bypassing vasopressin receptor signaling pathways in
TRANSCRIPT
Wogw
M
S
A
0©
2
Bypassing VasopressinReceptor Signaling Pathways
in Nephrogenic Diabetes Insipidus
Richard Bouley, PhD, Udo Hasler, PhD, Hua A. J. Lu, MD, PhD, Paula Nunes, BSc, andDennis Brown, PhD
Summary: Water reabsorption in the kidney represents a critical physiological event in themaintenance of body water homeostasis. This highly regulated process relies largely onvasopressin (VP) action and on the VP-sensitive water channel (AQP2) that is expressed inprincipal cells of the kidney collecting duct. Defects in the VP signaling pathway and/or inAQP2 cell surface expression can lead to an inappropriate reduction in renal water reabsorp-tion and the development of nephrogenic diabetes insipidus, a disease characterized bypolyuria and polydipsia. This review focuses on the major regulatory steps that are involved inAQP2 trafficking and function. Specifically, we begin with a discussion on VP-receptor–indepen-dent mechanisms of AQP2 trafficking, with special emphasis on the nitric oxide–cyclic guanosinemonophosphate signaling pathway, followed by a review of the mechanisms that govern AQP2endocytosis and exocytosis. We then discuss emerging data illustrating roles played by the actincytoskeleton on AQP2 trafficking, and lastly we consider elements that affect AQP2 proteinexpression in cells. Recent advances in each topic are summarized and are presented in thecontext of their potential to serve as a basis for the development of novel therapies that mayultimately improve life quality of nephrogenic diabetes insipidus patients.Semin Nephrol 28:266-278 © 2008 Elsevier Inc. All rights reserved.Keywords: Vasopressin receptor, AQP2, endocytosis, exocytosis, cGMP
iir(Ne2eodtItfimsufV
ater homeostasis and urine concen-tration via water reabsorption in theurinary tubule are integral functions
f the kidney. In normal human subjects, thelomerular system can filter 180 L/d of fluid, ofhich 90% is reabsorbed back into the circulat-
assachusetts General Hospital-Center for Systems Biology, Program inMembrane Biology and Nephrology Division, Massachusetts General Hos-pital and Harvard Medical School, Boston, MA.
upported by National Institutes of Health grant DK38452. Also supported bya Young Investigator Award from the National Kidney Foundation (R.B.), aSwiss Fondation Suisse pour les Bourses en Médicine et Biologie Fellowshipand an Executive Committee on Research Fellowship from MassachusettsGeneral Hospital (U.H.), a National Institutes of Health KO8 grantDK075940-01 (H.A.J.L.), and a doctoral level postgraduate scholarship fromNational Sciences and Engineering Council of Canada (P.N.). The Micros-copy Core facility of the MGH Program in Membrane Biology receivesadditional support from the Boston Area Diabetes and Endocrinology Re-search Center (DK57521) and the Center for the Study of InflammatoryBowel Disease (DK43341).
ddress reprint requests to Dr. Dennis Brown, Center for Systems Biology,Massachusetts General Hospital, 185 Cambridge St, Simches Research BldgSuite 8202, Boston, MA 02114. E-mail: [email protected]
270-9295/08/$ - see front matter
m2008 Elsevier Inc. All rights reserved. doi:10.1016/j.semnephrol.2008.03.010Semina66
ng system in the proximal tubule and descend-ng limb of Henle’s loop. The remaining 10% iseabsorbed under the regulation of vasopressinVP) at the level of the collecting duct (CD).ephrogenic diabetes insipidus (NDI) is a dis-ase characterized by massive water loss (up to0 L/d) via the kidneys, and the disease canither be acquired or inherited. Acquired NDI isften observed in patients suffering from disor-ers such as hypokalemia, hypercalcemia, ure-eral obstruction, and secondary aldosteronism.n addition, at least 20% of bipolar patientsreated with lithium acquire NDI. In inheritedorms of NDI, early symptoms include fever, vom-ting, anorexia, growth retardation, and develop-
ental delay, while later in life polyuria, polydip-ia, and even mental retardation can ensue inntreated patients. Both acquired and congenitalorms of NDI have been linked to defects in theP hormone signaling system, which, under nor-
al conditions, increases both apical cell surfacers in Nephrology, Vol 28, No 3, May 2008, pp 266-278
epciht(t
mioeVbfcVt
ecpAhtbctydccttCia
Ftsspntbl f activ
VP receptor signaling pathways in NDI 267
xpression and whole-cell abundance of the aqua-orin-2 (AQP2) water channel in CD principalells. The most severe forms of NDI are observedn congenital cases, in which patients most oftenarbor mutations in the vasopressin type-2 recep-or (V2R) gene, although a small percentage10%) bear recessive or autosomal-dominant mu-ations in the AQP2 gene.1
The release of VP, a cyclic nonapeptide hor-one secreted by the posterior pituitary gland,
s regulated in the brain in response to serumsmolality and body volume status. In the pres-nce of high serum osmolality or hypovolemia,P is released into the blood stream where itinds to V2R expressed on the basolateral sur-ace of CD principal cells. V2R is a G-protein–oupled receptor, and VP binding initiates the2R signaling cascade, inducing a conforma-
igure 1. AQP2 trafficking in principal cells. This modehe major pathways that affect AQP2, and summarizesignaling pathway is depicted, with VP stimulation ofubsequently altering the balance between exocytosis alasma membrane. Also shown are the NO-cGMP pathegatively modulate AQP2 trafficking, respectively, as weranscription factors that mediate this regulation. AP1, ainding protein; Gi�, G-protein i � subunit; Gs�, G-pr
ymphocyte-associated protein; NFATc, nuclear factor o
ional change that promotes Gs� dissociation, ad- r
nylyl cyclase activation, and consequently an in-rease in intracellular cyclic adenosine mono-hosphate (cAMP) levels. The classical view ofQP2 trafficking, based on the so-called “shuttleypothesis” originally proposed to explain wa-er channel trafficking in the toad urinaryladder,2 postulates that an increase in cAMPoncentration and ensuing activation of pro-ein kinase type A (PKA) leads to the phosphor-lation of AQP2 at S256 located in its C-terminalomain, promoting exocytosis and cell-surface ac-umulation of AQP2.3 In addition to cAMP in-rease, an increase in intracellular Ca2� concen-ration and Ca2� oscillations are also a part ofhe VP response.4,5 Although it is likely thata2� plays a role in AQP2 plasma membrane
nsertion in addition to cAMP, the exact mech-nisms and relative contributions of each signal
s the interactions between the components of some off the points outlined in this review. The canonical V2R
leading to the phosphorylation of AQP2 by PKA anddocytosis, leading to AQP2 accumulation at the apicalnd the PGE2 receptor EP3 that can also positively andctors that affect AQP2 abundance and the elements andr protein-1 element; AC, adenylate cyclase; CREB, CRE
s � subunit; GC, guanylate cyclase; MAL, myelin andated T-cells c; Rho, Rho family small GTPase.
l showmost oV2Rnd enway all as factivatootein
emain to be elucidated (Fig. 1). Expression of
Aoocasit
radtrhdsrpuetsvsmVclucovsiatoaoathnhi
shd
rocarpmscutdAfuvwAtcso
T
TpbpwfittdNffplniopcigt(c
268 R. Bouley et al
QP2 at the apical membrane leads to an influxf water into the cell driven by the interstitialsmotic gradient generated by urea and sodiumhloride. Water then exits the cell via AQP3nd AQP4 water channels located at the ba-olateral side of the cell, allowing its re-entrynto the interstitium and the circulatory sys-em.
Polyuria in acquired NDI can be partiallyeduced by a combination of treatments such asdequate hydration, low-salt and/or low-proteiniet, diuretics, and nonsteroidal anti-inflamma-ory drugs.6 However, congenital patients oftenespond poorly to such therapies. Some studiesave focused on rescuing misfolded V2R byeveloping nonpeptidic lipid-soluble vasopres-in ligands that cross the plasma membrane andeach misfolded receptors trapped in the endo-lasmic reticulum. The ligand acts like a molec-lar chaperone, helping V2R refold, escape thendoplasmic reticulum quality control, and reachhe plasma membrane where endogenous VP canubsequently displace the VP analogue and acti-ate the receptor.7 These compounds have shownome positive effects on patients bearing specificissense mutations or small insertion/deletion2R mutations, but are not effective against trun-ated proteins, and furthermore, they did not al-eviate polyuria completely. A second strategy hassed aminoglycoside antibiotics such as gentami-in. This class of antibiotics causes read-throughf some nonsense V2R mutations in vitro and inivo,8 but the beneficial effect of aminoglyco-ides unfortunately is overshadowed by its tox-city to the kidneys. Another disadvantage ofttempts at rescuing mutant receptors is thathey often are heavily dependent on the naturef the mutation and, therefore, any such ther-py may not be widely applicable. Finally, the usef cAMP phosphodiesterase (PDE) inhibitors suchs rolipram has so far been unsuccessful for thereatment of NDI. Although beneficial effectsave been observed in mouse models, they haveot been reproduced in human subjects, per-aps reflecting a difference in cAMP PDE local-
zation.9,10
The search for more effective therapeutictrategies for both acquired and congenital NDIas, thus, motivated many advances in our un-
erstanding of the V2R signaling cascade that eegulates AQP2 trafficking. In addition, variousther physiological factors that modulate AQP2ell-surface localization as well as its abundancere now beginning to be uncovered. In thiseview, we begin by describing the V2R-inde-endent nitric oxide (NO)/cyclic guanosineonophosphate (cGMP) pathway and recent
tudies that have linked this important signalingascade to AQP2 shuttling. We then discuss these of PDE inhibitors such as Viagra as potentialherapeutic agents. Next, we summarize novelevelopments on the role of endocytosis onQP2 trafficking and consider plausible targets
or bypassing V2R, as well as the possibility ofsing statins in NDI treatment. Finally, we pro-ide a brief overview of other signaling path-ays currently being investigated that regulateQP2 abundance and that may be explored in
he future in the search for specific targets thatan be readily exploited in novel therapeutictrategies in the treatment of NDI, as well asther forms of water imbalances.
HE NO-cGMP SIGNALING PATHWAY
ogether with the canonical cAMP-inducedathway, the NO-cGMP signaling pathway haseen shown to play a role in AQP2 trafficking,rompting investigation of this signaling path-ay as a means to develop alternative therapies
or treatment of NDI. NO is a free radical result-ng from the enzymatic conversion of L-arginineo L-citrulline by 1 of 3 isoforms of NO syn-hetase (NOS) expressed in the kidney. NO pro-uced by endothelial NOS (eNOS), inducibleOS, and neuronal NOS can diffuse and, there-
ore, act in both an autocrine and paracrineashion in the kidney. The classic NO signalingathway depends on activation of soluble guany-
ate cyclase, located in several segments of theephron including the CD, where it is expressed
n principal cells. The major effect of an increasef cGMP concentration is cGMP-dependentrotein kinase (PKG) activation. However,GMP also affects PKA and p21ras kinase activ-ty11,12 and can directly regulate Na-channel andlucose transporter (GLUT4) activity, as well asrafficking of type 1 and type 5 water channelsAQP1 and 5).13-16 All components of the NO-GMP signaling pathway are expressed in renal
pithelial cells, supporting the notion that the
Npp
sdtosssacltslNtppohraTsiortts(emacr
datsAdioha
pimtiwupapiaA
V
BsotsAaogescsttibePrAptipt4mAnmCam
VP receptor signaling pathways in NDI 269
O-cGMP signaling pathway plays a key role inrincipal cell physiology, including fluid trans-ort in the kidney.The effect of NO-cGMP in CD water reab-
orption still is controversial. In one study, NOonors were found to decrease VP-induced wa-er reabsorption as a consequence of reducedsmotic water permeability and sodium reab-orption, whereas another report failed to ob-erve this inhibitory effect.17,18 Another studyhowed that NO antagonizes the effect of VP byltering cAMP levels.19 In our hands, VP in-reased the conversion of L-arginine to L-citrul-ine. However, increased NOS activity appearso result from an indirect effect and VP mayimply increase substrate availability. Neverthe-ess, increasing evidence supports the idea thatO-cGMP is involved in renal water reabsorp-
ion. VP increases neuronal NOS and eNOS ex-ression in water-deprived rats, and eNOS ex-ression is accompanied by a reduction of urineutput, suggesting that it plays a role in wateromeostatic mechanisms.20,21 Simultaneous dis-uption of all 3 NOS isoforms led to NDI in mice22
nd a reduction of AQP2 whole-cell abundance.he mechanism that regulates AQP2 abundancetill is elusive, but may be related to low cAMPntracellular levels detected in knock-out micewing to increased prostacyclin activity.22 Aeduction of basal intracellular cGMP concen-ration owing to the absence of NOS may leado decreased levels of functional PKG that sub-equently may affect cAMP-response-elementCRE)-dependent transcription.23 Thus, AQP2xpression, which is regulated chiefly by VP,ay be stimulated additionally by NOS basal
ctivity. This may partially explain the signifi-ant amount of AQP2 expression in Brattleboroats, which do not express circulating VP.
Our study provided evidence that both so-ium nitroprusside, a NO donor, and L-arginine,precursor of NO, are able to shift the localiza-
ion of AQP2 from the cytoplasm to the apicalide of rat CD principal cells.24 This increase ofQP2 membrane insertion was cGMP-depen-ent but cAMP-independent. The role of cGMP
n AQP2 trafficking was confirmed by analysisf the atrial natriuretic peptide receptor, whichas intrinsic guanylate cyclase activity, and by
nalysis of the effects produced by the cell- oermeant dibutyryl cGMP analogue. Both agentsnduced translocation of AQP2 to the plasma
embrane. In addition, atrial natriuretic pep-ide infusion in rat induced a marked increasen AQP2 apical targeting.25 The mechanism by
hich cGMP induces AQP2 trafficking still isnclear. Our study showed that AQP2 can behosphorylated by PKG, but we cannot reason-bly eliminate the possibility that PKG phos-horylates PKA, nor can we yet rule out that an
ncrease of cGMP concentration activates PKAnd subsequently induces an accumulation ofQP2 at the plasma membrane.
IAGRA STIMULATES AQP2 TRAFFICKING
ecause cGMP increases AQP2 membrane in-ertion in rat kidney, we investigated the effectf selective cyclic-3’,5’-nucleotide PDE inhibi-ors, which abolish cGMP catabolism and sub-equently increase cGMP concentration, onQP2 cell surface expression. Intracellular cAMPnd cGMP levels are regulated strongly by oner more members of 11 PDE families that to-ether account for more than 60 isoforms. Sev-ral isoforms are expressed along the nephron,uch as cAMP-sensitive PDEs (PDE 3 and 4),GMP-sensitive PDE (PDE 5), and cAMP/cGMP-elective PDE (PDE 1). We used sildenafil ci-rate (Viagra), a selective cGMP phosphodies-erase (PDE5) inhibitor, as a means to increasentracellular cGMP concentration. Sildenafil haseen used successfully in clinical treatment ofrectile dysfunction. We studied the effect ofDE5 inhibition on AQP2 trafficking in LLC-PK1enal epithelial cells expressing c-myc–taggedQP2. Western blot analysis showed that theresence of sildenafil or 3-isobutyl-1-methylxan-hine (IBMX), a nonselective cAMP/cGMP PDEnhibitor, enhanced AQP2 expression at thelasma membrane (Fig. 2A).26 We also observedhat both PDE5 inhibitors sildenafil and-([3=,4=-(methylenedioxy)benzyl]amino)-6-ethoxyquinazoline modulate endogenousQP2 trafficking in rat kidney (Fig. 2B).24 Silde-afil increased insertion of AQP2 in the apicalembrane of principal cells of outer medullaryD both in vitro and in vivo, but did not haven effect on AQP2 localization in cortical CD, aajor site of water reabsorption.26 Although
ur studies suggest that PDE inhibition may
otttiia
A
AwVmrcptmptscrtlwpabi
CA
Tm
FmprdpwstdclimnpdrDdttC
270 R. Bouley et al
ffer a promising approach for X-linked NDIherapy, further studies need to be performedo determine whether prolonged cGMP inhibi-ion, or a combined therapeutic approach, canmprove water reabsorption in patients suffer-ng from NDI who may express variablemounts of AQP2 in their principal cells.
t
QP2 ENDOCYTOSIS AND EXOCYTOSIS
bout 10% of congenital NDI cases are associatedith AQP2 mutations rather than with defects of2R signaling. Interestingly, the majority of theseutations are manifested as misrouting errors
ather than as structural defects that affect thehannel’s water permeability. Therefore, properlasma membrane insertion/trafficking is integralo proper AQP2 functioning and correct waterolecule transport. A reagent that affects thisrocess may consequently represent a potentialarget for modulating water absorption. Severaltudies from our group have revealed that besidesAMP- and cGMP-stimulated trafficking, AQP2apidly recycles between an intracellular pool andhe plasma membrane under baseline (nonstimu-ated) conditions. In light of this constitutive path-
ay, steady-state accumulation of AQP2 at thelasma membrane may be mediated not only byn increase in exocytosis, as originally postulatedy the shuttle hypothesis, but also by a reduction
n endocytosis.
LATHRIN-MEDIATED ENDOCYTOSIS OFQP2
he clathrin-mediated pathway is one of theajor routes of endocytosis in eukaryotic cells,
igure 2. (A) Western blot detection of AQP2 in plasmaembrane–enriched fractions from LLC-PK1 cells ex-
ressing c-myc–tagged AQP2. (B) Indirect immunofluo-escence microscopy of tissue slices showing AQP2 re-istribution in the inner stripe (outer medulla) of CDrincipal cells in response to PDE V inhibition. (A) Cellsere incubated for 45 minutes in the presence of the
elective PDE5 inhibitor (sildenafil) or with the nonselec-ive PDE inhibitor (IBMX) at a 0.1-, 1-, or 10-fold higherose than that corresponding to the EC50 of eitherhemical agent. A plasma membrane fraction was iso-ated from the cells and probed with anti-AQP2 antibod-es. The same plot was reprobed with an anti–pan-actin
onoclonal antibody as a loading control. Both silde-afil and IBMX induce the appearance of AQP2 in thelasma membrane fraction of the cells in a dose-depen-ent manner. (B) Kidney slices from a Sprague-Dawleyat were incubated for 10 minutes with (Deamino-Cys1,-arg8)VP (DDAVP, 10 nmol/L) or 45 minutes with sil-enafil (0.5 �mol/L) before fixation by immersion, sec-ioning, and immunostaining for AQP2. (A) Diffuse in-racellular distribution of AQP2 in a control medullaryD; apical membrane accumulation is induced in tissues
reated with (B) DDAVP or (C) sildenafil. Bar � 25 �m.
atoptot(pttAtpifirttd(faitmnhasAoaAlv
cpi
tdmiAilspiSsAmvptBoccatsp
A
Acp
FeetpAt lls. Ba
VP receptor signaling pathways in NDI 271
nd is characterized by the selective internaliza-ion of proteins from the cell surface. An elab-rate series of protein–protein interactions im-arts selectivity on this highly dynamic processhat involves the rapid assembly or disassemblyf transient protein complexes. Both adenosineriphosphatases and guanosine triphosphatasesGTPases) modulate assembly of these com-lexes,27,28 requiring multiple interactions be-
ween clathrin, dynamin, heat shock cognate pro-ein 70 (hsc70), the adaptor proteins AP2 andP180, Esp15, and many other accessory pro-
eins such as auxillin, endophilin, and am-hiphysin.28 The presence of molecules later
dentified as AQP2 in clathrin-coated pits wasrst observed more than 20 years ago.29 Moreecently, endocytosis blockade achieved byransfecting a dominant-negative dynamin mu-ant into cultured cells was found to induceramatic membrane accumulation of AQP2Fig. 3).30 Members of the heat-shock proteinamily (hsc70 and hsp70) were found to inter-ct directly with AQP2 and regulate its traffick-ng. Immunogold electron microscopy showedhat hsc70 colocalized with AQP2 in the plasmaembrane. In addition, inhibition of endoge-
ous hsc70 activity using a dominant-negativesc70 mutant also caused dramatic membraneccumulation of AQP2 in cells (Fig. 3).31 Thisuggests that hsc70 is likely to be involved inQP2 endocytosis, although it cannot be ruledut that other biological functions are associ-ted with the interaction of hsc/hsp70 andQP2. A recent observation suggests that mye-
in and lymphocyte-associated protein is in-
igure 3. AQP2 membrane accumulation can be indxpressing AQP2 displayed baseline perinuclear AQP2 stxpression at the plasma membrane (B). Endocytosis wareatment (C), expressing a dominant interfering dynamhatase–deficient hsc70 mutant (T204V) (E). All 3 approQP2 expression at the plasma membrane. Immunostahe c-myc tag of AQP2 in stably transfected LLC-PK1 ce
olved in regulated AQP2 trafficking by physi- m
ally interacting with AQP2. It increases AQP2lasma membrane expression by attenuating its
nternalization.32
An interesting aspect of AQP2 endocytosis ishat AQP2 constitutive recycling is indepen-ent of S256 phosphorylation. AQP2-S256Dutants are localized in the plasma membrane
n the absence of VP stimulation33 whereasQP2-S256A mutant expression is restricted to
ntracellular compartments.34 However, a cho-esterol-depleting agent that inhibits endocyto-is caused a large accumulation of AQP2 at thelasma membrane both in cell cultures35 and in
solated perfused rat kidney.36 Even AQP2-256A mutants rapidly accumulated at the cellurface under these conditions, indicating thatQP2 dephosphorylated at S256 can also accu-ulate at the plasma membrane. Other obser-
ations additionally indicate that S256 phos-horylation alone is not sufficient to induceranslocation of AQP2 to the cell surface. Inrattleboro rats, which display decreased levelsf AQP2 abundance because of the lack of cir-ulating VP, AQP2 is expressed mostly in intra-ellular pools despite the fact that a significantmount of AQP2 is phosphorylated at S256. VPreatment increased AQP2 cell surface expres-ion but did not appear to increase its phos-horylation.37
N ASSAY FOR AQP2 EXOCYTOSIS
lthough it was assumed initially that an in-rease of AQP2 exocytosis arises from AQP2hosphorylation at S256, most assays to date
by inhibiting endocytosis. (A) Control LLC-PK1 cells(A), whereas cells exposed to VP showed strong AQP2ked in LLC-PK1 cells by methyl-�-cyclodextrin (m�CD)tant (dynamin 2 DK44A) (D), or an adenosine triphos-to reduce endocytosis resulted in a dramatic increase ofas performed using an anti–c-myc antibody to detect
r � 20 �m.
ucedainings blocin muachesining w
easured only AQP2 cell surface accumulation
rheuwsysTmtu
sctscstef
ttcAiAsoPGtromphitepct
PS
Bleipcrtptpued
FtLvptssislrs1csiw0flac
272 R. Bouley et al
ather than bona fide exocytosis. We have,owever, recently shown that a burst of AQP2xocytosis occurs during the first 15 to 20 min-tes of VP stimulation (Fig. 4). This observationas made using a novel fluorescence-based as-
ay that relies on expression of secreted solubleellow fluorescence protein (ssYFP) that pas-ively labels biosynthetic/post-Golgi vesicles.his assay provides an indirect but quantitativeeans to measure AQP2 exocytosis. In addi-
ion, a recent study has shown that VP stim-
igure 4. VP/FK (forskolin) treatment increases exocy-osis in AQP2-expressing cells, but not in control cells.LC-PK1 cells expressing AQP2 were transfected with aector encoding a soluble, secreted form of YFP (kindlyrovided by Jennifer Lippincott-Schwartz, National Insti-utes of Health). The amount of ssYFP produced in LLC-sYFP (which express YFP but not AQP2) and LLC-AQP2–sYFP cells (which express AQP2 and YFP) and secretedn the extracellular medium after 15 minutes was mea-ured by fluorimetry, and is similar between both cellines under baseline conditions (bars 1 and 3 from left toight). When VP/FK is applied, AQP2-expressing cellshow a large increase in ssYFP secretion within the first5 minutes of stimulation, as compared with controlells (bars 2 and 4, respectively). These results are con-istent with a large burst of exocytosis of AQP2-contain-ng vesicles in response to VP/FK stimulation. Valuesere calculated as the relative increase from the-minute baseline control and are expressed in relativeuorescence units (RFU). Each bar represents the aver-ge of 5 independent experiments performed in tripli-ate.
lation increases AQP2 expression at the cell t
urface by inducing its accumulation in “endo-ytosis-resistant” membrane domains,38 indicatinghat although VP enhances AQP2 exocytosis, itignificantly reduces AQP2 endocytosis. In-reased expression of water channels at the cellurface through altered exocytotic and endocy-otic activity already was suggested by math-matic modeling more than a decade ago be-ore the molecular identification of AQPs.39
It is well known that protein phosphoryla-ion and dephosphorylation markedly affecthe biological activity of proteins. As dis-ussed earlier regarding the S256 residue,QP2 phosphorylation plays an important role
n AQP2 trafficking/membrane accumulation.nalysis of potential AQP2 phosphorylationites in addition to S256 suggests the presencef putative sites for at least 4 kinases, namelyKG, protein kinase C, casein kinase II, andolgi casein kinase in addition to PKA. The po-
ential role of these phosphorylation sites is cur-ently under investigation by our group as well asthers.33,34,37,40-42 AQP2 targeting to the apicalembrane may be achieved by manipulating itshosphorylation state, and pharmacological in-ibition of phosphatase activity by okadaic acid
s sufficient to increase expression of AQP2 athe plasma membrane of cultured cells.43 Thevents governing AQP2 phosphorylation and de-hosphorylation undoubtedly will lead to the dis-overy of potential targets for the development ofherapeutic reagents.
OTENTIAL ROLE OFTATINS IN AQP2 TRAFFICKING
y reducing cholesterol-containing atherogenicipoproteins,44 3-hydroxy-3-methylglutaryl co-nzyme A reductase inhibitors (statins) dramat-cally improve cardiovascular outcome. Studieserformed on cell cultures of proximal tubularells have provided evidence that statins reduceeceptor-mediated endocytosis (RME) and thathis is a consequence of statin-induced impairedrenylation and, therefore, membrane associa-ion, of one or more GTP-binding proteins thatlay a key role in RME.45,46 By reducing proteinptake, statins may even exert a renoprotectiveffect as suggested by animal models of kidneyisease47 and by a meta-analysis of randomized
rials.48 Various observations point to the pos-
siRqpwtcimeciwaAwt(cidatTgRcAtAt
RCA
AdaosbaccGcdoA
brsTrtdAmtii
pactctqPiirsvtAoda
PC
PdatepiThrwmmso
VP receptor signaling pathways in NDI 273
ibility that statins might be used as a means toncrease AQP2 cell surface expression. First,ME is a clathrin-mediated process, which re-uires the participation of several GTP-bindingroteins, such as Rho, Rac, and Rab. Consistentith this, prenylation of the GTP-binding pro-
ein Rap1A was found to be reduced by statins inultured proximal cells.45 An effect similar to thatnduced by statins on RME in proximal cells also
ay occur in CD cells and could affect AQP2ndocytosis. In this respect, statins may in-rease AQP2 cell surface expression by reduc-ng its endocytosis. Second, low doses of statins
ere found to increase eNOS phosphorylationnd activation in endothelial cells via increasedkt activation,49,50 and high doses of statinsere found to increase eNOS protein synthesis
hrough an increase of eNOS messenger RNAmRNA) stability.51 Statins consequently may in-rease AQP2 cell surface expression by enhanc-ng eNOS activity (see earlier). Third, highoses of statins were found to decrease Rhoctivity,50 a key player of actin reorganizationhat affects AQP2 trafficking, as discussed later.his effect may be attributed to inhibition of theeranylgeranylation and membrane localization ofhoA and by alterations in RhoA-dependentell-signaling pathways, such as flk-1/KDR andkt.52 Based on these observations, our labora-
ory currently is evaluating the role of statins inQP2 trafficking with a view to developing po-
ential novel therapies for the treatment of NDI.
EARRANGEMENT OFYTOSKELETAL COMPONENTSND REGULATION OF SMALL G-PROTEINS
ctin polymerization and depolymerization is aynamic and tightly regulated process that playsn important role in protein trafficking. Actin re-rganization is controlled by the Rho family ofmall GTP-binding proteins. This includes mem-ers of the RhoA-G, Cdc42, and Rac1 family thatre activated after guanosine diphosphate is ex-hanged with GTP. The nucleotide exchange pro-ess is controlled by various proteins such asTPase activating protein, guanine nucleotide ex-hange factors (GEFs), and guanine nucleotideissociation inhibitors (GDI). Depolymerizationf the actin network results in an increase of
QP2 expression at the cell surface whereas tlockade of VP-induced AQP2 translocation inesponse to Rho activation was shown to be as-ociated with increased actin polymerization.53-55
hus, modulation of the actin cytoskeleton mightepresent a therapeutic approach for NDI, despitehe omnipresence of Rho that makes this proteinifficult to target specifically in CD principal cells.t the very least, a better understanding of theechanisms that regulate cytoskeletal reorganiza-
ion and AQP2 trafficking undoubtedly will helpdentify therapeutic targets whose modified activ-ties may provide the basis for future therapies.
A shift of the equilibrium between V2R androstaglandin E2 (PGE2) receptor stimulationffects the polymerization state of the actinytoskeleton and consequently affects AQP2rafficking to the plasma membrane. An in-rease of cAMP concentration after V2R activa-ion results in Rho inhibition56 and the subse-uent depolymerization of the actin cytoskeleton.GE2, on the other hand, counteracts the VP-nduced increase of osmotic water permeabilityn the renal CD. When PGE2 binds to the EP3
eceptor, adenylate cyclase is inactivated, re-ulting in an increase of actin polymerizationia Rho activation. PGE2 also may counteracthe intrinsic actin reorganization capability ofQP2-bearing vesicles, as suggested by a recentbservation that shows that AQP2 can interactirectly with actin and SPA-1, a specific Rap GTP-se activating protein.57
ROSTAGLANDINS AND URINEONCENTRATION
GE2 is expressed abundantly in the kidney. Iterives from arachidonic acid via cyclooxygen-se (COX)58 and PGE synthetase (PGES) activi-ies. Two COX isoforms, COX-1 and COX-2, arexpressed in the kidney. Interestingly, COX-2 ex-ression, which is known to be induced by phys-
ologic stress, is increased in NDI patients.59,60
he development of selective COX inhibitorsas raised several expectations. For example,ofecoxib (a COX-2 inhibitor) in combinationith hydrochlorothiazide and a low-salt for-ula reduced urine volume in a 1-month-oldale infant.61 However, COX-2 inhibitors
hould be used with extreme caution becausef the high risk of developing myocardial infarc-
ion.62 The adverse effects associated with this
fst
rtpmihaatmTdiauiitagdaivrtcfrt
bmtssliaBrBHrpt
aduAc(dccGp
MA
ItafcNflieAaimfdadbtdfCdmdN
riitbmpt
274 R. Bouley et al
amily of inhibitors suggest that more researchhould focus on the downstream effectors ofhe COX/PGE2 signaling pathway.
Three isomers of PGES have been describedecently. Interestingly, the microsomal PGE syn-hase 1 (mPGE1) isoform is inducible and its ex-ression is tightly related to COX-2 expression.PGE1 is expressed in the CD and is increased
n type 2 diabetes. The role of mPGES in NDIas not been investigated fully but the recentvailability of selective mPGE1 inhibitors willllow us to investigate in-depth their potentialherapeutic benefits.63 Several efforts have beenade to develop PGE-receptor antagonists.hree of 4 PGE-receptor subtypes (prostaglan-in E2 receptor types 1, 3, and 4) are expressed
n different regions of the kidney. EP1 and EP4
re expressed in the glomerulus, whereas EP3 isndetectable in this region. However, 2 EP3
soforms are expressed in the CD.64 Some inhib-tors of the PGE receptor have been developedhat show interesting effects. An EP1-selectiventagonist has been shown to prevent the pro-ression of nephropathy in streptozotocin-in-uced diabetic rats.65 In that study, Makino etl65 showed that aspirin, a nonselective COXnhibitor, has more beneficial effects on urineolume than a COX-selective antagonist. Thisesult indicates that selective PGE-receptor an-agonism may represent an efficient means ofontrolling water excretion and that every ef-ort should be made to develop other PGE-eceptor inhibitors that target other PGE-recep-or isoforms such as EP3.
Other alternative mechanisms recently haveeen reported to regulate AQP2 trafficking thatay provide potential targets for future NDI
herapies. Both bradykinin and Epac have beenhown to increase AQP2 membrane expres-ion. Bradykinin binds to the B2 receptor andeads to Rho activation, subsequently attenuat-ng AQP2 trafficking by stabilizing polymerizedctin.66 Bradykinin binds 2 receptor subtypes:1 and B2. B2 is expressed constitutively in theenal CD whereas B1 expression is inducible.oth receptors share similar signaling pathways.67
owever, little information is available on theole that the B1 receptor plays in NDI patho-hysiology. The B1 receptor is associated with
he progression of insulin-dependent diabetes pnd has a protective role in renal ischemia. Theevelopment of selective antagonists may helps to better understand its possible link to NDI.QP2 trafficking is additionally affected byAMP activation of the exchange proteinEpac).68 Epac can be activated selectively andirectly by a cAMP analogue (8-pCPT-2=-O-Me-AMP). We speculate that activated Epac ex-hanges bound guanosine diphosphate withTP in both Rap1 and Rap2 proteins, whichlay a role in cytoskeletal rearrangement.
ECHANISMS THAT REGULATEQP2 WHOLE-CELL ABUNDANCE
n addition to controlled AQP2 expression athe cell surface, an increase of AQP2 whole-cellbundance represents an attractive approachor NDI therapy. Indeed, down-regulated AQP2ell surface expression occurring in acquiredDI and in some cases of congenital NDI re-ects down-regulated AQP2 abundance, which
n some patients may limit the efficacy of strat-gies aimed simply at targeting presynthesizedQP2 to the cell surface. Although reduced AQP2bundance is associated with reduced V2R activ-ty in some cases of NDI, such as hypercalce-
ia,69 other conditions of NDI appear to ariserom VP-independent mechanisms. Recent evi-ence has shown that lithium-induced NDI isssociated with an adenylyl cyclase–indepen-ent decrease of AQP2 mRNA expression, possi-ly resulting from decreased AQP2 transcrip-ion.70 In ureteral obstruction, VP-independentown-regulation of AQP2 abundance and cell sur-ace expression was found to arise from increasedOX-2 activity and PGE2 synthesis.71 VP-indepen-ent mechanisms that increase AQP2 abundanceay, thus, prove to be extremely valuable for
esigning new therapeutic strategies to treatDI, as illustrated later.In the kidney, the expression of AQP2 is
estricted to the renal collecting system72,73 ands modulated by both VP and factors that actndependently of VP. Several regulatory motifshat induce AQP2 transcriptional activity haveeen identified in the AQP2 promoter. Theost well documented of these are activatorrotein-1 element and CRE sites that respec-ively bind cAMP-induced c-fos and the phos-
horylated adenosine CRE binding protein
(rlAwhit(tTOssVfsaioNtcaAtaapc
iatdiarVtimtftercbs
AtrmvtemgAut
S
RcptscstFoqsqptmedtewc
R
VP receptor signaling pathways in NDI 275
CREB).74-76 In this respect, p-CREB plays a dualole in regulating AQP2 by inducing its accumu-ation at the cell surface and by enhancingQP2 transcription. AQP2 abundance increasesith interstitial tonicity and recent findingsave shown that this up-regulation arises from
ncreased transcription of the AQP2 gene. Theonicity-responsive enhancer binding proteinTonEBP) has been shown to play a key role inhis event, most likely by binding to at least oneonE element present in the AQP2 promoter.77
f particular interest to the present review, thetimulatory effect of TonEBP on AQP2 tran-cription was found to occur independently ofP.77 Moreover, a stimulatory effect of nuclear
actor of activated T-cells c (NFATc), a tran-cription factor that belongs to the same familys TonEBP, on AQP2 transcription was shownn cultured renal cells together with cross-talkccurring between TonEBP and calcineurin–FATc pathways that further enhance AQP2
ranscription.78 Inhibition of TonEBP activity byalcineurin inhibitors, including cyclosporine And its derivatives, has been shown to reduceQP2 expression.78 Consequently, environmen-
al signals that increase intracellular calcium,nd calcineurin activation in particular, providettractive targets for the promotion of AQP2 ex-ression at the cell surface resulting from in-reased AQP2 whole-cell abundance.
Several pieces of evidence have shown thatn addition to transcriptional regulation, AQP2bundance also is modulated by posttranscrip-ional processing. AQP2 degradation is depen-ent on both lysosomal and proteasomal activ-
ty.79 Observations made from both in vitro andnimal studies indicate that AQP2 protein deg-adation is associated inversely with changes in2R activity.80 Moreover, both dihydrotachys-
erol-treated and fasted animals displayed a VP-ndependent decrease of AQP2 protein but not
RNA abundance, indicating that AQP2 pro-ein degradation is regulated by both VP andactors acting independently of VP.81,82 In addi-ion to controlled AQP2 protein degradation,nhanced translation of AQP2 mRNA may rep-esent another means of increasing AQP2 whole-ell abundance. In vitro studies revealed a feed-ack mechanism that is dependent on tran-
criptional activity, that acts independently ofQP2 degradation, and that involves rapid syn-hesis of regulatory protein(s) that continuouslyeduce AQP2 mRNA translation.83 Aldosteroneay enhance AQP2 protein abundance by alle-
iating such negative control on AQP2 mRNAranslation.83 Further dissection of molecularlements involved in AQP2 degradation andRNA translation may uncover potential tar-
ets delimiting therapies based on controlledQP2 degradation/mRNA translation that wouldltimately increase the expression of AQP2 athe cell surface.
UMMARY
ecent advances in our understanding of theell biology of AQP2 recycling and the signalingathways that lead to the membrane accumula-ion of AQP2 in principal cells have opened upeveral possible strategies for inducing this pro-ess in the absence of conventional vasopressinignaling via its G-protein–coupled receptor,he V2R, which is defective in X-linked NDI.urthermore, these strategies also may apply tother types of NDI, including some of the ac-uired forms. Superimposed on the need totimulate AQP2 membrane trafficking is the re-uirement that sufficient AQP2 be expressed inrincipal cells of NDI patients to achieve effec-ive therapy. We, therefore, also discuss someechanisms that regulate AQP2 expression lev-
ls in target cells. Depending on the nature of theefect leading to NDI, it is likely that a combina-ion of approaches, directed by the basic researchndeavors that are ongoing in many laboratories,ill be required to achieve a positive clinical out-
ome.
EFERENCES1. Deen PM. Mouse models for congenital nephrogenic
diabetes insipidus: what can we learn from them?Nephrol Dial Transplant. 2007;22:1023-6.
2. Wade JB, Stetson DL, Lewis SA. ADH action: evidencefor a membrane shuttle mechanism. Ann NY Acad Sci.1981;372:107-17.
3. Kamsteeg EJ, Heijnen I, van Os CH, et al. The subcel-lular localization of an aquaporin-2 tetramer dependson the stoichiometry of phosphorylated and non-phosphorylated monomers. J Cell Biol. 2000;151:919-30.
4. Yip KP. Coupling of vasopressin-induced intracellularCa2� mobilization and apical exocytosis in perfused
rat kidney collecting duct. J Physiol. 2002;538:891-9.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
276 R. Bouley et al
5. Nickols HH, Shah VN, Chazin WJ, et al. Calmodulininteracts with the V2 vasopressin receptor: elimina-tion of binding to the C terminus also eliminatesarginine vasopressin-stimulated elevation of intracel-lular calcium. J Biol Chem. 2004;279:46969-80.
6. Robben JH, Knoers NV, Deen PM. Cell biologicalaspects of the vasopressin type-2 receptor and aqua-porin 2 water channel in nephrogenic diabetes insip-idus. Am J Physiol Renal Physiol. 2006;291:F257-70.
7. Morello JP, Salahpour A, Laperriere A, et al. Pharma-cological chaperones rescue cell-surface expressionand function of misfolded V2 vasopressin receptormutants. J Clin Invest. 2000;105:887-95.
8. Schulz A, Sangkuhl K, Lennert T, et al. Aminoglyco-side pretreatment partially restores the function oftruncated V(2) vasopressin receptors found in pa-tients with nephrogenic diabetes insipidus. J ClinEndocrinol Metab. 2002;87:5247-57.
9. Deen PTP. Mouse models for congenital nephrogenicdiabetes insipidus: what can we learn from them?Nephrol Dial Transplant. 2007;22:1023-6.
0. Bichet DG, Ruel N, Arthus MF, et al. Rolipram, aphosphodiesterase inhibitor, in the treatment of twomale patients with congenital nephrogenic diabetesinsipidus. Nephron. 56:449-50.
1. Oliveira CJ, Schindler F, Ventura AM, et al. Nitricoxide and cGMP activate the Ras-MAP kinase path-way-stimulating protein tyrosine phosphorylation inrabbit aortic endothelial cells. Free Radic Biol Med.2003;35:381-96.
2. Yamada T, Matsuda K, Uchiyama M. Atrial natriureticpeptide and cGMP activate sodium transport throughPKA-dependent pathway in the urinary bladder of theJapanese tree frog. J Comp Physiol [B]. 2006;176:203-12.
3. Anthony TL, Brooks HL, Boassa D, et al. Cloned hu-man aquaporin-1 is a cyclic GMP-gated ion channel.Mol Pharmacol. 2000;57:576-88.
4. Das S, Garepapaghi M, Palmer LG. Stimulation bycGMP of apical Na channels and cation channels intoad urinary bladder. Am J Physiol Cell Physiol. 1991;260:C234-41.
5. Etgen GJ Jr, Fryburg DA, Gibbs EM. Nitric oxide stimu-lates skeletal muscle glucose transport through a calci-um/contraction- and phosphatidylinositol-3-kinase-inde-pendent pathway. Diabetes. 1997;46:1915-9.
6. Ishikawa Y, Iida H, Ishida H. The muscarinic acetyl-choline receptor-stimulated increase in aquaporin-5levels in the apical plasma membrane in rat parotidacinar cells is coupled with activation of nitric oxide/cGMP signal transduction. Mol Pharmacol. 2002;61:1423-34.
7. Garcia NH, Stoos BA, Carretero OA, et al. Mechanism ofthe nitric oxide-induced blockade of collecting ductwater permeability. Hypertension. 1996;27:679-83.
8. Hirsch JR, Cermak R, Forssmann WG, et al. Effects ofsodium nitroprusside in the rat cortical collectingduct are independent of the NO pathway. Kidney Int.
1997;51:473-6.9. Ortiz PA, Garvin JL. Role of nitric oxide in the regu-lation of nephron transport. Am J Physiol RenalPhysiol. 2002;282:F777-84.
0. Martin PY, Bianchi M, Roger F, et al. Arginine vaso-pressin modulates expression of neuronal NOS in ratrenal medulla. Am J Physiol Renal Physiol. 2002;283:F559-68.
1. Shin SJ, Lai FJ, Wen JD, et al. Increased nitric oxidesynthase mRNA expression in the renal medulla ofwater-deprived rats. Kidney Int. 1999;56:2191-202.
2. Morishita T, Tsutsui M, Shimokawa H, et al. Nephro-genic diabetes insipidus in mice lacking all nitricoxide synthase isoforms. Proc Natl Acad Sci U S A.2005;102:10616-21.
3. Pilz RB, Casteel DE. Regulation of gene expression bycyclic GMP. Circ Res. 2003;93:1034-46.
4. Bouley R, Breton S, Sun T, et al. Nitric oxide and atrialnatriuretic factor stimulate cGMP-dependent membraneinsertion of aquaporin 2 in renal epithelial cells. J ClinInvest. 2000;106:1115-26.
5. Wang W, Li C, Nejsum LN, et al. Biphasic effects ofANP infusion in conscious, euvolumic rats: roles ofAQP2 and ENaC trafficking. Am J Physiol RenalPhysiol. 2006;290:F530-41.
6. Bouley R, Pastor-Soler N, Cohen O, et al. Stimulationof AQP2 membrane insertion in renal epithelial cellsin vitro and in vivo by the cGMP phosphodiesteraseinhibitor sildenafil citrate (Viagra). Am J Physiol RenalPhysiol. 2005;288:F1103-12.
7. Brodsky FM, Chen CY, Knuehl C, et al. Biologicalbasket weaving: formation and function of clathrin-coated vesicles. Annu Rev Cell Dev Biol. 2001;17:517-68.
8. Conner SD, Schmid SL. Regulated portals of entry intothe cell. Nature. 2003;422:37-44.
9. Brown D, Orci L. Vasopressin stimulates formation ofcoated pits in rat kidney collecting ducts. Nature.1983;302:253-5.
0. Sun TX, Van Hoek A, Huang Y, et al. Aquaporin-2 local-ization in clathrin-coated pits: inhibition of endocytosisby dominant-negative dynamin. Am J Physiol RenalPhysiol. 2002;282:F998-1011.
1. Lu HA, Sun TX, Matsuzaki T, et al. Heat shock protein70 interacts with aquaporin-2 and regulates its traf-ficking. J Biol Chem. 2007;282:28721-32.
2. Kamsteeg EJ, Duffield AS, Konings IB, et al. MALdecreases the internalization of the aquaporin-2 waterchannel. Proc Natl Acad Sci U S A. 2007;104:16696-701.
3. Nejsum LN, Zelenina M, Aperia A, et al. Bidirectionalregulation of AQP2 trafficking and recycling: involve-ment of AQP2-S256 phosphorylation. Am J PhysiolRenal Physiol. 2005;288:F930-8.
4. Procino G, Carmosino M, Marin O, et al. Ser-256phosphorylation dynamics of aquaporin 2 during mat-uration from the ER to the vesicular compartment inrenal cells. FASEB J. 2003;17:1886-8.
5. Lu H, Sun TX, Bouley R, et al. Inhibition of endocyto-
sis causes phosphorylation (S256)-independent plasma
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
VP receptor signaling pathways in NDI 277
membrane accumulation of AQP2. Am J Physiol RenalPhysiol. 2004;286:F233-43.
6. Russo LM, McKee M, Brown D. Methyl-beta-cyclodex-trin induces vasopressin-independent apical accumu-lation of aquaporin-2 in the isolated, perfused rat kid-ney. Am J Physiol Renal Physiol. 2006;291:F246-53.
7. Christensen BM, Zelenina M, Aperia A, et al. Localiza-tion and regulation of PKA-phosphorylated AQP2 inresponse to V(2)-receptor agonist/antagonist treat-ment. Am J Physiol Renal Physiol. 2000;278:F29-42.
8. Bouley R, Hawthorn G, Russo LM, et al. Aquaporin 2(AQP2) and vasopressin type 2 receptor (V2R) endo-cytosis in kidney epithelial cells: AQP2 is located in‘endocytosis-resistant’ membrane domains after vaso-pressin treatment. Biol Cell. 2006;98:215-32.
9. Knepper MA, Nielsen S. Kinetic model of water andurea permeability regulation by vasopressin in col-lecting duct. Am J Physiol Renal Physiol. 1993;265:F214-24.
0. Fushimi K, Sasaki S, Marumo F. Phosphorylation ofserine 256 is required for cAMP-dependent regulatoryexocytosis of the aquaporin-2 water channel. J BiolChem. 1997;272:14800-4.
1. Katsura T, Gustafson CE, Ausiello DA. Protein kinaseA phosphorylation is involved in regulated exocytosisof aquaporin-2 in transfected LLC-PK1 cells. Am JPhysiol Renal Physiol. 1997;272:F817-22.
2. Stefan E, Wiesner B, Baillie GS, et al. Compartmental-ization of cAMP-dependent signaling by phosphodies-terase-4D is involved in the regulation of vasopressin-mediated water reabsorption in renal principal cells.J Am Soc Nephrol. 2007;18:199-212.
3. Brown D. The ins and outs of aquaporin-2 trafficking.Am J Physiol Renal Physiol. 2003;284:F893-901.
4. Tobert JA. Lovastatin and beyond: the history of theHMG-CoA reductase inhibitors. Nat Rev Drug Discov.2003;2:517-26.
5. Sidaway JE, Davidson RG, McTaggart F, et al. Inhibi-tors of 3-hydroxy-3-methylglutaryl-CoA reductase re-duce receptor-mediated endocytosis in opossum kid-ney cells. J Am Soc Nephrol. 2004;15:2258-65.
6. Verhulst A, D’Haese PC, De Broe ME. Inhibitors ofHMG-CoA reductase reduce receptor-mediated endo-cytosis in human kidney proximal tubular cells. J AmSoc Nephrol. 2004;15:2249-57.
7. McFarlane SI, Muniyappa R, Francisco R, et al. Clinicalreview 145: pleiotropic effects of statins: lipid reduc-tion and beyond. J Clin Endocrinol Metab. 2002;87:1451-8.
8. Fried LF, Orchard TJ, Kasiske BL. Effect of lipid re-duction on the progression of renal disease: a meta-analysis. Kidney Int. 2001;59:260-9.
9. Kureishi Y, Luo Z, Shiojima I, et al. The HMG-CoAreductase inhibitor simvastatin activates the proteinkinase Akt and promotes angiogenesis in normocho-lesterolemic animals. Nat Med. 2000;6:1004-10.
0. Urbich C, Dernbach E, Zeiher AM, et al. Double-edged role of statins in angiogenesis signaling. Circ
Res. 2002;90:737-44.1. Laufs U, La Fata V, Plutzky J, et al. Upregulation ofendothelial nitric oxide synthase by HMG CoA reduc-tase inhibitors. Circulation. 1998;97:1129-35.
2. Park HJ, Kong D, Iruela-Arispe L, et al. 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors inter-fere with angiogenesis by inhibiting the geranylgera-nylation of RhoA. Circ Res. 2002;91:143-50.
3. Tamma G, Klussmann E, Maric K, et al. Rho inhibitscAMP-induced translocation of aquaporin-2 into theapical membrane of renal cells. Am J Physiol RenalPhysiol. 2001;281:F1092-101.
4. Klussmann E, Maric K, Wiesner B, et al. Protein kinaseA anchoring proteins are required for vasopressin-mediated translocation of aquaporin-2 into cell mem-branes of renal principal cells. J Biol Chem. 1999;274:4934-8.
5. Klussmann E, Tamma G, Lorenz D, et al. An inhibitoryrole of Rho in the vasopressin-mediated translocationof aquaporin-2 into cell membranes of renal principalcells. J Biol Chem. 2001;276:20451-7.
6. Tamma G, Klussmann E, Procino G, et al. cAMP-induced AQP2 translocation is associated with RhoAinhibition through RhoA phosphorylation and inter-action with RhoGDI. J Cell Sci. 2003;116:1519-25.
7. Noda Y, Sasaki S. Regulation of aquaporin-2 traffick-ing and its binding protein complex. Biochim Bio-phys Acta. 2006;1758:1117-25.
8. Cox RA, Kanagalingam K. A spectrophotometricstudy of the denaturation of deoxyribonucleic acid inthe presence of urea or formaldehyde and its rele-vance to the secondary structure of single-strandedpolynucleotides. Biochem J. 1968;108:599-610.
9. Kotnik P, Nielsen J, Kwon TH, et al. Altered expres-sion of COX-1, COX-2, and mPGES in rats with neph-rogenic and central diabetes insipidus. Am J PhysiolRenal Physiol. 2005;288:F1053-68.
0. Rao R, Zhang MZ, Zhao M, et al. Lithium treatmentinhibits renal GSK-3 activity and promotes cyclooxy-genase 2-dependent polyuria. Am J Physiol RenalPhysiol. 2005;288:F642-9.
1. Pattaragarn A, Alon US. Treatment of congenital neph-rogenic diabetes insipidus by hydrochlorothiazideand cyclooxygenase-2 inhibitor. Pediatr Nephrol.2003;18:1073-6.
2. Johnsen SP, Larsson H, Tarone RE, et al. Risk ofhospitalization for myocardial infarction among usersof rofecoxib, celecoxib, and other NSAIDs: a popula-tion-based case-control study. Arch Intern Med. 2005;165:978-84.
3. Samuelsson B, Morgenstern R, Jakobsson PJ. Mem-brane prostaglandin E synthase-1: a novel therapeutictarget. Pharmacol Rev. 2007;59:207-24.
4. Takeuchi K, Takahashi N, Abe T, et al. Two isoformsof the rat kidney EP3 receptor derived by alternativeRNA splicing: intrarenal expression co-localization.Biochem Biophys Res Commun. 1994;199:834-40.
5. Makino H, Tanaka I, Mukoyama M, et al. Prevention of
diabetic nephropathy in rats by prostaglandin E re-
6
6
6
6
7
7
7
7
7
7
7
7
7
7
8
8
8
8
278 R. Bouley et al
ceptor EP1-selective antagonist. J Am Soc Nephrol.2002;13:1757-65.
6. Tamma G, Carmosino M, Svelto M, et al. Bradykininsignaling counteracts cAMP-elicited aquaporin 2 trans-location in renal cells. J Am Soc Nephrol. 2005;16:2881-9.
7. Leeb-Lundberg LM, Marceau F, Muller-Esterl W, et al.International union of pharmacology. XLV. Classifica-tion of the kinin receptor family: from molecular mech-anisms to pathophysiological consequences. Pharma-col Rev. 2005;57:27-77.
8. Fukuhara S, Sakurai A, Sano H, et al. Cyclic AMPpotentiates vascular endothelial cadherin-mediatedcell-cell contact to enhance endothelial barrier func-tion through an Epac-Rap1 signaling pathway. MolCell Biol. 2005;25:136-46.
9. Procino G, Carmosino M, Tamma G, et al. Extracellu-lar calcium antagonizes forskolin-induced aquaporin2 trafficking in collecting duct cells. Kidney Int. 2004;66:2245-55.
0. Li Y, Shaw S, Kamsteeg EJ, et al. Development oflithium-induced nephrogenic diabetes insipidus is dis-sociated from adenylyl cyclase activity. J Am SocNephrol. 2006;17:1063-72.
1. Norregaard R, Jensen BL, Li C, et al. COX-2 inhibitionprevents downregulation of key renal water and so-dium transport proteins in response to bilateral ure-teral obstruction. Am J Physiol Renal Physiol. 2005;289:F322-33.
2. Fushimi K, Uchida S, Hara Y, et al. Cloning andexpression of apical membrane water channel of ratkidney collecting tubule. Nature. 1993;361:549-52.
3. Nielsen S, DiGiovanni SR, Christensen EI, et al. Cellu-lar and subcellular immunolocalization of vasopres-sin-regulated water channel in rat kidney. Proc NatlAcad Sci U S A. 1993;90:11663-7.
4. Yasui M, Zelenin SM, Celsi G, et al. Adenylate cyclase-
coupled vasopressin receptor activates AQP2promoter via a dual effect on CRE and AP1 elements.Am J Physiol Renal Physiol. 1997;272:F443-50.
5. Hozawa S, Holtzman EJ, Ausiello DA. cAMP motifsregulating transcription in the aquaporin-2 gene. Am JPhysiol Cell Physiol. 1996;270:C1695-702.
6. Frokiaer J, Marples D, Valtin H, et al. Low aquaporin-2levels in polyuric DI �/� severe mice with constitu-tively high cAMP-phosphodiesterase activity. Am JPhysiol Renal Physiol. 1999;276:F179-90.
7. Hasler U, Jeon US, Kim JA, et al. Tonicity-responsiveenhancer binding protein is an essential regulator ofaquaporin-2 expression in renal collecting duct prin-cipal cells. J Am Soc Nephrol. 2006;17:1521-31.
8. Li SZ, McDill BW, Kovach PA, et al. Calcineurin-NFATc signaling pathway regulates AQP2 expressionin response to calcium signals and osmotic stress. Am JPhysiol Cell Physiol. 2007;292:C1606-16.
9. Hasler U, Mordasini D, Bens M, et al. Long termregulation of aquaporin-2 expression in vasopressin-responsive renal collecting duct principal cells. J BiolChem. 2002;277:10379-86.
0. Hasler U, Nielsen S, Feraille E, et al. Posttranscrip-tional control of aquaporin-2 abundance by vasopres-sin in renal collecting duct principal cells. Am J PhysiolRenal Physiol. 2006;290:F177-87.
1. Puliyanda DP, Ward DT, Baum MA, et al. Calpain-mediated AQP2 proteolysis in inner medullary col-lecting duct. Biochem Biophys Res Commun. 2003;303:52-8.
2. Wilke C, Sheriff S, Soleimani M, et al. Vasopressin-independent regulation of collecting duct aquaporin-2in food deprivation. Kidney Int. 2005;67:201-16.
3. Hasler U, Mordasini D, Bianchi M, et al. Dual influ-ence of aldosterone on AQP2 expression in culturedrenal collecting duct principal cells. J Biol Chem.
2003;278:21639-48.