1 was reported to translocate from intracellular vesicles to the apical membrane in cholangiocytes to increase ductal bile secretion (Tietz et al., 2006). In our study, AQP1 gold-labeling found in the cytoplasm was often seen in vesicles and colocalized with lysosome-associated membrane protein Lamp2 suggesting that endocytosis of AQP1 had occurred in choroidal cells in response to hydrocephalus. Whether our observations reflect a compensatory mechanism for hydrocephalus or simply reflect damage to the choroidal epithelial cells requires further study. In order to examine the potential role of AQP1 in the development of hydrocephalus we applied the cisternal kaolin model to AQP1 KO mice. Use of AQP1 KO mice was necessary as there are no non-toxic pharmacological agents for AQP1 blockade. The lower baseline ventricular size in AQP1 KO mice is consistent with the role of AQP1 in CSF formation. Although there was some ventricular enlargement after cisternal kaolin injection in AQP1 KO mice, it was much less than that observed in WT mice. The implication of these findings supports a role for AQP1 in CSF secretion and in the pathogenesis of hydrocephalus. While AQP1 is not essential for CSF formation it appears to contribute to a significant degree. Even though AQP1 KO mice have only a 30% reduction in CSF production (Oshio et al., 2005), this may be significant enough to influence hydrocephalus formation and the degree of symptoms particularly when the pressure volume characteristics of the intracranial compartment are considered. Acetazolamide, a carbonic anhydrase inhibitor, also reduces CSF production to a similar degree but remains clinically useful in conditions such as pseudotumor cerebri, although its benefits are limited by side effects such as paraesthesias and renal calculi. An AQP1 blocking agent may be a useful alternative or have additive effects to acetazolamide. We suggest that, while there may be a compensatory mechanism in hydrocephalus in the form of choroidal AQP1 endocytosis, this mechanism appears to be overwhelmed. The results in the AQP1 KO mice suggest that the ability to block the contribution of the remaining apical AQP1 may have clinical benefit. Conclusion We found a significant redistribution of AQP1 protein to intracellular vesicles following kaolin-induced hydrocephalus in
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
mice that appears to involve AQP1 endocytosis. The reduced plasma membrane AQP1 expression, and consequential reduced cell osmotic water permeability may be a compensatory mechanism resulting in reduced CSF secretion. AQP1 deficiency results in reduced baseline ventricular size and less severe ventriculomegaly than in WT mice. We conclude that AQP1 is a potential therapeutic target for the treatment of hydrocephalus. Experimental methods Wild-type (WT) Quackenbush Swiss (QS) mice were purchased from the Australian Animal Resources Centre (ARC). AQP1 KO mice were generated by targeted gene disruption was described previously (Ma et al., 1998). A total of 123 WT and 31 AQP1 KO mice aged 6 to 8 weeks (weight N 21 g) were used in this study. All procedures were performed in accordance with National Health and Medical Research Council (NHMRC) of Australia guidelines. All experiments were approved by the local Animal Care and Ethics Committee (Ethics No. K265/07 and K277/08). Cisternal kaolin hydrocephalus mouse model Hydrocephalic WT and AQP1 KO mice were generated as described below, and phenotypes were compared with saline injected control WT and control AQP1 KO mice. Under ketamine/xylazine anaesthesia, the mice were positioned with the head gently flexed and a 1 cm midline vertical incision was made over the posterior neck. The posterior cervical muscles were split in the midline and the occipitocervical membrane visualized. The cisterna magna, which is immediately located underneath the membrane, was injected with 10 μL of 5% kaolin suspension (Sigma-Aldrich, Cat#: K7375) in saline through a 30 G needle. The skin was sutured using a 4-0 silk suture and the mice recovered. Control mice received a 10 μL cisternal injection of sterile saline. Histological analysis Mice were sacrificed by CO2 asphyxia 3 or 5 days post injection (dpi). Mice were perfused with 4% paraformaldehyde and the brains were removed and processed for paraffin embedding. Serial coronal sections (7 μm) were cut. In order to compare the ventricular sizes, the coronal sections with the anatomical landmarks of anterior commissure and optic nerve were stained with haemoxylin and eosin (H&E). Images of H&E stained sections were taken at 12.5× magnification by using a Leica MZ8 Dissecting Microscope (Leica Microsystems Ltd, Heerbrugg, Switzerland) equipped with a DC500 camera and Image Manager (IM50) image capture software (Leica Microsystems Ltd , Heerbrugg, Switzerland). Images were analyzed with Image J software (http://rsb.info.nih.gov/ij) and resized by using Photoshop (Version 7.0). Immunohistochemical (IHC) analysis Coronal paraffin sections of mice were stained with anti-AQP1 antibodies (rabbit anti-rat AQP1, Cat#: AQP11-A, from AlphaDiagnostics, at 1: 200 dilution) and standard IHC procedures (Au et al., 2008, 2004) were followed. IHC images were analyzed with an Olympus BX50 epifluorescence microscope (Olympus Optical CO., Ltd. Japan), and a Jenoptik ProgRes CF Scan digital imaging camera (Jenoptik Laser Optik Systeme GmbH, Jena) was used to capture images at standard exposure and gain. Western blot analysis The lateral choroid plexuses from hydrocephalic and control mice at 3 and 5 dpi were collected and stored in a freezer at −80 °C. The
323
whole choroid plexus was homogenized in ice-cold whole cell lysis buffer [50 mM Tris (pH 8.0), 150 mM NaCl, 0.1% sodium n-dodecyl sulfate (SDS), 1% NP-40, 1 mM phenylmethyl sulfonyl fluoride (PMSF)] containing protease inhibitor cocktail (Sigma-Aldrich, St Louis, MO). The homogenate was briefly spun at 1000 ×g to remove cell debris. Protein concentrations in lysates were determined with Bio-Rad protein assay kit (Bio-Rad, Hercules, CA). Ten micrograms of proteins were used following the standard Western-Blot procedure for AQP1 protein as described (Au et al., 2008). Protein bands were scanned and analyzed with the GS-800 Calibrated Densitometer (Bio-Rad) and Quantity One program v4.2.2. Arbitrary densitometry reading of AQP1 protein band was normalized to the control protein glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and normalized to the average of the control group.
Immunofluorescence staining and confocal microscopy Frozen sections of 6-μm thickness were incubated with the primary AQP1 antibody (1:200 dilution) as described above and Lamp2 antibody (ABL-93, Developmental Studies Hybridoma Bank, Iowa City, IA) at 1:100 dilution. Immunoreactivity was detected with Alexa Fluor 488-conjugated goat anti-rabbit IgG (A-21426, Molecular Probes, Eugene, OR) and Cy3-conjugated goat anti-rat IgG (Jackson ImmunoResearch Laboratories, Westgrove, PA). Confocal microscopy was performed using a Leica TCS SP2 Scanning Confocal Microscope equipped with HCX Plan Apo (PH3) 63×/1. 4 and 100×/1.32 oil immersion objective lenses. Alexa 488 and Cy3 were excited at 488 and 543 nm, respectively. Images were merged using Leica LCS software and figures were assembled using Adobe Photoshop (Version 7). Confocal images were taken at the same photomultiplier tube (PMT) gain and threshold, and same frame and line average. Regions of interest (ROI) were selected manually in the intracellular region (including nucleus and basolateral membrane) and the apical membrane region of 30–50 choroidal cells from each control (n = 6) and hydrocephalic mice (n = 10) at 5 dpi. Mean gray values from each ROI were measured with Image J software.
Immuno-electron microscopy (Immuno-EM) At 3 and 5 dpi, choroid plexus from WT mice were collected and fixed with 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M phosphate buffer and then dehydrated in ethanol followed by embedded in LR White resin. Polymerization was performed for 2.5 h at 55 °C (Bowling and Vaughn, 2008), and ultrathin sections (70 nm) were cut. Immunostaining was carried out in a Leica IGL Immunostainer (Leica Microsystems, Vienna, Austria). After blocking, sections were incubated with rabbit anti-rat AQP1 (1:200) at room temperature for 2 h, followed by goat-anti-rabbit IgG conjugated to 10-nm gold particles (BBI, UK) (1:40). Omission of primary antibody served as the negative control. After staining with uranyl acetate and Reynold's lead citrate, sections were examined using a transmission electron microscope (Philips CM120 Biotwin TEM, FEI Worldwide Corporate Headquarters, Oregon, USA). Sections were visualized by using SIS iTEM TEM imaging software and SIViewer program (Soft Imaging Systems, analySIS, GmBh, Germany). For quantification of gold particles and microvilli area, 20 randomly selected fields (at ×24500) from apical membranes of microvilli from each section were taken. Numbers of immunogold particles were counted by using ImageTool software (http://ddsdx.uthscsa.edu/dig/itdesc.html). Total microvilli areas were measured with ImageJ by using the “threshold” and “find particles” command. Numbers of gold particles in cytoplasm were counted from 25 to 45 choroidal cells in each section from each mouse. Assessment was performed by two observers blinded to the genotype and treatment history of the specimen.
324
D. Wang et al. / Molecular and Cellular Neuroscience 46 (2011) 318–324
Statistical analysis All data were expressed as mean± SEM. SPSS for Windows (version 15.0) was used for analysis. The statistical significance of differences was determined by non-parametric Mann–Whitney test (for 2 independent samples) and Kruskal–Wallis Test (for 3 independent samples). p b 0.05 was considered statistically significant. Acknowledgments This work was funded by the Madeline Foundation for Neurosurgical Research and the Brain Foundation, Australia. We thank Mr. Ross Boadle for his assistance with electron microscopy and Ms. Rebecca Reilly, Manager, Transgenic Animal Facility. References Au, C.G., Cooper, S.T., Lo, H.P., Compton, A.G., Yang, N., Wintour, E.M., North, K.N., Winlaw, D.S., Au, C.G., Cooper, S.T., Lo, H.P., Compton, A.G., Yang, N., Wintour, E.M., North, K.N., Winlaw, D.S., 2004. Expression of aquaporin 1 in human cardiac and skeletal muscle. J. Mol. Cell. Cardiol. 36, 655–662. Au, C.G., Butler, T.L., Egan, J.R., Cooper, S.T., Lo, H.P., Compton, A.G., North, K.N., Winlaw, D.S., Au, C.G., Butler, T.L., Egan, J.R., Cooper, S.T., Lo, H.P., Compton, A.G., North, K.N., Winlaw, D.S., 2008. Changes in skeletal muscle expression of AQP1 and AQP4 in dystrophinopathy and dysferlinopathy patients. Acta Neuropathol. 116, 235–246. Bloch, O., Auguste, K.I., Manley, G.T., Verkman, A.S., 2006. Accelerated progression of kaolin-induced hydrocephalus in aquaporin-4-deficient mice. J. Cereb. Blood Flow Metab. 26, 1527–1537. Bowling, A.J., Vaughn, K.C., 2008. A simple technique to minimize heat damage to specimens during thermal polymerization of LR White in plastic and gelatin capsules. J. Microsc. 231, 186–189. Cosan, T.E., Tel, E., Durmaz, R., Gulec, S., Baycu, C., 2000. Non-hindbrain-related syringomyelia. Obstruction of the subarachnoid space and the central canal in rats. An experimental study. J. Neurosurg. Sci. 44, 123–127.
Del Bigio, M.R., Zhang, Y.W., 1998. Cell death, axonal damage, and cell birth in the immature rat brain following induction of hydrocephalus. Exp. Neurol. 154, 157–169. Ding, Y., McAllister II, J.P., Yao, B., Yan, N., Canady, A.I., 2001. Neuron tolerance during hydrocephalus. Neuroscience 106, 659–667. Hatta, J., Hatta, T., Moritake, K., Otani, H., 2006. Heavy water inhibiting the expression of transforming growth factor-beta1 and the development of kaolin-induced hydrocephalus in mice. J. Neurosurg. 104, 251–258. Ishizaki, R., Tashiro, Y., Inomoto, T., Hashimoto, N., 2000. Acute and subacute hydrocephalus in a rat neonatal model: correlation with functional injury of neurotransmitter systems. Pediatr. Neurosurg. 33, 298–305. Longatti, P., Basaldella, L., Orvieto, E., Dei Tos, A., Martinuzzi, A., Longatti, P., Basaldella, L., Orvieto, E., Dei Tos, A., Martinuzzi, A., 2006. Aquaporin(s) expression in choroid plexus tumours. Pediatr. Neurosurg. 42, 228–233. Lopes Lda, S., Slobodian, I., Del Bigio, M.R., 2009. Characterization of juvenile and young adult mice following induction of hydrocephalus with kaolin. Exp. Neurol. 219, 187–196. Ma, T., Yang, B., Gillespie, A., Carlson, E.J., Epstein, C.J., Verkman, A.S., 1998. Severely impaired urinary concentrating ability in transgenic mice lacking aquaporin-1 water channels. J. Biol. Chem. 273, 4296–4299. Mao, X., Enno, T.L., Del Bigio, M.R., Mao, X., Enno, T.L., Del Bigio, M.R., 2006. Aquaporin 4 changes in rat brain with severe hydrocephalus. Eur. J. Neurosci. 23, 2929–2936. Oshio, K., Watanabe, H., Song, Y., Verkman, A.S., Manley, G.T., Oshio, K., Watanabe, H., Song, Y., Manley, G.T., 2005. Reduced cerebrospinal fluid production and intracranial pressure in mice lacking choroid plexus water channel Aquaporin-1. FASEB J. 19, 76–78. Smith, Z.A., Moftakhar, P., Malkasian, D., Xiong, Z., Vinters, H.V., Lazareff, J.A., Smith, Z.A., Moftakhar, P., Malkasian, D., Xiong, Z., Vinters, H.V., Lazareff, J.A., 2007. Choroid plexus hyperplasia: surgical treatment and immunohistochemical results. Case report. J. Neurosurg. 107, 255–262. Tashiro, Y., Chakrabortty, S., Drake, J.M., Hattori, T., 1997a. Progressive loss of glutamic acid decarboxylase, parvalbumin, and calbindin D28K immunoreactive neurons in the cerebral cortex and hippocampus of adult rat with experimental hydrocephalus. J. Neurosurg. 86, 263–271. Tashiro, Y., Drake, J.M., Chakrabortty, S., Hattori, T., 1997b. Functional injury of cholinergic, GABAergic and dopaminergic systems in the basal ganglia of adult rat with kaolin-induced hydrocephalus. Brain Res. 770, 45–52. Tietz, P.S., McNiven, M.A., Splinter, P.L., Huang, B.Q., Larusso, N.F., Tietz, P.S., McNiven, M.A., Splinter, P.L., Huang, B.Q., Larusso, N.F., 2006. Cytoskeletal and motor proteins facilitate trafficking of AQP1-containing vesicles in cholangiocytes. Biol. Cell 98, 43–52.