Proinflammatory role of aquaporin-4 in autoimmune neuroinflammation

The FASEB Journal • Research Communication

Proinflammatory role of aquaporin-4 in autoimmuneneuroinflammation

Lihua Li,*,† Hua Zhang,*,† Michel Varrin-Doyer,‡ Scott S. Zamvil,‡ and A. S. Verkman*,†,1

*Department of Medicine, †Department of Physiology, and ‡Department of Neurology, University ofCalifornia, San Francisco, California, USA

ABSTRACT Aquaporin-4 (AQP4) deficiency in micereduces neuroinflammation in experimental autoim-mune encephalomyelitis (EAE) produced by active im-munization with myelin oligodendrocyte glycoproteinpeptide (MOG). Potential mechanisms for the protec-tive effect of AQP4 deficiency were investigated, in-cluding AQP4-dependent leukocyte and microglia cellfunction, immune cell entry in the central nervoussystem (CNS), intrinsic neuroinflammation, and hu-moral immune response. As we found with active-immunization EAE, neuroinflammation was greatly re-duced in AQP4-knockout mice in adoptive-transferEAE. AQP4 was absent in immune cells, includingactivated T lymphocytes. The CNS migration of fluo-rescently labeled, MOG-sensitized T lymphocytes wascomparable in wild-type and AQP4-knockout mice.Microglia did not express AQP4. Serum anti-AQP4antibodies were absent in EAE. Remarkably, intracere-bral injection of LPS produced much greater neuroin-flammation in wild-type than in AQP4-knockout mice,and cytokine (TNF-� and IL-6) secretion was reducedin astrocyte cultures from AQP4-knockout mice. Ade-novirus-mediated expression of AQP4, or of an unre-lated aquaporin, AQP1, increased cytokine secretion inastrocyte and nonastrocyte cell cultures, supporting theinvolvement of aquaporin water permeability in cyto-kine secretion. Our data suggest an intrinsic proinflam-matory role of AQP4 involving AQP4-dependent astro-cyte swelling and cytokine release. Reduction in AQP4water transport may be protective in neuroinflamma-tory CNS diseases.—Li, L., Zhang, H., Varrin-Doyer,M., Zamvil, S. S., Verkman, A. S. Proinflammatory roleof aquaporin-4 in autoimmune neuroinflammation.FASEB J. 25, 000–000 (2011). www.fasebj.org

Key Words: water channel � neuromyelitis optics � astrocyte� EAE

Experimental autoimmune encephalomyelitis (EAE)is an extensively used model to study the patho-genesis of neuroinflammatory demyelinating diseases,such as multiple sclerosis, and to evaluate potentialtherapies (1). Many EAE variants using different animalmodels and immunization approaches have been de-veloped that produce, to differing extents, characteris-tic neuroinflammatory lesions in the central nervous

system (CNS) with demyelination and clinical motordysfunction. EAE is mediated by myelin-specific Th1 orTh17 cells (2, 3), although a humoral response is alsoimportant in certain models (4, 5).

Motivated by the discovery that aquaporin-4 (AQP4)is the target antigen in the multiple sclerosis variantneuromyelitis optica (NMO; ref. 6) and that AQP4autoantibodies (NMO-IgG) are involved in NMO dis-ease pathogenesis (7–9), we recently investigated theAQP4 dependence of EAE severity. AQP4 is a water-selective membrane transport protein expressed inastrocytes throughout the CNS that is involved in brainwater balance, neuroexcitation and astrocyte migration(10). We found that compared with wild-type mice,AQP4-knockout mice showed remarkably attenuatedEAE following active immunization with myelin oligo-dendrocyte glycoprotein (MOG35–55) peptide, with re-duced motor dysfunction, CNS inflammation, and de-myelination (11).

The purpose of this study was to investigate potentialmechanisms responsible for the neuroprotective effectof AQP4 deficiency in EAE. We systematically investi-gated the various steps in EAE pathogenesis involvingimmune cell function, T-lymphocyte penetration intothe CNS, and the consequent CNS response, as well asAQP4-dependent humoral responses. Our studies weredone using wild-type and AQP4-null mice bred on theC57BL/6 genetic background. AQP4-null mice are notdifferent from wild-type mice in their survival, growth,and behavior, CNS gross and microscopic anatomy, andbasal intracranial pressure and blood-brain barrier in-tegrity (12–14). Biophysical studies indicated a mildlyexpanded extracellular space in the brain cortex ofAQP4-null mice at baseline (15, 16). AQP4-null micemanifest significant phenotypes in response to stresses,including reduced cytotoxic brain swelling in waterintoxication, stroke, and bacterial meningitis (12, 17);increased vasogenic brain swelling in tumor, abscess,and hydrocephalus (13, 18); reduced astrocyte migra-tion and glial scarring following injury (19, 20); andprolonged seizure and cortical spreading neuroexcita-tion following mechanical or chemical stimuli (21).

1 Correspondence: 1246 Health Sciences East Tower,UCSF, Box 0521, San Francisco, CA 94143-0521, USA. E-mail:[email protected]

doi: 10.1096/fj.10-177279

10892-6638/11/0025-0001 © FASEB

The FASEB Journal article fj.10-177279. Published online January 21, 2011.

Each of these phenotypes is likely explicable on thebasis of reduced astrocyte cell osmotic water permea-bility in AQP4 deficiency but provides no useful cluesabout the mechanisms of reduced EAE in AQP4 defi-ciency.

We found here that the neuroprotective effect ofAQP4 deficiency in EAE cannot be explained by AQP4-dependent differences in immune cell function or CNSpenetration, or microglia cell function or humoralresponse. We found evidence, however, for an intrinsicproinflammatory role for AQP4 in the CNS, which mayhave broad implications to CNS diseases associated withinflammation.

MATERIALS AND METHODS

Mice

AQP4-null mice in a C57BL/6 genetic background wereobtained by �10 back-crosses of AQP4-null mice generatedoriginally by targeted gene disruption in a CD1 background(22). Studies were performed on 8-to-10-wk-old weight-matched, female C57BL/6 wild-type and AQP4-null mice.Mice were maintained in air-filtered cages and fed normalmouse chow in the University of California, San Francisco(UCSF) Animal Care facility. All procedures were approvedby the UCSF Committee on Animal Research.

Active-immunization and adoptive-transfer EAE

MOG35–55 peptide (NH2-MEVGWYRSPFSRVVHLYRNGK-COOH), �98% pure, was synthesized by Biomatik (Wilming-ton, DE, USA). For active-immunization EAE, MOG35–55 wasdissolved in PBS at 4 mg/ml, then 1:1 emulsified withcomplete Freund’s adjuvant supplemented with 4 mg/mlheat-inactivated Mycobacterium tuberculosis H37Ra (Sigma, St.Louis, MO, USA). According to standard procedures (11),mice were injected with 0.1 ml of the emulsion subcutane-ously, distributed over 3 sites along the midline of the backbetween the shoulders. Mice received 200 ng pertussis toxin(Sigma) in 200 �l PBS intraperitoneally at the time ofimmunization and 2 d later. Mice were assessed daily forclinical signs using the following scoring: 0, normal mouse,no signs of disease; 1, limp tail or hindlimb weakness, but notboth; 2, limp tail and hindlimb weakness; 3, partial hindlimbparalysis; and 4, complete hindlimb paralysis. For adoptive-transfer EAE, mice were immunized as described above. At10–14 d, the draining lymph nodes were isolated, and col-lected cells were resuspended at 6 � 106 cells/ml in RPMI1640 medium supplemented with 10% FBS, penicillin/streptomycin, l-glutamine, sodium pyruvate, nonessentialamino acids, and �-mercaptoethanol, according to standardprocedures (23). MOG35–55 was added at 10 �g/ml togetherwith recombinant murine IL-12 (Sigma) at 5 ng/ml andrecombinant human IL-2 (Sigma) at 50 U/ml. After 3 d inculture (37°C, 5% CO2), collected cells were suspended inHBSS for transfer. Mice received 2 � 107 viable cells byretro-orbital sinus injection. Pertussis toxin (200 ng) in 200 �lPBS was given intraperitoneally just after cell transfer andagain 48 h later. Clinical score was assessed daily as describedabove. After adoptive transfer, at d 15, brain and spinal cordwere removed for staining with hematoxylin and eosin(H&E), Luxor fast blue (for myelin), and CD45 immunore-activity. H&E and CD45-stained sections were scored for theseverity of inflammation using the following scale: 0, no

inflammation; 1, mild inflammation with few mononuclearcells infiltrates; 2, marked inflammation with multiple infil-trates per �100 field; and 3, severe inflammation with exten-sive infiltrates in both white and gray mater.

Intracerebral LPS

Mice were anesthetized by intraperitoneal 2,2,2-tribromoeth-anol (125 mg/kg, Sigma), and the head was immobilized in astereotactic frame (Benchmark, Neurolab, St. Louis, MO,USA). Core temperature was maintained at 37–38°C using aheating lamp and rectal temperature probe. The skin abovethe brain was shaved and disinfected with betadine. A midlinescalp incision was made to expose the bregma and lambda. Aburr hole was made on the right side, at a location 1.2 mmposterior to bregma and 1 mm right lateral to midline usinga high-speed drill (0.7 mm burr; Foredom, Bethel, CT, USA).LPS (Escherichia coli 0111:B4; Sigma) at 1 �g/�l in sterile PBSwas injected at a depth of 2.5 mm through a 30-gauge needleattached to a 10-�l Hamilton syringe (Hamilton, Reno, NV,USA). The injection volume was 2 �l, and the needle was keptin place for 5–10 min after injection. The scalp was closedwith 5-0 silk suture. Mice were killed 24 h later by anestheticoverdose and underwent transcardiac perfusion with PBSfollowed by 4% paraformaldehyde. Brains were removed,fixed in 4% paraformaldehyde for 24 h, and embedded inparaffin. In some experiments, at 2 h after LPS injection,brains were homogenized in ice-cold lysis buffer containing25 mM HEPES (pH 7.4), 0.1% 3-[(3-cholamidopropyl) di-methyl-ammonio]1-propanesulfonate; 5 mM MgCl2; 1.3 mMEDTA; 1 mM EGTA; 10 �g/ml pepstatin, aprotinin andleupeptin; and 1 mM PMSF. Homogenates were centrifuged(15 min at 50,000 rpm) and stored at �80°C for mouse TNF-�ELISA assay (Invitrogen, Carlsbad, CA, USA).

RT-PCR and immunostaining

For RT-PCR, mice (control and MOG sensitized) were killedby anesthetic overdose, and lymph nodes and spleen wereremoved. Total RNA was isolated by homogenization inTRIzol reagent (Invitrogen). cDNA was reverse-transcribedfrom 5 �g mRNA with oligo(dT) (Super-Script first-strandsynthesis system for RT-PCR; Invitrogen), using kidney as apositive control. After reverse transcription, PCR was carriedout using gene-specific primers designed to amplify the AQP4coding sequence. Fluorescence-based real-time reverse tran-scription-PCR (RT-PCR) using 2 �g cDNA from astrocytecultures was carried out using a LightCycler with FastStartDNA MasterPLUS SYBR Green I kit (Roche Diagnostics,Indianapolis, IN, USA). PCR conditions comprised an initialstep at 95°C for 5 min followed by 40 cycles at 95°C for 10 s, 60°Cfor 10 s, and 72°C for 10 s. Primers were as follows: �-actin:5�-TGTATGCCTCTGGTCGTACC-3� (sense), 5�-CAGGTCC-AGACGCAGGATG-3� (antisense); TLR4: 5�-CAAGTTTAG-AGAATCTGGTGGCTGTGG-3� (sense), 5 � -TGAAA-GGCTTGGTCTTGAAT GAAGTCA-3� (antisense); TNF-�:5�-CTGTAGCCCACGTCGTAGC-3� (sense), 5�-TTGAGA-TCCATGCCGTTG-3� (antisense); NF-�B: 5�-CAGCTCTTC-TCAAAGCAGCA-3� (sense), 5�-TCCAGGTCATAGAGAGGC-TCA-3� (antisense). �-Actin was used as the reference gene,and pooled wild-type and AQP4-null astrocyte cDNA wasused as the calibrator. Results are reported as normalized,calibrated ratios.

For immunofluorescence, tissues and astrocyte cultureswere fixed in the 4% paraformaldehyde. Paraffin sectionswere stained with H&E or Luxol fast blue. For CD45 immu-nocytochemistry, paraffin-embedded sections were cut at5-�m thickness and deparaffinized, then treated with citrate

2 Vol. 25 May 2011 LI ET AL.The FASEB Journal � www.fasebj.org

buffer using microwave antigen retrieval and 3% hydrogenperoxide. Sections were incubated for 2 h with a primaryCD45 antibody (1:200; Abcam, Cambridge, MA, USA) atroom temperature, then incubated with biotinylated second-ary goat anti-rabbit antibody (1:1000; Vector Laboratories,Burlingame, CA, USA), followed by avidin-biotin peroxidasecomplex (1:1000). Peroxidase labeling was visualized withdiaminobenzidine to yield a brown color. For Iba-1, AQP4,and glial fibrillary acidic protein (GFAP) immunofluores-cence, sections or cells were incubated with rabbit anti-Iba-1(1:500; Wako, Osaka, Japan), rabbit anti-AQP4 (Santa CruzBiotechnology, Santa Cruz, CA, USA; 1:200) or mouse anti-GFAP (1:500; Millipore, Billerica, MA, USA) primary anti-body and goat anti-rabbit or anti-mouse secondary antibody(1:200; Molecular Probes, Eugene, OR, USA). Cell nucleiwere stained blue with DAPI.

CNS penetration of fluorescently labeled, MOG-sensitizedT lymphocytes

To fluorescently label lymphocytes, a stock solution of car-boxyfluorescein succinimidyl ester (CFSE; 10 mM in DMSO,Invitrogen) was diluted in PBS to 25 �M. Lymphocytes, ascultured and activated for adoptive-transfer EAE, were sus-pended in PBS containing 1% BSA at 2 � 106 cells/ml andincubated with CFSE for 15 min at 37°C. Cells were washedand resuspended in culture medium for 15 min, and thenresuspended in PBS on ice. Gentle trituration was used tominimize cell clumping. Cells were injected into wild-typeand AQP4-null mice as done for adoptive-transfer EAE.Brains were harvested at 24 and 48 h after intracardiacperfusion with PBS and 10-�m-thick frozen sections were cutfor counting of fluorescent cells.

Blood-brain barrier permeability

Control and MOG-immunized mice (after 12 d) were injectedintravenously with Evans blue dye (4% in PBS, 160 mg/kg,Sigma). After 90 min, mice were anesthetized, and the leftcardiac ventricle was perfused with 20 ml PBS. Brains wereremoved and immersed in 1 ml formamide (Sigma) at 55°Covernight to extract the Evans blue dye. Extracted dye wasquantified by optical absorbance at 610 nm against Evansblue/formamide standards.

Lymphocyte proliferation assay

Splenic CD3 T lymphocytes from MOG35–55-specific T-lym-phocyte receptor transgenic (2D2) mice were purified(�95%) by MACS (Miltenyi Biotec, Bergisch Gladbach, Ger-many), then cultured with irradiated splenocytes from naivewild-type or AQP4-deficient mice with different concentra-tions of MOG35–55. At 72 h, cultures were pulsed with 1 �Ci[3H]-thymidine and harvested 18 h later for measurement of[3H]-thymidine incorporation.

Anti-AQP4 autoantibody assay

Wild-type mice were immunized with MOG35–55 peptide. Atspecified times, blood was collected for serum isolation. Forimmunofluorescence, FRT-AQP4 cells were fixed in 4% para-formaldehyde for 15 min, and blocked with 5% BSA for 1 h.Cells were incubated with EAE (or control) serum at 1:50 or1:200 for 2 h at room temperature, then incubated with goatanti-mouse secondary antibody (1:200, Molecular Probes). Asa positive control, cells were immunostained with humanserum from a seropositive NMO patient who met the estab-

lished criteria for diagnosis of NMO, using anti-human sec-ondary antibody (1:200; Molecular Probes).

Cell culture and adenovirus infection

Primary astrocyte cultures were prepared from cortex ofwild-type and AQP4-null neonatal mice, as described previ-ously (24). Briefly, the cerebral hemispheres were isolated,the meninges were dissected away, and the hippocampus,basal ganglia, and olfactory bulb were removed. Corticaltissues were minced with forceps and incubated for 15 min at37°C in DMEM containing 0.25% trypsin-EDTA. Dissociatedcells were centrifuged at 500 g for 5 min and resuspended inDMEM containing 10% FBS and 1% penicillin/streptomycinin polylysine-coated 75 cm2 flasks and grown at 37°C in a 5%CO2 incubator with a change of medium 2�/wk. After cellconfluence (d 8–10), flasks were shaken in rotator at 180 rpmovernight, then at 220 rpm for 4 h to purify astrocyes. Atconfluence (d 12–15), cultures were treated with 10 �Mcytosine arabinoside for 48 h to prevent proliferation of othercell types, and the medium was replaced with DMEM contain-ing 3% FBS and 0.15 mM dibutyryl cAMP to induce differen-tiation. Cultures were maintained for up to 2 wk longer witha change of medium 2�/wk. Immunocytochemistry showedthat �95% of the cells stained positively for the astrocyticmarker, GFAP. T24 cells (human bladder carcinoma cell line,HTB-4; American Type Culture Collection, Manassas, VA,USA) were cultured at 37°C in 5% CO2 in complete DMEMmedium containing 10% FBS, 100 U/ml penicillin, and l00�g/ml streptomycin. Fisher rat thyroid (FRT) cells stablytransfected with AQP4 were cultured at 37°C (5% CO2) withCoon’s F-12 growth medium containing 10% FBS, 10 U/mlpenicillin, 100 �g/ml streptomycin, 4 mM L-glutamine, and0.6 mg/ml zeocin.

Adenoviruses containing full-length coding sequences ofAQP1, M1-AQP4, M23-AQP4, and green fluorescent protein(GFP) were generated by Vector Biolaboratories. For virusinfection, AQP4-null astrocyte cell cultures were seeded on24-well plates. After 1 wk, cells were incubated in 1 ml DMEMmedium with 5% FBS containing adenovirus at a multiplicityof infection (MOI) of 20, which was determined to give strongexpression without affecting cell viability. The virus-contain-ing medium was removed 3 h later and replaced with freshDMEM medium containing 10% FBS. Cells were used 2 or 3 dlater. For infection of T24 cells, cells on 24-well plates at 50%confluence were incubated for 36 h in 1 ml DMEM mediumwith 5% FBS containing adenovirus at a MOI of 500, then themedium was replaced with fresh DMEM/H-16 medium con-taining 10% FBS.

In some experiments, cytokine secretion was induced incell cultures by a 2-h incubation with LPS (100 ng/ml) with orwithout 20 mM d-mannitol, or with manganese (III) acetatedihydrate (Sigma). Supernatants and/or cells were collectedand kept at �80°C for cytokine assay. Protein content wasassayed by the bicinchoninic (BCA) procedure, and tumornecrosis factor � (TNF-�) and interleukin-6 (IL-6) wereassayed by ELISA (BD Bioscience, San Jose, CA, USA).

RESULTS

Adoptive-transfer EAE

Our previous study of EAE in AQP4 deficiency involvedEAE production by active immunization with MOG35–55oligopeptide. Here, we first determined whether AQP4deficiency is also protective in an adoptive-transfer

3PRONEUROINFLAMMATORY ROLE OF AQUAPORIN-4

model of EAE in which sensitized T lymphocytes fromMOG-treated wild-type mice were cultured and sensi-tized in vitro, and then introduced intravenously intonaive wild-type or AQP4-null recipient mice. As shownin Fig. 1A, clinical signs of EAE were seen by 6 d afterintravenous transfer. EAE disease severity, as assessedby clinical score, was much reduced in AQP4-null vs.wild-type mice. Representative histology in Fig. 1B (leftpanels) shows greater inflammation in spinal cord ofwild-type mice at 10 d after T-lymphocyte transfer.Figure 1B shows corresponding significant loss of my-elin (middle panels), and that the inflammatory cellinfiltrates are primarily CD45-positive mononuclearcells (right panels), as expected in EAE. Histologyscores are summarized in Fig. 1C. These experimentsshow that AQP4 deficiency in recipient mice is protec-tive in adoptive-transfer EAE.

AQP4 expression in immune cells

We investigated whether AQP4 expression in immunecells might account for the reduced severity of EAE inAQP4-null mice. AQP4 expression was studied by RT-PCR and immunofluorescence. Figure 2A shows nodetectable AQP4 transcript by RT-PCR of T lympho-cytes before or after in vitro culture, as done in adop-tive-transfer EAE. Mouse kidney cDNA was the positivecontrol. Figure 2B shows no detectable AQP4 protein inT lymphocytes by immunofluorescence, with AQP4-expressing Chinese hamster ovary (CHO) cells as apositive control. Figure 2C shows no AQP4 immunofluo-rescence in lymph nodes from control or MOG-immu-nized mice (as harvested for adoptive-transfer EAE).AQP4 immunofluorescence of mouse kidney was thepositive control. Lymph node size in untreated wild-

type and AQP4-null mice was comparable, as was theincrease in lymph node size after MOG treatment (datanot shown). Further, splenic antigen-presenting cellsfrom wild-type and AQP4-null mice presentedMOG35–55 to naive 2D2 T lymphocytes with comparableefficiency (Fig. 2D). These results indicate that AQP4expression in immune cells is not responsible for theattenuated EAE in AQP4 deficiency.

CNS penetration of MOG-sensitized T lymphocytes

The possibility was tested that reduced immune cellpenetration into the CNS might be responsible for thereduced severity of EAE in AQP4-null mice. Activated Tlymphocytes are thought to efficiently permeate theblood-brain barrier by mechanisms involving specificreceptors (25). We generated MOG-sensitized T lym-phocytes from wild-type mice in vitro, as done foradoptive-transfer EAE, which were labeled with a greenfluorescent dye and injected intravenously in wild-typeand AQP4-null mice. The presence of fluorescentlylabeled T lymphocytes in brain was measured at 24 and48 h after injection, examining 10-�m-thick frozenbrain sections. Figure 3A shows the fluorescent, MOG-activated T lymphocytes used for injection, along with arepresentative frozen section of brain containing thefluorescent cells. Figure 3B shows no significant differ-ence in the number of fluorescent T lymphocytescounted in brain slices at 24 or 48 h after injection.

Blood-brain barrier integrity was also examined. Theblood-brain barrier at baseline was comparably tight inwild-type and AQP4-null mice as assessed by Evan’s bluedye extravasation (Fig. 3C), in agreement with priorresults for wild type vs. AQP4-null mice in a CD1genetic background (12, 14). Blood-brain barrier per-

Figure 1. Attenuated adoptive-transfer EAE in AQP4-null mice. A) Activated, MOG-sensitized T lymphocytes from wild-type micewere transferred to naive wild-type (/) or AQP4-null (�/�) mice. Means se of clinical scores (4 mice/group).B) Representative spinal cord H&E staining (left panels), Luxol fast blue staining (middle panels), and CD45 immunocyto-chemistry (right panels) at 15 d after T-lymphocyte transfer into wild-type or AQP4-null mice. Arrows denote EAE lesion.C) Inflammatory score determined by masked assessment of H&E and CD45 sections (se, 3 mice/group, 5 sectionsexamined/mouse). *P � 0.001.

4 Vol. 25 May 2011 LI ET AL.The FASEB Journal � www.fasebj.org

meability was also not different in wild-type vs. AQP4-null mice when measured at a relatively early time (12d) after active immunization with MOG35–55 oligopep-tide, prior to development of clinical weakness. Thesestudies suggest that differences in blood-brain barrierproperties cannot account for the reduced EAE severityin AQP4 deficiency.

AQP4 humoral response

On the basis of the finding that AQP4 autoantibodiesproduce NMO-like neuroinflammatory lesions wheninjected directly into brain of naive mice (9) andexacerbate neuroinflammatory lesions in rats with pre-existing EAE (7, 8), we tested whether AQP4 autoanti-

Figure 2. Absence of AQP4expression in immune cells.A) RNA was isolated from Tlymphocytes, without (con-trol) or after MOG sensitiza-tion (as done for adoptive-transfer EAE), and frommouse kidney (positive con-trol). AQP4 and �-actin wereamplified by RT-PCR usingspecific primers. Representativeof 3 sets of amplifications on dif-ferent cultures. B) AQP4 immu-nofluorescence in T lymphocytes,without or after MOG sensitiza-tion. AQP4-expressing Chinesehamster ovary (CHO) cells(CHO-AQP4) as the positive con-

trol and nontransfected CHO cells (CHO) as the negative control. C) CD45 and AQP4 immunofluorescence in lymph nodes ofcontrol mice and at 15 d after MOG immunization. Mouse kidney shown as a positive control for AQP4. D) Splenocytes from wild-typeand AQP4-deficient mice stimulate proliferation of MOG35–55-specific T lymphocyte in a similar manner. MACS-separated 2D2 Tlymphocytes were cultured with irradiated splenic antigen-presenting cells from naive wild-type or AQP4-deficient mice in the presenceof MOG35–55. Proliferation was measured by [3H]-thymidine incorporation (sd, 3 sets of cultures, differences not significant).

Figure 3. AQP4 deletion does not alter penetration of T lymphocytes into the CNS. MOG-sensitized T lymphocytes (as preparedfor adoptive-transfer EAE) were fluorescently labeled with CFSE. Naive wild-type and AQP4-null mice were injectedintravenously with 2 � 107 labeled lymphocytes, and brains were perfused and harvested at 24 and 48 h. A) Left panel:fluorescently labeled MOG-activated T lymphocytes. Right panel: representative frozen section of brain, showing fluorescent Tlymphocytes that crossed the blood-brain barrier. B) Number of fluorescent T lymphocytes counted in 10-�m-thick brainsections per mouse at 24 and 48 h (se, 3 mice/group at each time point, differences not significant). C) Extravasated Evans bluedye in brain in control mice and at 12 d after MOG immunization (se, 4 mice/group at each time point, differences notsignificant).

5PRONEUROINFLAMMATORY ROLE OF AQUAPORIN-4

bodies are produced in EAE in mice. EAE-associatedAQP4 autoantibodies in wild-type mice could contrib-ute to their greater EAE disease severity. Serum frommice was obtained at 13 d after active immunizationwith MOG, at which time significant neurological im-pairment was seen (Fig. 4A). The presence of AQP4autoantibodies was assayed using FRT cells overexpress-ing mouse AQP4 by transient transfection. Figure 4Bshows no detectable AQP4 autoantibodies in control orEAE sera. All serum samples from immunized andcontrol mice were negative, tested up to a 1:50 dilution.As a positive control a 1:200 dilution of serum from anNMO patient produced strong immunofluorescence inthe AQP4-expressing cells but not in nontransfectedcells.

Intrinsic CNS inflammatory response

We tested whether intrinsic differences in the CNS hostresponse to inflammatory stimuli might be responsiblefor the reduced severity of EAE in AQP4 deficiency.Neuroinflammation was produced by injection of LPSdirectly into brain parenchyma in order to bypass theblood-brain barrier penetration step. Neuroinflamma-tion was assessed at 24 h after LPS injection. Figure 5Ashows brain H&E staining and CD45 immunocyto-chemistry in control (saline-injected) and LPS-injectedmice. Extensive leukocyte infiltrates were seen in LPS-injected wild-type mice, which were relatively infre-quent in AQP4-null mice. Figure 5B shows a gallery ofbrain sections from 3 wild-type and 3 AQP4-null mice.Figure 5C summarizes inflammation scores obtained bytwo independent observers who assessed the H&E andCD45-stained sections in a masked procedure. Therewas significantly greater inflammation in the LPS-treated wild-type mice than the LPS-treated AQP4-nullmice. A small amount of inflammation was seen incontrol mice as a consequence of the needle insertion.

A prior report suggested the expression of AQP4 inmicroglia after LPS administration (26), which, if cor-rect, would raise the possibility of AQP4-dependent

microglia cell function as a determinant of EAE sever-ity. Using Iba-1 as a selective marker of microglia, wefound microglial cell activation in LPS-treated mice,greater in wild-type than AQP4-null mice (Fig. 6A), asexpected from their greater neuroinflammatory re-sponse. However, the expression patterns of AQP4and Iba-1 were nonoverlapping, indicating absence ofAQP4 expression in microglia. Absence of AQP4 ex-pression in microglia was also found in mice followingactive-immunization and adoptive-transfer EAE (datanot shown). The absence of AQP4 in microglia indi-cates that AQP4-dependent differences in intrinsic mi-croglia cell responses cannot account for the reducedEAE severity in AQP4 deficiency. This conclusion wassupported by measurements of the release of a cytokine(TNF-�) in brain at 3 h after LPS administration. At thisshort time, TNF-� release is primarily from microglialcells and is minimally confounded by secondary effectsof neuroinflammation, such as astrocyte activation,local blood-brain barrier breakdown, and cell extrava-sation. Figure 6B shows significant TNF-� elevation inLPS vs. saline-injected brain, but no differences in wildtype vs. AQP4-null mice.

Cytokine secretion in cell cultures

To investigate the possible involvement of AQP4 in theinflammatory response of astrocytes, measurements ofcytokine release were made on differentiated primaryastrocyte cultures from brain cortex of neonatal wild-type and AQP4-null mice. Figure 7A shows similarmorphology and GFAP immunoreactivity of the cul-tures (left panels), with strong AQP4 expression in thecultures from wild-type mice (right panels). Cultureswere judged to be free of microglia by Iba-1 staining(not shown). Initial screening of a panel of 12 cytokinesshowed robust elevations in TNF-� and IL-6 at 2 h afterLPS addition in astrocyte cultures from wild-type mice.TNF-� and IL-6 were, therefore, measured in subse-quent experiments. Figure 7B shows significantly greaterrelease of both TNF-� and IL-6 into the medium of

Figure 4. Absence of anti-AQP4 autoantibodies in EAE. A) Active-immunization EAE was produced in wild-type mice byMOG35–55 peptide. Clinical score. B) AQP4-expressing FRT cells (FRT-AQP4) and control nontransfected cells (FRT) stainedwith serum from control and EAE mice, and green fluorescent secondary anti-mouse antibody. Positive control is human serumfrom patient with NMO (EAE serum). AQP4 stained red. Representative of 3 sets of staining studies.

6 Vol. 25 May 2011 LI ET AL.The FASEB Journal � www.fasebj.org

cultures from wild-type vs. AQP4-null mice at 2 h afterLPS addition, with the levels of these cytokines near thelimit of detection in the absence of LPS. This was arobust finding seen in 5 separate sets of cultures.However, cell-associated TNF-� and IL-6 were not dif-ferent in cultures from wild-type vs. AQP4-null mice at2 h after LPS (Fig. 7C, left and middle panels). TheAQP4-dependent difference in cytokine release was notrelated to reduced TLR-4 expression in the AQP4-deficient astrocytes, as demonstrated by quantitativeRT-PCR (Fig. 7C, right panel) and immunofluores-cence (not shown). Figure 7C (right panel) also showscomparable increases in TNF-� and NF-�B transcript

expression following LPS. These data suggest AQP4involvement in cytokine release/secretion rather thanin transcription or translation. Reduced TNF-� releasewas also found at 24 h after addition of Mn3, whichcauses cytokine release in a TLR-4-independent man-ner (Fig. 7D, right panel). We found greater swelling inMn3-treated astrocytes from wild-type vs. AQP4-nullmice (Fig. 7D, left panels), suggesting a possible mech-anism for AQP4-dependent cytokine release involvingastrocyte water permeability and cell swelling. In sup-port of a water permeability/cell-swelling mechanismwas the finding of reduced TNF-� and IL-6 release fromastrocytes when the culture medium was made mildly

Figure 5. Reduced brain inflammation in AQP4-null mice after intracerebral LPS injection.Mice were injected intracerebrally with PBS (control) or LPS. A) Brain histology at 1 d by H&Estaining (top panels) and CD45 immunocytochemistry (bottom panels). B) Gallery ofCD45-immunostained sections of hippocampus from 3 wild-type and 3 AQP4-null mice.C) Quantification of inflammation done using H&E and CD45-stained sections (se). *P �0.01 vs. /.

Figure 6. Microglial cells do not expressAQP4. A) Absence of AQP4 expression inreactive microglia following intracerebralinjection of LPS. Immunofluorescence ofbrain sections stained for microglial mar-ker anti-Iba1 (red) and AQP4 (green).Representative of 3 sets of experiments.Scale bars � 50 �m. B) TNF-� in whole-brain homogenates at 2 h after intracere-bral injection of saline (control) or LPS(se, n�4, differences between / and�/� not significant).

7PRONEUROINFLAMMATORY ROLE OF AQUAPORIN-4

hyperosmolar with mannitol to minimize cell swelling(Fig. 7E).

To investigate whether the increase in cytokine se-cretion is related to the water permeability function ofAQP4, we measured LPS-induced secretion of TNF-�and IL-6 in AQP4-null astrocytes after adenovirus-me-diated expression of aquaporins. Figure 8A shows ex-pression of M1-AQP4, M23-AQP4, and AQP1 at 3 dafter adenovirus infection. AQP1 was studied as anunrelated aquaporin with similar water permeabilityfunction to AQP4. At an MOI of 20, expression ofaquaporins was seen in most astrocytes, without evi-dence of cytotoxicity. By immunofluorescence, theexpression of M1 and M23 AQP4 was comparable tothat seen in astrocyte cultures from wild-type mice.Figure 8B show significantly increased TNF-� and IL-6secretion at 2 h after LPS in the AQP4-null astrocytesafter adenoviral expression of M1-AQP4, M23-AQP4, orAQP1, compared to noninfected or GFP adenovirus-infected cultures. These results suggest the involvementof AQP4 water permeability in cytokine release. To testthe generality of aquaporin-facilitated cytokine secre-tion, we identified a nonastrocyte cell line that showedrobust LPS-induced IL-6 secretion and does not expressaquaporins, T24 urinary bladder cells. Figure 8C shows

expression of M1-AQP4, M23-AQP4, AQP1, and GFP inT24 cells after adenovirus infection. Figure 8D (toppanel) shows significantly increased LPS-stimulatedIL-6 secretion in the aquaporin-expressing T24 cellscompared with control or GFP-expressing cells, withoutincreased cell-associated IL-6 (Fig. 8D, bottom panel).These data suggest that aquaporin-facilitated cytokinesecretion is not cell type specific.

DISCUSSION

Neuroinflammation and demyelination in MOG-in-duced EAE in mice involves a series of steps, includingT-lymphocyte sensitization and activation, passage intothe CNS across the blood-brain barrier, and initiationof a neuroinflammatory cascade involving multipleinflammatory factors and immune cell actions. There isalso a potential role for humoral factors, as antibodiesagainst various CNS targets have been identified in EAE(27). Here, we systematically dissected these processesin order to determine the AQP4-dependent steps thatmight account for the reduced severity in AQP4 defi-ciency in our original study of EAE produced by activeimmunization with MOG35–55 oligopeptide (11), and,

Figure 7. Reduced cytokine release by astrocyte cultures in AQP4 deficiency. A) GFAP (left panels) and AQP4 (right panels)immunofluorescence of differentiated, primary cultures of astrocytes from neonatal mouse brain cortex. B) TNF-� and IL-6 inculture medium at 2 h after LPS (100 ng/ml) or saline addition (se, n�6). Representative of 5 sets of cultures. C) Cell-associatedTNF-� and IL-6 at 2 h after LPS (se, n�4, differences not significant comparing �LPS and LPS). Representative of 3 sets ofcultures. (right) Quantitative real-time RT-PCR of indicated transcripts from astrocyte cultures (se, n�4, differences notsignificant comparing �LPS and LPS). D) Right panel: TNF-� in culture medium at 24 h after Mn3 exposure (se, n�4). Leftpanels: GFAP immunofluorescence of astrocytes after 24 h Mn3 exposure. E) TNF-� and IL-6 in culture medium at 2 h afterLPS in control medium (290 mosmol) or hyperosmolar (310 mosmol) medium containing excess 20 mM mannitol (se, n�4).*P � 0.01 vs. corresponding /.

8 Vol. 25 May 2011 LI ET AL.The FASEB Journal � www.fasebj.org

as found here, in EAE produced adoptive-transfer ofMOG-sensitized T lymphocytes. AQP4 expression wasabsent in the relevant immune cells, as well as inmicroglia. The CNS penetration of intravenously ad-ministered, MOG-sensitized T lymphocytes was compa-rable in wild-type and AQP4-null recipient mice, as wasEvan’s blue dye extravasation. Anti-AQP4 antibodieswere not found in the EAE model. The remarkablepositive finding was a greater inflammatory response inbrain of wild-type vs. AQP4-null mice following LPSchallenge. Our data implicate AQP4 as a novel deter-minant of neuroinflammation that likely accounts forthe reduced EAE severity in AQP4 deficiency, and forthe greatly reduced neuroinflammatory response inAQP4 deficiency in a mouse model of bacterial menin-gitis (17).

The known biology of AQP4 and the findings here inastrocyte cultures suggest the cellular mechanisms re-sponsible for the involvement of AQP4 in neuroinflam-mation. AQP4 facilitates water movement into the brainthrough an intact blood-brain barrier in cytotoxic brainedema, as seen in water intoxication and ischemicstroke (12); as a bidirectional water transporter AQP4also facilitates exit of excess water from the brain invasogenic brain edema, as seen in brain tumor, abscess,and obstructive hydrocephalus (13). AQP4-facilitatedwater transport by astrocytes at the blood-brain andcerebrospinal fluid-brain barriers likely accounts forthe AQP4-dependent water movement in brain. AQP4plays a similar role in spinal cord, as seen in studies ofclinical and histological outcomes following compres-sion (28) and impact (29) spinal cord injury. Astrocytewater permeability is at least 5–10 times reduced inAQP4 deficiency, as demonstrated in astrocyte cultures

and in vivo (30). We found here that cytokine releasefrom astrocyte cultures was reduced in AQP4 deficiencyin response to very different stimuli, including LPS andMn3. The increased swelling of astrocytes from wild-type mice and the reduced cytokine release in hyper-osmolar media suggest a mechanism involving AQP4-dependent astrocyte water permeability and consequentcell swelling to account for the reduced cytokine releasein AQP4 deficiency. The increased cytokine release fol-lowing astrocyte and nonastrocyte cell transfections withAQP4, or an unrelated aquaporin, AQP1, supports theinvolvement of aquaporin water permeability in cytokinesecretion. Evidence for involvement of aquaporins in ananalogous process, secretory vesicle exocytosis, has beenreported (31, 32), though the biophysical mechanismsremain speculative on how aquaporin water transportfacilitates fusion of secretory vesicles with the cell plasmamembrane. Although other mechanisms to account forthe proneuroinflammatory effect of AQP4 cannot beruled out, such as differences in expression of variousgenes in AQP4 deficiency, our proposed mechanismprovides a direct link between the unique molecularfunction of AQP4, osmotic water transport, and neuroin-flammation. We propose that the proinflammatory role ofAQP4 involves a positive-feedback cycle of local brainswelling (cytotoxic edema) and secretion of proinflamma-tory cytokines, which, at the molecular level, hinges onAQP4-dependent osmotic water permeability and astro-cyte swelling.

A second major role of AQP4 in brain is in astrocytemigration, which we have proposed involves AQP4-facilitated water transport in lamellipodia at the leadingedge of migrating cells. AQP4-null astrocytes migratemuch slower than wild-type astrocytes in vitro (20) and

Figure 8. Evidence for aquaporin-facilitated cyto-kine secretion. A) Immunofluorescence of AQP4(M1 and M23 isoforms) and AQP1 in adenovirus-treated primary astrocyte cultures from AQP4-nullmice (AQP4�/� astrocytes). AQP4 immunofluores-cence of culture from wild-type mice shown at right.

B) TNF-� and IL-6 in culture medium at 2 h after LPS (100 ng/ml) or saline addition (se, n�4). Representative of 3 setsof cultures/infections. *P � 0.01 vs. �/� control. C) AQP4 and AQP1 immunofluorescence, and GFP fluorescence, ofadenovirus-treated T24 (bladder) cells. D) Secreted IL-6 (in culture medium, top panel) and cell-associated IL-6 (bottompanel) at 2 h after addition of LPS (se, n�4). *P � 0.01 vs. LPS-treated control.

9PRONEUROINFLAMMATORY ROLE OF AQUAPORIN-4

in vivo (19), and glial scarring is reduced in AQP4-nullmice (20). AQP4-dependent astrocyte migration is un-likely to play a role in LPS-induced neuroinflammationhere, as little astrocyte migration occurs in 24–48 hfollowing a neuroinflammatory stimulus. A third role ofAQP4 is in neuroexcitation, where AQP4-facilitatedwater transport in astrocytes modulates extracellularspace water and K handling. Mice lacking AQP4manifest increased seizure severity (21), delayed K

reuptake from the extracellular space following neuro-excitation (33), and impaired visual (34), auditory(35), and olfactory (36) signal transduction. AQP4-dependent neuroexcitation is unlikely to play a role inneuroinflammation, as AQP4 deficiency would be pre-dicted to increase rather than reduce neuroexcitotox-icity and downstream neuroinflammatory responses.Finally, the mild extracellular space expansion inAQP4 deficiency (15, 16) is unlikely to be responsiblefor reduced neuroinflammation because the CNSdiffusion of inflammatory cells and soluble factorswould be increased, rather than reduced, in AQP4deficiency.

Our original motivation for studying EAE in AQP4deficiency was to obtain insight into the pathogenesis ofNMO, where AQP4 autoantibodies are found in themajority of NMO patients (37). A central question inNMO has been the role of NMO-IgG in disease patho-genesis, versus, or perhaps in addition to, activated Tlymphocytes as in EAE. Evidence was reported recentlythat NMO-IgG, when injected together with humancomplement into brain parenchyma of naive mice,produced, in 10 d, characteristic NMO lesions withperivascular neuroinflammation and complement de-position, demyelination, and loss of AQP4 and GFAPimmunoreactivity (9). Neuroinflammatory lesions werenot produced in AQP4-null mice. Our findings heresuggest a second, AQP4-dependent mechanism formodulation of neuroinflammation in NMO and otherneuroinflammatory diseases. We note that thoughAQP4 expression is generally low within active NMOlesions, it is increased in surrounding brain in NMO; incontrast, AQP4 expression remains intact within plaquelesions in multiple sclerosis (38, 39).

In summary, our results establish a novel role forAQP4 in neuroinflammation, which we propose at thecellular level involves AQP4-dependent differences inastrocyte water permeability and consequent cell swell-ing and cytokine release. Pharmacological inhibition ofAQP4 water permeability or reduction in AQP4 plasmamembrane expression may thus be of benefit in thetherapy of neuroinflammatory diseases.

This work was supported by grants from the Guthy-Jackson Charitable Foundation (A.S.V. and S.S.Z.); U.S.National Institutes of Health (NIH) grants DK35124,EY13574, HL73856, DK72517, DK86125, and EB00415(A.S.V.); and NIH grants AI73737 and NS63008, NationalMultiple Sclerosis Society grant RG4124, and the MaisinFoundation (S.S.Z.).

REFERENCES

1. Ransohoff, R. M. (2006) A mighty mouse: building a bettermodel of multiple sclerosis. J. Clin. Invest. 116, 2313–2316

2. Korn, T., Bettelli, E., Oukka, M., and Kuchroo, V. K. (2009)IL-17 and Th17 cells. Annu. Rev. Immunol. 27, 485–517

3. Kroenke, M. A., Carlson, T. J., Andjelkovic, A. V., and Segal,B. M. (2008) IL-12- and IL-23-modulated T cells induce distincttypes of EAE based on histology, CNS chemokine profile, andresponse to cytokine inhibition. J. Exp. Med. 205, 1535–1541

4. Lyons, J. A., Ramsbottom, M. J., and Cross, A. H. (2002) Criticalrole of antigen-specific antibody in experimental autoimmuneencephalomyelitis induced by recombinant myelin oligoden-drocyte glycoprotein. Eur. J. Immunol. 32, 1905–1913

5. Weber, M. S., Prod’homme, T., Patarroyo, J. C., Molnarfi, N.,Karnezis, T., Lehmann-Horn, K., Danilenko, D. M., Eastham-Anderson, J., Slavin, A. J., Linington, C., Bernard, C. C., Martin,F., and Zamvil, S. S. (2010) B-cell activation influences T-cellpolarization and outcome of anti-CD20 B-cell depletion incentral nervous system autoimmunity. Ann. Neurol. 68, 369–383

6. Lennon, V. A., Kryzer, T. J., Pittock, S. J., Verkman, A. S., andHinson, S. R. (2005) IgG marker of optic-spinal multiplesclerosis binds to the aquaporin-4 water channel. J. Exp. Med.202, 473–477

7. Bennett, J. L., Lam, C., Kalluri, S. R., Saikali, P., Bautista, K.,Dupree, C., Glogowska, M., Case, D., Antel, J. P., Owens, G. P.,Gilden, D., Nessler, S., Stadelmann, C., and Hemmer, B. (2009)Intrathecal pathogenic anti-aquaporin-4 antibodies in early neu-romyelitis optica. Ann. Neurol. 66, 617–629

8. Kinoshita, M., Nakatsuji, Y., Kimura, T., Moriya, M., Takata, K.,Okuno, T., Kumanogoh, A., Kajiyama, K., Yoshikawa, H., andSakoda, S. (2009) Neuromyelitis optica: passive transfer to ratsby human immunoglobulin. Biochem. Biophys. Res. Commun. 386,623–627

9. Saadoun, S., Waters, P., Bell, B. A., Vincent, A., Verkman, A. S.,and Papadopoulos, M. C. (2010) Intra-cerebral injection ofneuromyelitis optica immunoglobulin G and human comple-ment produces neuromyelitis optica lesions in mice. Brain 133,349–361

10. Verkman, A. S., Binder, D. K., Bloch, O., Auguste, K., andPapadopoulos, M. C. (2006) Three distinct roles of aquaporin-4in brain function revealed by knockout mice. Biochim. Biophys.Acta 1758, 1085–1093

11. Li, L., Zhang, H., and Verkman, A. S. (2009) Greatly attenuatedexperimental autoimmune encephalomyelitis in aquaporin-4knockout mice. BMC Neurosci. 10, 94

12. Manley, G. T., Fujimura, M., Ma, T., Noshita, N., Filiz, F., Bollen,A. W., Chan, P., and Verkman, A. S. (2000) Aquaporin-4deletion in mice reduces brain edema after acute water intoxi-cation and ischemic stroke. Nat. Med. 6, 159–163

13. Papadopoulos, M. C., Manley, G. T., Krishna, S., and Verkman,A. S. (2004) Aquaporin-4 facilitates reabsorption of excess fluidin vasogenic brain edema. FASEB J. 18, 1291–1293

14. Saadoun, S., Tait, M. J., Reza, A., Davies, D. C., Bell, B. A.,Verkman, A. S., and Papadopoulos, M. C. (2009) AQP4 genedeletion in mice does not alter blood-brain barrier integrity orbrain morphology. Neuroscience 161, 764–772

15. Yao, X., Hrabetova, S., Nicholson, C., and Manley, G. T. (2008)Aquaporin-4-deficient mice have increased extracellular spacewithout tortuosity change. J. Neurosci. 28, 5460–5464

16. Zhang, H., and Verkman, A. S. (2010) Microfiberoptic mea-surements of extracellular space volume in brain and tumorslices based on fluorescent dye partitioning. Biophys. J. 99,1284 –1291

17. Papadopoulos, M. C., and Verkman, A. S. (2005) Aquaporin-4gene disruption in mice reduces brain swelling and mortality inpneumococcal meningitis. J. Biol. Chem. 280, 13906–13912

18. Bloch, O., Auguste, K. I., Manley, G. T., and Verkman, A. S.(2006) Accelerated progression of kaolin-induced hydrocepha-lus in aquaporin-4-deficient mice. J. Cereb. Blood Flow Metab. 26,1527–1537

19. Auguste, K. I., Jin, S., Uchida, K., Yan, D., Manley, G. T.,Papadopoulos, M. C., and Verkman, A. S. (2007) Greatlyimpaired migration of implanted aquaporin-4-deficient astro-glial cells in mouse brain toward a site of injury. FASEB J. 21,108–116

10 Vol. 25 May 2011 LI ET AL.The FASEB Journal � www.fasebj.org

20. Saadoun, S., Papadopoulos, M. C., Watanabe, H., Yan, D.,Manley, G. T., and Verkman, A. S. (2005) Involvement ofaquaporin-4 in astroglial cell migration and glial scar formation.J. Cell Sci. 118, 5691–5698

21. Binder, D. K., Yao, X., Zador, Z., Sick, T. J., Verkman, A. S., andManley, G. T. (2006) Increased seizure duration and slowedpotassium kinetics in mice lacking aquaporin-4 water channels.Glia 53, 631–636

22. Ma, T., Yang, B., Gillespie, A., Carlson, E. J., Epstein, C. J., andVerkman, A. S. (1997) Generation and phenotype of a trans-genic knockout mouse lacking the mercurial-insensitive waterchannel aquaporin-4. J. Clin. Invest. 100, 957–962

23. Miller, S. D., Karpus, W. J., and Davidson, T. S. (2010).Experimental autoimmune encephalomyelitis in the mouse.Curr. Protoc. Immunol. Chap. 15, unit 15–11

24. Swanson, R. A., Liu, J., Miller, J. W., Rothstein, J. D., Farrell, K.,Stein, B. A., and Longuemare, M. C. (1997) Neuronal regula-tion of glutamate transporter subtype expression in astrocytes.J. Neurosci. 17, 932–940

25. Bartholomaus, I., Kawakami, N., Odoardi, F., Schlager, C.,Miljkovic, D., Ellwart, J. W., Klinkert, W. E., Flugel-Koch, C.,Issekutz, T. B., Wekerle, H., and Flugel, A. (2009) Effector T cellinteractions with meningeal vascular structures in nascent auto-immune CNS lesions. Nature 462, 94–98

26. Tomas-Camardiel, M., Venero, J. L., de Pablos, R. M., Rite, I.,Machado, A., and Cano, J. (2004) In vivo expression of aqua-porin-4 by reactive microglia. J. Neurochem. 91, 891–899

27. Martin Mdel, P., and Monson, N. L. (2007) Potential role ofhumoral immunity in the pathogenesis of multiple sclerosis(MS) and experimental autoimmune encephalomyelitis (EAE).Front. Biosci. 12, 2735–2749

28. Saadoun, S., Bell, B. A., Verkman, A. S., and Papadopoulos, M. C.(2008) Greatly improved neurological outcome after spinal cordcompression injury in AQP4-deficient mice. Brain 131, 1087–1098

29. Kimura, A., Hsu, M., Seldin, M., Verkman, A. S., Scharfman, H. E., andBinder, D. K. (2010) Protective role of aquaporin-4 water channelsafter contusion spinal cord injury. Ann. Neurol. 67, 794–801

30. Solenov, E., Watanabe, H., Manley, G. T., and Verkman, A. S.(2004) Sevenfold-reduced osmotic water permeability in pri-

mary astrocyte cultures from AQP-4-deficient mice, measured bya fluorescence quenching method. Am. J. Physiol. Cell Physiol.286, C426–C432

31. Cho, S. J., Sattar, A. K., Jeong, E. H., Satchi, M., Cho, J. A., Dash,S., Mayes, M. S., Stromer, M. H., and Jena, B. P. (2002)Aquaporin 1 regulates GTP-induced rapid gating of water insecretory vesicles. Proc. Natl. Acad. Sci. U. S. A. 99, 4720–4724

32. Sugiya, H., and Matsuki, M. (2006) AQPs and control of vesiclevolume in secretory cells. J. Membr. Biol. 210, 155–159

33. Padmawar, P., Yao, X., Bloch, O., Manley, G. T., and Verkman,A. S. (2005) K waves in brain cortex visualized using along-wavelength K-sensing fluorescent indicator. Nat. Meth. 2,825–827

34. Li, J., Patil, R. V., and Verkman, A. S. (2002) Mildly abnormalretinal function in transgenic mice without Muller cell aqua-porin-4 water channels. Invest. Ophthalmol. Vis. Sci. 43, 573–579

35. Li, J., and Verkman, A. S. (2001) Impaired hearing in micelacking aquaporin-4 water channels. J. Biol. Chem. 276, 31233–31237

36. Lu, D. C., Zhang, H., Zador, Z., and Verkman, A. S. (2008)Impaired olfaction in mice lacking aquaporin-4 water channels.FASEB J. 22, 3216–3223

37. Weinshenker, B. G., Wingerchuk, D. M., Pittock, S. J., Lucchi-netti, C. F., and Lennon, V. A. (2006) NMO-IgG: a specificbiomarker for neuromyelitis optica. Dis. Markers 22, 197–206

38. Roemer, S. F., Parisi, J. E., Lennon, V. A., Benarroch, E. E.,Lassmann, H., Bruck, W., Mandler, R. N., Weinshenker, B. G.,Pittock, S. J., Wingerchuk, D. M., and Lucchinetti, C. F. (2007)Pattern-specific loss of aquaporin-4 immunoreactivity distin-guishes neuromyelitis optica from multiple sclerosis. Brain 130,194–205

39. Misu, T., Fujihara, K., Kakita, A., Konno, H., Nakamura, M.,Watanabe, S., Takahashi, T., Nakashima, I., Takahashi, H., andItoyama, Y. (2007) Loss of aquaporin 4 in lesions of neuromy-elitis optica: distinction from multiple sclerosis. Brain 130,1224–1234

Received for publication November 22, 2010.Accepted for publication January 6, 2011.

11PRONEUROINFLAMMATORY ROLE OF AQUAPORIN-4