Journal of Cancer Research and Therapeutics

: 2012  |  Volume : 8  |  Issue : 4  |  Page : 571--577

Correlation of expression pattern of aquaporin-1 in primary central nervous system tumors with tumor type, grade, proliferation, microvessel density, contrast-enhancement and perilesional edema

Prabal Deb1, Seerat Pal1, Vibha Dutta1, Dibyajyoti Boruah1, Vijai M Chandran2, Harjinder S Bhatoe3,  
1 Department of Pathology, Armed Forces Medical College, Pune, India
2 Department of Radiology, Armed Forces Medical College, Pune, India
3 Department of Neurosurgery, ­Command Hospital, Pune, India

Correspondence Address:
Prabal Deb
Department of Pathology, Armed Forces Medical College, Pune - 411040


Objectives: Brain edema, a hallmark of malignant brain tumors, continues to be a major cause of mortality. The underlying molecular mechanisms are poorly understood and thought to be mediated through membrane water-channels: aquaporins (AQP1,4,9). The abnormal upregulation of AQP1 in certain glial neoplasms has suggested a potential role in tumor pathogenesis, apart from being a novel target for newer therapeutic regimen. This study was undertaken to evaluate the expression of AQP1 in primary CNS tumors of various histologic types and grades, and its correlation with contrast-enhancement, perilesional edema, histomorphology, proliferation index and microvessel density. Materials and Methods: Biopsy tissues from 30 patients (10 each from gliomas, meningiomas and other primary CNS tumors) were studied. Autopsy brain sections served as control. AQP1-immunoreactivity was correlated with histomorphology, radiology, proliferation index and microvessel density (MVD). Results: AQP1 expression was increased in gliomas and ependymal tumors as compared to meningiomas. Intratumoral expression was homogenous in high-grade and membranous in low-grade neoplasms, while peritumoral areas showed expression around vessels and reactive astrocytes. High-grade tumors showed peritumoral upregulation, while low-grade had intense intratumoral expression. A trend of positive correlation was observed between AQP1-immunopositivity and increasing grade, higher MIB-1LI, increasing contrast-enhancement and more perilesional edema, and elevated MVD with raised AQP1:MVD ratio. Conclusions: AQP1-immunoexpression had a good correlation with high-grade tumors. AQP-upregulation in perilesional areas of high-grade tumors suggests its role in vasogenic edema. Further studies involving other AQP molecules, vascular endothelial growth factor (VEGF) and hypoxia inducible factor-1 α (HIF-1α) should be undertaken to evaluate its possible role as a potential surrogate marker of high-grade tumors heralding poor outcome, inhibition of which may serve as the basis for future targeted therapy.

How to cite this article:
Deb P, Pal S, Dutta V, Boruah D, Chandran VM, Bhatoe HS. Correlation of expression pattern of aquaporin-1 in primary central nervous system tumors with tumor type, grade, proliferation, microvessel density, contrast-enhancement and perilesional edema.J Can Res Ther 2012;8:571-577

How to cite this URL:
Deb P, Pal S, Dutta V, Boruah D, Chandran VM, Bhatoe HS. Correlation of expression pattern of aquaporin-1 in primary central nervous system tumors with tumor type, grade, proliferation, microvessel density, contrast-enhancement and perilesional edema. J Can Res Ther [serial online] 2012 [cited 2022 Jun 26 ];8:571-577
Available from:

Full Text


Majority of the primary malignant central nervous system (CNS) tumors are high-grade gliomas. These are characterized by rapid growth rate, high glucose consumption, microvascular proliferation, intratumoral necrosis and hypoxia, breakdown of blood-brain barrier and vasogenic brain edema, and perivascular infiltration of glioma cells. Out of these, brain edema constitutes one of the serious problems, since it significantly contributes to increased patient morbidity and mortality. [1]

Despite intense search, the molecular mechanisms underlying brain edema remains poorly understood. Recent developments in the field of tumor biology of CNS neoplasms have unraveled a wide array of markers, like aquaporins, matrix metalloproteinases, vascular endothelial growth factors, angioproteins and hypoxia inducible factor 1α, that not only give an insight into the tumor pathogenesis, but also have possible prognostic and therapeutic connotations. [2]

Aquaporins (AQP) are water-selective transmembrane transport channel proteins, which are thought to have a major function in the development of vasogenic brain edema. The aquaporin family consists of 13 highly conserved water channel proteins that play a critical role in controlling the water contents of cells. Among the AQP subtypes cloned in mammals, only AQP1, AQP4, and AQP9 have been described in the brain. [3]

AQP1 is classically expressed in the apical surface of the epithelium of the choroid plexus, and may have a function in cerebrospinal fluid (CSF) production. [4] Apart from development of vasogenic edema, AQP1 is also thought to significantly contribute to a number of important processes in the brain and the body, like angiogenesis, cell migration, development and neuropathological diseases. [5] In tumors, AQP1-immunoexpression tends to decrease in the center, compared to the periphery, which may correlate with the water transport within the neoplasm. [6],[7],[8],[9]

Enhancement in a CNS tumor following administration of contrast depicts perturbation in the functional integrity of the blood-brain barrier (BBB), which is often a feature of high-grade tumor. [10]

Literature pertaining to AQP1 expression in brain tumors is limited, with the majority of the studies being directed towards high-grade glial neoplasms. [1],[2],[3],[6],[8],[9] The present study was therefore undertaken to correlate the expression pattern of AQP1 in CNS tumors of various types and grades, and also to assess its correlation with contrast-enhancement and perilesional edema, along with histological features and proliferation index.

 Materials and Methods

The study included biopsy tissue from 30 cases of primary brain tumors that were managed at this tertiary care institute during 2009-10. This included: (a) 10 cases of gliomas [seven glioblastoma multiformes (GBM) and one case each of diffuse astrocytoma (DA), oligoastrocytoma (OA) and anaplastic oligoastrocytoma (AOA)]; (b) 10 cases of meningiomas [seven meningothelial; two fibroblastic; and one psammomatous meningioma]; and (c) 10 cases of other tumors (two cases of ependymomas and pituitary adenomas respectively; and one case each of anaplastic ependymoma, dysembryoplastic neuroepithelial tumor (DNET), primitive neuroepithelial tumour (PNET), germ cell tumor, craniopharyngioma, hemangioblastoma). All metastatic lesions to the brain and cases without available corresponding neuroimages were excluded from the study. Relevant clinical findings were noted for all the cases.

Contrast enhancement was evaluated using gadolinium-enhanced T-1 weighted magnetic resonance imaging (T1WI), while perilesional edema was assessed on FLAIR and T2WI.

Two pathologists (PD and VD) independently reviewed the slides (hematoxylin and eosin: H and E along with relevant immunohistochemical stains) of all the cases, and reconfirmed the original histopathological diagnoses, viz. tumor type and grade, as per WHO 2007 classification system for CNS tumors.

One representative block of formalin-fixed paraffin-embedded tissue was selected, 5-micron thick sections cut and immunohistochemical (IHC) staining performed by LSAB technique (LSAB Kit, M/s Dakopatts, Denmark) using monoclonal antibodies AQP1 (1 : 50, M/s Abcam, Germany) for 4 h at room temperature in a moist chamber. In all cases, high-temperature antigen retrieval was performed by autoclaving the sections in 0.01 M citrate buffer (pH 6.0) at 121°C for 10 min. Visualization was done by treating the sections with a solution containing 0.06 mM 3,30-diaminobenzidine and 2 mM hydrogen peroxide in 0.05% Tris-HCl buffered at pH 7.6 for 10 min. Slides were counterstained with hematoxylin, dehydrated and mounted.

Assessment was based on overall staining intensity; microvessel staining; pattern of staining; and percentage of immunopositive cells. Pattern of immunoexpression of cells was either cell membrane type or combined membrane and cytoplasmic. The intensity of staining was graded semi-quantitatively as mild (+), moderate (++) and severe (+++), based on visual inspection. The percentage of immunopositive cells was divided into four groups (0-25%; 26-50%; 51-75%; 76-100%). Microvessel staining was evaluated as 0, + and ++ based on ≤33%, 34-66% and >67% of total vessels.

Proliferation index was evaluated using MIB-1 antibody staining (1 : 300, M/s Dakopatts, Denmark), and manually counting 1000 cells from at least five representative microscopic fields, at high-power magnification (400×).

A similar procedure, as described above, using mouse anti-human CD34 monoclonal antibody (dilution 1 : 200, M/S Dakopatts, Denmark) was followed. Microvessel density (MVD) was calculated by scanning the immunostained sections at low-power (×100) magnification, and selecting the area with the greatest number of distinctly CD34-immunostained microvessels ("hot spot"). The mean value of vessel count as evaluated under high-power magnification (×400) in five fields within the hot spot was defined as microvessel density (MVD). Image morphometry was performed by using a computerized digital photomicrograph system (Dewinter Optical Inc. with Digi Eye 330 digital photomicrography camera and Biowizard 4.2 Image analysis software, New Delhi).

Control tissues comprised brain sections from ten autopsy cases (three of ischemic stroke; two of hemorrhagic stroke; and five without any gross or microscopic brain pathology).


The present study consisted of 30 cases of primary brain tumor that were diagnosed during 2009-10, in which slides and blocks were available. The cases were divided into gliomas, meningiomas and other primary CNS tumors.

The mean age at surgery was 49 years (range: 16-70 years) for gliomas; 48.1 years (range: 16-72 years) for meningiomas; and 29.6 years (range: 8-58 years) for other neoplasms. There were a total of four cases in the pediatric age (<18 years) group (one each of oligoastrocytoma, fibroblastic meningioma, ependymoma and PNET).

Meningiomas showed a female preponderance (4 : 1) unlike other primary CNS tumors where majority were males (7 : 3). In gliomas the distribution was, however, equal.

All except one GBM and both the mixed gliomas showed contrast enhancement with variable degrees of perilesional edema. The solitary case of diffuse astrocytoma and a GBM (case No G6) showed mild perilesional edema despite lacking enhancement [Table 1]. Meningiomas were characterized by isodense contrast-enhancing dural masses. Perilesional edema was absent in all but a single case of fibroblastic meningioma [Table 2]. All high-grade tumors (PNET, germinoma and anaplastic ependymoma) showed increased enhancement with intense perilesional edema, while all the Grade I tumors, except hemangioblastoma, displayed variable enhancement without any perilesional edema. The single case of hemangioblastoma showed a rim of edema around the tumor [Table 3].{Table 1}{Table 2}{Table 3}

Seven of the ten gliomas were GBMs (World Health Organization (WHO) Grade IV) that were characterized by nuclear pleomorphism, frequent atypical mitoses, microvascular proliferation and necrosis. Of the remaining, one (anaplastic oligoastrocytoma) was WHO Grade III, and two (oligoastrocytoma and diffuse astrocytoma) were Grade II neoplasms. The mean MIB-1 labeling index (MIB-1 LI) was 11 (range: 2-18). All meningiomas were WHO Grade I tumors, none of which showed any evidence of anaplasia. The mean MIB-1 LI was 1.1 (range: <1-2). Majority of the other primary CNS tumors were WHO Grade I (two pituitary adenomas, and one each of DNET, craniopharyngioma and hemangioblastoma) neoplasm. Of the remaining, two were Grade II (ependymomas), one Grade III (anaplastic ependymoma) and two Grade IV neoplasms (PNET and germinoma). The mean MIB-1 LI of tumors in this group was 5.6 (range: 1-24).

Overall AQP1-immunoreactivity was present in 100% of the 'glioma' and 'other primary brain tumor' groups unlike meningiomas, where immunoreactivity was only 20%. The AQP1-expression in the intratumoral and peritumoral areas was 100% and 100% in gliomas; 20% and 20% in meningiomas; and 50% and 80% in other tumors. All the glial tumors and 60% of other neoplasms displayed immunopositivity in >50% cells unlike both meningiomas, where AQP1 upregulation was focal and seen in ≤50% cells. Expression in microvessels graded as 0, + and ++ was seen in 0, 40% and 60% respectively in gliomas; 100%, 0 and 0 respectively in meningiomas; and 50%, 30% and 20% respectively in other tumors [Table 1],[Table 2],[Table 3] and [Table 4].

In general, AQP1-immunoreactivity displayed specific patterns of expression. The intratumoral areas comprised large areas of homogenous immunostaining, where individual perikarya or cell processes could not be identified ("diffuse") [Figure 1]a. Lower grade tumors were characterized by cell membrane ("membrane") reactivity [Figure 1]b, while other tumors displayed positive immunostaining either in the cell processes and perikarya [Figure 1]c, or consisted of mixed expression both in the cell membrane and in the perikarya ("mixed") [Figure 1]d. In contrast, the peritumoral regions were characterized by a clustered ("bushy") appearance concentrated around the microvessels and within the reactive astrocytes [Figure 1]e and f. Both intratumoral and peritumoral areas, however, showed perivascular staining [Figure 1]g, with more intense expression in the latter. Interestingly, high-grade gliomas showed AQP1-staining around the vessel with sparing of the proliferating endothelial cells [Figure 1]h. Sections containing interface between normal brain and tumor periphery were conspicuous by the absence of AQP1-immunoexpression in the former and upregulation in the latter areas [Figure 1]i.

Low-grade gliomas showed higher intratumoral expression [Figure 2]a and b while GBMs showed intense peritumoural upregulation [Figure 2]c and d. GBMs showed "diffuse" "mixed" pattern of immunostaining while low-grade glioma showed predilection for "membrane" pattern. In the peritumoral location both showed perivascular clumping or "bushy" pattern along with AQP1-immunopositive reactive astrocytes. All six meningothelial meningiomas were negative for AQP1 either in the intratumoral or in the peritumoral areas. However, both cases of fibroblastic meningiomas showed mild to moderate "mixed" membrane and cytoplasmic positivity in the tumor cells [Figure 2]e, and also in the vascular endothelia in the intratumoral and peritumoral locations. All the other primary CNS tumors showed "mixed" staining pattern. Both ependymomas [Figure 2]f and g and hemangioblastoma showed intense intratumoral expression, in contrast to anaplastic ependymoma, PNET [Figure 2]h and germinoma, which displayed moderate to severe expression in the peritumoral region, while being largely absent within the tumor cells.{Figure 1}{Figure 2}{Table 4}

The MVD in gliomas, meningiomas, and other tumors ranged between 12.2-29.8 (mean : 18.72); 5-10 (mean : 8.04); and 9.8-35.6 (mean : 18.3). The mean MVD in Grade I, II, III, IV tumors was 8.3; 12.6; 16.8; and 19.4. Pituitary adenomas, despite being Grade I neoplasms expressed high MVD, owing to which the data were excluded while calculating mean MVD of Grade I tumors. In order to have a better predictive value, especially in low-grade tumors with high intratumoral vascular density, AQP1:MVD ratio was calculated. The AQP1 : MVD ratio in gliomas, meningiomas, and other tumors ranged between 0.56-0.83 (mean: 0.68); 0.23-0.32 (mean: 0.28); and 0.2-0.62 (mean: 0.39).

Overall it was noted that upregulation of Aquaporin1 was a prominent feature in all gliomas, unlike meningeal tumors. The expression of AQP1 correlated well with increasing grade of tumor, which was corroborated by histological features (anapalsia, mitoses, vascular proliferation, necrosis), higher MIB-1 labeling, contrast-enhancement and perilesional edema, and increasing AQP1 : MVD ratio. AQP1-upregulation in peritumoral areas of high-grade gliomas correlated with increased perilesional edema, unlike low-grade tumors. Though AQP1 expression and perilesional edema was not a distressing feature in most meningiomas, fibroblastic variants despite being of lower grade had a tendency to overexpress AQP1 and develop perilesional edema. Despite the restricted number of individual tumors in the third group, AQP1 expression in ependymal tumors and other high-grade neoplasms (PNET, germinoma) displayed a direct correlation with tumor grade, MIB-1LI, contrast-enhancement and perilesional edema, and AQP1 : MVD ratio similar to GBMs. Hemangioblastoma, a WHO Grade I tumor, interestingly showed an intense AQP1-immunoreactivity both in the intratumoral stromal cells while being largely negative in the microvessels.

Sections from all five autopsy cases with ischemic or hemorrhagic stroke displayed diffuse immunoreactivity both within the vascular channels in lesional and perilesional areas, and also in the reactive astrocytes [Figure 2]i. In contrast, sections from all five autopsy cases without any brain pathology were AQP1-immunonegative. The MVD in this subset ranged between 7.6-9.6.


Literature on the role of AQP as a marker in brain tumors is sparse with most studies focusing on AQP4 and that too in high-grade gliomas. [1],[2],[3],[6],[8],[9],[10]

Endo et al., performed one of the initial studies of AQP1 expression in gliobalstoma multiforme and breast cancer transplanted into mouse brain. [11] They noted that AQP1 expression was confined to the GBM cells and microvascular endothelial cells, and was absent in normal brain parenchyma and normal microvessel endothelium. Saadoun et al., further reconfirmed this observation. [9] This was comparable to the present study, where a similar pattern was detected in GBM, and in the peritumoral normal brain parenchyma and controls consisting of normal brain tissue. In addition, the current study also noted the expression pattern in meningiomas and other primary CNS tumors, which had not been included in most of the series available in the literature. Meningiomas were conspicuous by their lack of AQP1 expression, except in transitional variant, while ependymomas, PNET and germinomas displayed a pattern akin to GBM. In all primary CNS tumors AQP1 expression increased with increasing grade of tumor. This was similar to that observed by other workers. [7],[8],[9] Since all meningiomas included in the present study were Grade I neoplasms, further studies recruiting angiomatous, atypical and anaplastic variants, should be done for better correlation.

Increased expression of AQP1 with increasing grade of brain tumors is possibly a reflection of the impaired blood-brain barrier function with possible increased permeability to water, culminating in vasogenic edema. Vascular endothelial growth factor (VEGF) and HIF-1α are few of the various factors that may induce AQP1 expression in the endothelium of malignant CNS tumors. [9],[12] Pan et al. assessed AQP1 : vascular density ratio in normal, hyperplastic and carcinomatous endometrial biopsies and noted a positive correlation with development of malignancy. [13] They suggested that AQP1 promotes the genesis and progression of malignancy by promoting angiogenesis.

In addition it was also evident that the distribution pattern of AQP1 expression, viz. peritumoral in high-grade and intratumoral in low-grade neoplasms, had a tendency to correlate with the grade of tumor. Though Hayashi et al.,[1] noted this aspect in their series of gliomas, the current study confirmed a similar trend in other types of primary brain tumors as well.

While analyzing the AQP1-immunoreactivity pattern, Hayashi et al.[1] noted a "homogenous" and a perivascular "clustered" distribution. However, in the present study it was observed that AQP1 expression in the intratumoral location was either "diffuse" homogenous in high-grade glial tumors, with "combined" cell membrane and cytoplasmic positivity; or "diffuse" with predominantly "cell membrane" pattern in low-grade gliomas. The peritumoral region however displayed "bushy" or perivascular clustered appearance in all grades of tumor, with difference in the intensity of expression.

The present study thus highlighted a distinctive pattern of quantitative and qualitative expression of AQP1 in various types and grades of primary CNS malignancies, and attempted to establish the utility of AQP1-immunostaining as a marker for malignant potential of the tumor. Correlation of AQP1 expression with various histological features (pleomorphism, cellularity, mitoses, vascular proliferation, necrosis), proliferation index and neuroimaging profile of contrast enhancement and peritumoral edema further substantiates this trend.

Thus apart from having a pathogenetic and prognostic connotation, evaluation of AQP1 expression in brain tumors has a potential for serving as a target for future therapeutic intervention for effective management of vasogenic edema. [14] Moon et al., had sequenced AQP1 promoter molecules from human erythroleukemia cells, and noted that they possess steroid-responsive elements. [15] This possibly explains the anti-edema action of glucocorticoids, which are effective in managing vasogenic edema in brain tumors that result from alterations in AQP1 expression. Discovery of arylsulphonamide AqB013 as an antagonist for AQP1 had further paved the path for translational research in treatment of diseases based on overexpression of AQP. [5]

However, a complex interaction between various molecules, including AQP4, AQP9, HIF-1α, VEGF and angiopoietins, possibly acts as a deterrent in the development of an effective targeted therapy. [16],[17],[18],[19],[20],[21] A testimony to this fact was the AQP1 expression profile exhibited by hemangioblastoma. Despite an intense contrast enhancement, AQP1 was expressed within the tumor and there was lack of peritumoral edema. However, close scrutiny revealed that immunoreactivity was confined to the stromal cells and was not found in the microvessels, a feature which has often been reported in the literature where most cases were associated with cyst formation. [22],[23],[24] On the contrary, pituitary adenomas despite demonstrating intense contrast enhancement and high MVD (35.6) lacked AQP1-immunoreactivity, and correlated well with its Grade I status.

To summarize, this study shows a trend of positive correlation of the intensity and distribution pattern of AQP1 expression with increasing grade of primary brain tumors of various types, and possibly suggests a pathogenetic relationship between AQP1 and vasogenic brain edema in malignant primary brain tumors. However, more number of cases of various grades in each of the three groups needs to be evaluated along with other molecules of vasogenic edema, viz. AQP4, AQP9 and HIF-1α, in order to understand the complex interactions in the pathogenesis of vasogenic brain edema in malignant CNS neoplasms, which will be imperative for developing targeted pharmacological interventions in the future.


1Hayashi Y, Edwards NA, Proescholdt MA, Oldfield EH, Merrill MJ. Regulation and function of aquaporin-1 in glioma cells. Neoplasia 2007;9:777-87.
2Nag S, Manias JL, Stewart DJ. Pathology and new players in the pathogenesis of brain edema. Acta Neuropathol 2009;118:197-217.
3Magni F, Chinello C, Raimondo F, Mocarelli P, Kienle MG, Pitto M. AQP1 expression analysis in human diseases: Implications for proteomic characterization. Expert Rev Proteomics 2008;5:29-43.
4Verkman AS, Mitra AK. Structure and function of aquaporin water channels. Am J Physiol Renal Physiol 2000;278:F13-28.
5Yool AJ, Brown EA, Flynn GA. Roles for novel pharmacological blockers of aquaporins in the treatment of brain oedema and cancer. Clin Exp Pharmacol Physiol 2010;37:403-9.
6Papadopoulos M, Saadoun S, Krishna S, Bell B, Davies D. The aquaporin-1 water channel protein is abnormally expressed in oedematous human brain tumours. J Anat 2002;200:531-2.
7Boon K, Edwards JB, Eberhart CG, Riggins GJ. Identification of astrocytoma associated genes including cell surface markers. BMC Cancer 2004;4:39-46.
8Oshio K, Binder DK, Liang Y, Bollen A, Feuerstein B, Berger MS, et al. Expression of the aquaporin-1 water channel in human glial tumors. Neurosurgery 2005;56:375-81.
9Saadoun S, Papadopoulos MC, Davies DC, Bell BA, Krishna S. Increased aquaporin 1 water channel expression in human brain tumours. Br J Cancer 2002;87:621-3.
10Dua RK, Devi I, Yasha TC. Increased expression of Aquaporin-4 and its correlation with contrast enhancement and perilesional edema in brain tumours. Br J Neurosurg 2010;24:454-9.
11Endo M, Jain RK, Witwer B, Brown D. Water channel (aquaporin 1) expression and distribution in mammary carcinomas and glioblastomas. Microvasc Res 1999;58:89-98.
12Zagzag D, Zhong H, Scalzitti JM, Laughner E, Simons JW, Semenza GL. Expression of hypoxia-inducible factor 1alpha in brain tumors: Association with angiogenesis, invasion, and progression. Cancer 2000;88:2606-18.
13Pan H, Sun C, Zhou C, Huang H. Expression of aquaporin-1 in normal, hyperplastic and carcinomatous endometrial. Int J Gynaecol Obstet 2008;101:239-44.
14Moon C, King LS, Agre P. Aqp1 expression in erythroleukemia cells: Genetic regulation of glucocorticoid and chemical induction. Am J Physiol 1997;273:C1562-70.
15Oshio K, Binder DK, Bollen A, Verkman AS, Berger MS, Manley GT. Aquaporin-1 expression in human glial tumors suggests a potential novel therapeutic target for tumor-associated edema. Acta Neurochir Suppl 2003;86:499-502.
16Badaut J, Brunet JF, Grollimund L, Hamou MF, Magistretti PJ, Villemure JG, et al. Aquaporin 1 and aquaporin 4 expression in human brain after subarachnoid hemorrhage and in peritumoral tissue. Acta Neurochir Suppl 2003;86:495-8.
17Badaut J, Lasbennes F, Magistretti P, Regli L. Aquaporins in brain: Distribution, physiology, and pathophysiology. J Cereb Blood Flow Metab 2002;22:367-8.
18McCoy E, Sontheimer H. Expression and function of water channels (aquaporins) in migrating malignant astrocytes. Glia 2007;55:1034-43.
19Ng WH, Hy JW, Tan WL, Liew D, Lim T, Ang BT, et al. Aquaporin-4 expression is increased in edematous meningiomas. J Clin Neuroscience 2009;16:441-3.
20Tan G, Sun SQ, Yuan DLJ. Expression of the water channel protein aquaporin-9 in human astrocytic tumours: Correlation with pathological grade. Int Med Res 2008;36:777-82.
21Xu M, Su W, Xu QP. Aquaporin-4 and traumatic brain edema. Chin J Traumatol 2010;13:103-10.
22Takei H, Powell SZ. Novel immunohistochemical markers in the diagnosis of nonglial tumors of nervous system. Adv Anat Pathol 2010;17:150-3.
23Chen Y, Tachibana O, Oda M, Xu R, Hamada J, Yamashita J, et al. Increased expression of aquaporin 1 in human hemangioblastomas and its correlation with cyst formation. J Neurooncol 2006;80:219-25.
24Longatti P, Basaldella L, Orvieto E, Dei Tos AP, Martinuzzi A. Aquaporin 1expression in cystic hemangioblastomas. Neurosci Lett 2006;392:178-80.