pkd1 haploinsufficiency causes a syndrome of inappropriate

PKD1 Haploinsufficiency Causes a Syndrome ofInappropriate Antidiuresis in Mice

Ali K. Ahrabi,* Sara Terryn,† Giovanna Valenti,‡ Nathalie Caron,§

Claudine Serradeil-Le Gal,� Danielle Raufaste,� Soren Nielsen,¶ Shigeo Horie,**Jean-Marc Verbavatz,†† and Olivier Devuyst**Division of Nephrology, Universite catholique de Louvain Medical School, Brussels, Belgium; †Laboratory of CellPhysiology, Center for Environmental Sciences, Hasselt University, Diepenbeek, Belgium; ‡Department of Physiology,University of Bari, Bari, Italy; §Department of Physiology and Pharmacology, University of Mons-Hainaut, Mons,Belgium; �Sanofi-Aventis, Toulouse, France; ¶The Water and Salt Research Center, University of Aarhus, Aarhus,Denmark; **Department of Urology, Teikyo University, Tokyo, Japan; and ††Cell and Molecular Imaging, CEA/Saclay,Gif-sur-Yvette, France

Mutations in PKD1 are associated with autosomal dominant polycystic kidney disease. Studies in mouse models suggest thatthe vasopressin (AVP) V2 receptor (V2R) pathway is involved in renal cyst progression, but potential changes beforecystogenesis are unknown. This study used a noncystic mouse model to investigate the effect of Pkd1 haploinsufficiency onwater handling and AVP signaling in the collecting duct (CD). In comparison with wild-type littermates, Pkd1�/� miceshowed inappropriate antidiuresis with higher urine osmolality and lower plasma osmolality at baseline, despite similar renalfunction and water intake. The Pkd1�/� mice had a decreased aquaretic response to both a water load and a selective V2Rantagonist, despite similar V2R distribution and affinity. They showed an inappropriate expression of AVP in brain,irrespective of the hypo-osmolality. The cAMP levels in kidney and urine were unchanged, as were the mRNA levels ofaquaporin-2 (AQP2), V2R, and cAMP-dependent mediators in kidney. However, the (Ser256) phosphorylated AQP2 wasupregulated in Pkd1�/� kidneys, with AQP2 recruitment to the apical plasma membrane of CD principal cells. The basalintracellular Ca2� concentration was significantly lower in isolated Pkd1�/� CD, with downregulated phosphorylatedextracellular signal–regulated kinase 1/2 and decreased RhoA activity. Thus, in absence of cystic changes, reduced Pkd1 genedosage is associated with a syndrome of inappropriate antidiuresis (positive water balance) reflecting decreased intracellularCa2� concentration, decreased activity of RhoA, recruitment of AQP2 in the CD, and inappropriate expression of AVP in thebrain. These data give new insights in the potential roles of polycystin-1 in the AVP and Ca2� signaling and the traffickingof AQP2 in the CD.

J Am Soc Nephrol 18: 1740–1753, 2007. doi: 10.1681/ASN.2006010052

A utosomal dominant polycystic kidney disease (AD-PKD) is the most frequent inherited nephropathy andan important cause of ESRD (1). Mutations in two

genes, PKD1 and PKD2, have been associated with ADPKD.Mutations in PKD1 account for approximately 85% of the af-fected families, and they are associated with a renal disease thatprogresses more rapidly than in families with PKD2 (2). PKD1and PKD2 encode integral membrane proteins, polycystin-1and polycystin-2, that interact in renal primary cilia and regu-late the proliferation and differentiation of renal tubular cellsvia different signaling pathways (3). Mutations in PKD1/PKD2disrupt these pathways, leading to cystogenesis by a combina-tion of increased cellular proliferation, abnormal fluid secre-

tion, and dedifferentiation (1,3). All nephron segments may beinvolved in cyst formation in ADPKD, but an important frac-tion of the cysts is derived from the collecting ducts (CD) (4,5).

In vitro studies have shown that cAMP plays a major role incystogenesis. Exposure to cAMP agonists stimulates fluid se-cretion across monolayers of ADPKD cyst-lining epithelial cells(6), as well as the proliferation of these cells (7). Furthermore,increased levels of cAMP resulting from the activation of vaso-pressin (AVP) V2 receptor (V2R) pathway in CD cells maycontribute to the progression of cystogenesis. In two cysticmodels that are orthologous to human autosomal recessivePKD (PCK rat) and nephronophthisis (pcy mouse) and onecystic model that is orthologous to human ADPKD type 2(Pkd2�/tm1Som mouse), increased renal cAMP levels comparedwith normal mice, paralleled with higher expression of aqua-porin-2 (AQP2) and V2R, have been reported (8–10). The ad-ministration of V2R antagonists to these models lowered renalcAMP and inhibited the development and progression of es-tablished renal cystic disease (8–10), motivating trials to test theefficacy of V2R antagonists in patients with ADPKD (11). It is

Received January 18, 2006. Accepted March 9, 2007.

Published online ahead of print. Publication date available at www.jasn.org.

Address correspondence to: Dr. Olivier Devuyst, Division of Nephrology, UCLMedical School, 10 Avenue Hippocrate, B-1200 Brussels, Belgium. Phone: �32-2-764-5450; Fax: �32-2-764-5455; E-mail: [email protected]

Copyright © 2007 by the American Society of Nephrology ISSN: 1046-6673/1806-1740

important to note that all rodent models tested so far developrenal cysts (and subsequent renal failure) within a few weeks ofage.

In normal CD cells, the stimulation of V2R by AVP leads tothe phosphorylation of AQP2 on the Ser256 residue and itssubsequent insertion in the apical plasma membrane, an essen-tial step to mediate final urine concentration (12,13). A mildimpairment in urinary concentrating ability, with increasedcirculating AVP levels, has been described in patients withADPKD and cystic kidneys (11,14,15). However, this urinaryconcentrating abnormality is probably not specific, because anymodification of the medullary architecture (e.g., cystic changes)impairs the constitution of the corticomedullary osmotic gradi-ent, resulting in nephrogenic diabetes insipidus (16). Consider-ing that an activation of the V2R pathway has been involved inPKD mouse models with cysts originating from the CD (8–10),we hypothesized that the complex chain of events that mediatesurinary concentration in the CD could be modified early, beforecystogenesis.

In this study, we used a well-established mouse model witha targeted deletion of Pkd1 (Pkd1�/�) (17) to test whether Pkd1haploinsufficiency causes abnormal water handling and AVPsignaling in the CD before cystogenesis and renal failure. Likeother Pkd1-null mutants, the homozygous Pkd1�/� mice die inutero with massive cystic kidneys, hydrops fetalis, and cardio-vascular defects (17–20). By contrast, there is no consistentphenotype in heterozygous Pkd1�/� mice that do not developrenal cysts until (in a few individuals) a very old age (21,22).Our investigations reveal for the first time that reduced Pkd1gene dosage results in inappropriate antidiuresis and positivewater balance, reflecting decreased intracellular calcium([Ca2�]i) levels with lowered RhoA activity and recruitment ofAQP2 in the apical membrane of CD principal cells and inap-propriate expression of AVP in the brain.

Materials and MethodsPkd1 Mice and Sampling

Experiments were conducted on age- and gender-matched adultmice (aged 20 to 35 wk) with a targeted deletion of the exons 2 to 5 andpart of exon 6 of Pkd1, resulting in a null allele (17). The mice weremaintained on a mixed 129/sv/C57BL/6J background. They werehoused in light- and temperature-controlled room with ad libitum accessto tap water and standard chow (Pavan, Oud-Turnhout, Belgium).Previous experiments showed that the Pkd1 mice had similar heart rateand BP (N. Morel, et al., unpublished data, 2007). Water handling atbaseline and during various protocols was assessed in individual met-abolic cages, after appropriate training. Blood and tissue samples wereobtained at time of killing, after anesthesia with Sevoflurane (Abbott,Ottignies, Belgium) and exsanguination. Blood was collected by venouspuncture, and plasma samples were kept at �20°C. The samplingprocedures were exactly similar in both groups. Tissue samples wereimmediately processed for fixation and mRNA/protein extraction. Theexperiments were conducted in accordance with the National ResearchCouncil Guide for the Care and Use of Laboratory Animals and wereapproved by the local Animal Ethics Committee.

Water Handling ProtocolsPlasma samples and 24-h urine collections were obtained at baseline,

and the urinary concentrating ability was tested after 24-h water de-

privation. The capacity to excrete a water load was tested after intra-peritoneal injection of 2 ml of sterile water; urine was collected undera plastic-wrapped container on an hourly basis for the next 6 h. Theaquaretic effect of the V2R antagonist SR121463B (Sanofi-Aventis,Chilly-Mazarin, France), which has a high affinity for renal V2R fromseveral species, including rat, mouse, and human (Ki � 0.26 � 0.04 nM)(23), was tested after intraperitoneal administration of dosages thatranged from 0.1 to 30 mg/kg and hourly determination of diuresis forthe next 6 h as described above.

V2R Binding Assays and AutoradiographyRenomedullary preparations from Pkd1�/� and Pkd1�/� mice (or

CHO membranes that expressed human V2R used as positive controls)were incubated in a 50-mM Tris-HCl buffer (pH 8.1) that contained 2mM MgCl2, 1 mM EDTA, 0.1% BSA, 0.1% bacitracin, and [3H]SR121463(0.8 to 28 nM for saturation experiments or 2 nM for binding studies).The reaction was started by the addition of membranes (7.5 �g/assayfor CHO and 100 to 130 �g/assay for mouse renal tissue) and incuba-tion for 45 min at 25°C, stopped by filtration through Whatman GF/Bfilters as described previously (24). Nonspecific binding was deter-mined in the presence of 1 �M SR121463B. Data for equilibrium bind-ing (apparent equilibrium dissociation constant [Kd] and maximumbinding density [Bmax]) were calculated using an interactive nonlinearregression program (25).

For performance of autoradiography, kidneys from Pkd1 mice werefrozen at �40°C in isopentane and further stored at �80°C. Serialsections (15 �m) were mounted onto gelatin chrome-alum slides, rinsedto eliminate endogenous AVP, and incubated with 1.5 nM[3H]SR121463 alone (total binding) and in the presence of 1 �M unla-beled SR121463B or AVP (nonspecific binding) as described previously(24). After incubation, the sections were washed three times for 10 mineach in ice-cold binding buffer, dipped in distilled water, and driedunder a stream of cold air. Rinsed labeled sections were placed on aphosphor-imaging plate for 4 d and further analyzed with a BAS5000Bio-Image Analyser (Fuji, Tokyo, Japan). SR121463B, monophosphatesalt, and [3H]SR121463 (47.5 Ci/mmol) were synthesized at Sanofi-Aventis, whereas AVP was obtained from Sigma Chemical Co. (L’Isled’Abeau, France).

Plasma and Urine AnalysesSodium, urea, creatinine, and calcium were measured using a Kodak

Ektachem DT60II Analyzer (Johnson & Johnson, New Brunswick, NJ),and osmolality was measured using a Fiske Osmometer (NeedhamHeights, MA). The nitrite/nitrate (NOx) concentrations were measuredin urine and plasma using a colorimetric assay (Cayman Chemical, AnnArbor, MI). Because sevoflurane may induce the release of AVP,thereby increasing plasma values in our protocols, we measured urineAVP levels, which were not obtained under sevoflurane anesthesia,using RIA (Peninsula Laboratories, San Carlos, CA). For cAMP deter-minations, whole kidneys were ground under liquid nitrogen andhomogenized in 10 volumes of 0.1 M HCl. The homogenate was cen-trifuged at 600 � g for 10 min, and the supernatant was collected,diluted (1:10) in 0.1 M HCl, and processed with acetylation using anenzyme immunoassay kit (Sigma-Aldrich, St. Louis, MO). The urinesamples were diluted (1:5000) in 0.1 M HCl and were processed with-out acetylation. The urinary prostaglandin E2 (PGE2) was measured byEIA (Amersham Biosciences, Piscataway, NJ).

Reverse Transcription–PCR and Real-Time ReverseTranscription–PCR

Total RNA from mouse kidney and brain (26) was extracted withTrizol (Invitrogen, Merelbeke, Belgium), treated with DNase I, and

J Am Soc Nephrol 18: 1740–1753, 2007 PKD1 Haploinsufficiency Causes SIAD 1741

reverse-transcribed into cDNA. The primers (Supplementary Table 1)were designed using Beacon Designer 2.0 (Premier Biosoft Interna-tional, Palo Alto, CA). Changes in target gene mRNA levels weredetermined by semiquantitative real-time reverse transcription–PCR(RT-PCR) with an iCycler IQ System (Bio-Rad Laboratories, Hercules,CA) using SYBR Green I. Real-time semiquantitative PCR analyseswere performed in duplicate as described previously (27). The PCRconditions were 94°C for 3 min followed by 31 cycles of 30 s at 95°C,30 s at 61°C and 1 min at 72°C. Negative controls excluded amplifica-tion from genomic DNA. For each assay, standard curves were pre-pared by serial four-fold dilutions of cDNA samples. The efficiency ofthe reactions was calculated from the slope of the standard curve[efficiency � (10�1/slope) � 1] (27).

AntibodiesRabbit polyclonal antibodies against AQP2 (Sigma-Aldrich), Ser256

phosphorylated AQP2 (p-AQP2) (12), AQP1 (Chemicon, Temecula,CA), AQP3 (a gift from J.-M. Verbavatz, CEA Saclay), extracellularsignal–regulated kinase 1/2 (ERK1/2; C16) and Tyr204 p-ERK1/2(Santa Cruz Biotechnologies, Santa Cruz, CA), mouse monoclonalRhoA and Ser188 p-RhoA (Santa Cruz Biotechnologies), and mousemAb against �-actin (Sigma-Aldrich) were used.

ImmunoblottingKidneys were ground under liquid nitrogen and homogenized as

described previously (27). The homogenate was centrifuged at 1000 �

g for 15 min at 4°C. The resulting supernatant was either kept at �80°C(as the “total extract” fraction) or centrifuged at 100,000 � g for 120 minat 4°C. The pellet (“membrane” fraction) was suspended in homogeni-zation buffer before determination of protein concentration and storageat �80°C. SDS-PAGE was performed under reduced (kidney) or non-reduced (urine) conditions. After blotting on nitrocellulose, the mem-branes were incubated overnight at 4°C with primary antibodies,washed, incubated for 1 h at room temperature with peroxidase-labeledantibodies (Dako, Glostrup, Denmark), and visualized with enhancedchemiluminescence. Normalization for �-actin was obtained after strip-ping and reprobing. Densitometry analysis was performed with aStudioStar Scanner (Agfa-Gevaert, Mortsel, Belgium) using the NIH-Image V1–57 software.

ImmunohistochemistryKidney samples were fixed in 4% paraformaldehyde (Boehringer

Ingelheim, Heidelberg, Germany) in 0.1 mol/L phosphate buffer (pH7.4) before embedding in paraffin. The 6-�m sections were stained withhemalum-eosin or incubated for 30 min with 0.3% H2O2, followed by 20min with 10% normal serum, and 45 min with the primary antibodiesdiluted in PBS that contained 2% BSA. After washing, sections weresuccessively incubated with biotinylated secondary anti-IgG antibod-ies, avidin-biotin peroxidase, and aminoethylcarbazole (VectastainElite; Vector Laboratories, Burlingame, CA). Sections were viewed un-der a Leica DMR coupled to a Leica DC300 digital camera (Leica,Heerbrugg, Switzerland).

Immunoelectron MicroscopyFor electron microscopy, kidney samples from Pkd1 mice were fixed

overnight in 4% paraformaldehyde and 0.1% glutaraldehyde in PBSand washed in PBS. Small samples, including the outer medulla and thetop of inner medulla, were embedded in unicryl, and 80-nm-thicksections were cut. Sections were preincubated in 20 mM Tris buffer (pH7.5) that contained 0.1% BSA, 0.1% fish gelatin, and 0.05% Tween 20(buffer-T), followed by a 90-min incubation in the same buffer-T that

contained a 1:100 dilution of anti-AQP2 polyclonal antibodies. Sectionswere washed three times in buffer-T, then incubated in a 1:25 dilutionof 10 nm of gold-conjugated secondary antibodies for 45 min. Afterwashing in Tris, sections were stained with uranyl-acetate and leadcitrate and photographed on a Philips EM 400 microscope (FEI, Eind-hoven, Netherlands). Three samples from three pairs of mice wereprocessed, and at least 10 pictures of outer medullary CD principal cellswere randomly taken for each sample. The data are expressed asnumber of gold particles per micron of apical membrane length. Ultra-structural examination of the vasa recta was performed on three pairsof kidney slices and fixed overnight in 2% glutaraldehyde before wash-ing in PBS.

Measurement of RhoA ActivityQuantification of active RhoA (GTP-bound) was measured by selec-

tive affinity precipitation of GTP-Rho (Upstate, Temecula, CA), follow-ing the procedure described in detail previously (28,29). The kidneyswere ground under liquid nitrogen and homogenized in ice-cold 1�

Mg2� lysis/wash buffer according to the manufacturer’s instructions(Upstate, Temecula, CA). The homogenates were centrifuged at15,000 � g for 30 min, and the supernatant of each sample was col-lected. A total of 30 �g of GST-tagged fusion protein, corresponding toresidues 7 to 89 of mouse Rhotekin Rho Binding Domain, bound toglutathione-agarose beads, was added to the supernatant of each sam-ple (500-�l exact) and were rotated overnight at 4°C. Beads werewashed three times with Mg2� lysis/wash buffer, and bound proteinswere separated by SDS-PAGE and detected by Western blotting usinga monoclonal RhoA antibody (28,29).

Measurement of Intracellular Ca2� ConcentrationInner medullary collecting ducts (IMCD) were isolated from collag-

enase-digested medulla from three pairs of Pkd1�/� and Pkd1�/� mousekidneys. The tubule segments were seeded onto glass coverslips andexamined using an inverted Nikon TMD35 epifluorescence microscope(Analis, Namur, Belgium) in a thermostated chamber at 37°C. Theintracellular Ca2� concentration ([Ca2�]i) was measured as describedpreviously (30). Briefly, after measurement of the background signal,isolated tubules were loaded with Fura-2 by incubation with the mem-brane-permeant acetoxymethyl (AM) ester form of the dye (10 �M) for1 h at 37°C. Tubules were excited at 340 and 380 nm, and the fluores-cence emission was recorded at 510 nm. Data collection time for animage was 2 s. Fura-2 was calibrated in vivo at the end of each exper-iment, according to the equation derived by [Ca2�]i � Kd Rbf [(r �

rmin)/(rmax � r)], where Kd is the dissociation constant of Fura-2 forCa2� (135 nM), Rbf is the maximum fluorescence intensity as a result ofexcitation at 380 nm (in the absence of Ca2�) divided by the minimumfluorescence intensity at 380 nm (in the presence of saturating Ca2�), ris the F340/F380 fluorescence ratio, rmax and rmin are the F340/F380 fluo-rescence ratios in the presence of saturating Ca2� and in the absence ofCa2�, respectively. rmax was obtained by permeabilization of the tu-bules with Ca2� ionophore ionomycin (10 �M), in the presence of 5 mMextracellular Ca2�. For obtaining subsequently the minimum ratio rmin,the tubules were exposed to a Ca2� free solution (containing 10 mMEGTA) with 10 �M ionomycin and 1,2-bis(o-aminophenoxy)ethane-N,N,N�,N�-tetraacetic acid (BAPTA)-acetoxymethyl (10 �M) to bufferintracellular Ca2�.

Statistical AnalysesData are means � SEM. LogED50 (the dosage of agonist that pro-

vokes 50% of the maximum response) values were calculated by non-linear curve fitting of the individual concentration-effect curve data

1742 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 1740–1753, 2007

(GraphPad, San Diego, CA). Comparisons between groups were per-formed using two-tailed unpaired t test. Significance level was P � 0.05.

ResultsKidney Structure and Baseline Parameters

Macroscopic and histology analyses (Supplementary Figure1) confirmed the normal kidney structure in heterozygousPkd1�/� versus Pkd1�/� mice (20 to 35 wk old). In particular, nocysts or tubular dilations were observed in any segment of thePkd1�/� kidneys. The baseline clinical and biologic parametersare shown in Table 1. The Pkd1�/� mice had similar body andkidney weight as wild-type (WT) littermates and similarplasma urea and creatinine values. The Pkd1�/� mice showed athree-fold decrease in urinary output, associated with a higherurine sodium/osmolality and a lower plasma sodium/osmola-lity in comparison with WT littermates, despite similar waterintake. These observations were confirmed in other sets of adultmice, irrespective of gender. The Pkd1�/� mice were also char-acterized by significantly lower cumulative urinary excretion ofAVP, calcium, and NOx and a trend for lower urinary PGE2,whereas the cAMP levels in kidney and urine and the plasmaNOx were unchanged. The urinary concentrations of calciumand AVP both were higher in Pkd1�/� mice, reflecting the netwater reabsorption. These data show that, in the absence ofcystic changes and renal failure, the Pkd1�/� mice are in posi-tive water balance, with similar water intake but lower plasmasodium and osmolality.

Urinary Concentrating Ability and Water HandlingA 24-h water deprivation was performed to assess the uri-

nary concentrating ability in the Pkd1 mice (Figure 1, A and B).

Confirming the previous observations, the Pkd1�/� mice had alower urine output and higher urine osmolality at baseline.Water deprivation resulted in a similar weight loss (averaging14 � 0.7% in Pkd1�/� and 13 � 0.2% in Pkd1�/�; n � 4 pairs),but the urinary concentrating ability was significantly higher inPkd1�/� mice, as indicated by lower volume and higher urineosmolality at the end of the test.

A test of acute water loading (2 ml intraperitoneally [approx-imately 70 ml/kg]) was performed to investigate the capacity ofPkd1 mice to eliminate water during a 6-h period (Figure 1C). Incomparison with WT littermates, Pkd1�/� mice showed a sig-nificant decrease in their ability to excrete water up to 3 h afterthe water load. Although Pkd1�/� mice were able to excretemore water than the WT mice during the last 3 h of the test, thetotal excreted volume of water after 6 h was slightly lower (total6-h urine output: 1720 � 70 �l in Pkd1�/� mice versus 1894 � 88�l in Pkd1�/� mice; n � 9 pairs; P � 0.88). Thus, the Pkd1�/�

mice have an inappropriate antidiuresis at baseline, a higherability to concentrate urine when challenged by water depriva-tion, and a decreased ability to excrete a water load.

Characterization of Renal V2R and Response to V2RAntagonist

To characterize the distribution and affinity of V2R, we per-formed autoradiography of Pkd1�/� and Pkd1�/� kidneys thatwere incubated with a highly selective V2R ligand, alone orwith an excess of unlabeled ligand or AVP (Figure 2A). Weobserved a dense specific labeling confined in the inner/outermedulla and papilla area, corresponding to the main localiza-tion of the V2R in rodent CD. The binding pattern, as well as

Table 1. Baseline parameters, renal function, and water metabolisma

Parameter Pkd1�/� Pkd1�/� P

Body weight (g) 28.7 � 0.12 (7) 27.9 � 0.43 (7) NSKidney weight (% body wt) 1.38 � 0.05 (7) 1.23 � 0.05 (7) NSPlasma urea (mg/dl) 27.4 � 1.91 (7) 28.4 � 1.17 (7) NSPlasma creatinine (mg/dl) 0.27 � 0.02 (7) 0.29 � 0.01 (7) NSWater intake (ml/24 h) 5.7 � 0.3 (15) 5.5 � 0.4 (15) NSUrine volume (�l/24 h) 1315 � 140 (15) 413 � 26 (15) �0.0001Urine osmolality (mosmol/kgH2O) 3034 � 227 (15) 4880 � 179 (15) �0.0001Urine sodium (mM) 128.0 � 13.4 (7) 260.0 � 20.6 (7) 0.0002Plasma osmolality (mosmol/kgH2O) 377 � 9 (12) 351 � 6 (13) 0.02Plasma sodium (mM) 151.0 � 0.86 (7) 147.0 � 1.17 (7) 0.01Urine AVP (pg/24 h) 1022 � 83 (15) 619 � 52 (15) 0.0003Urine AVP (pg/ml) 930 � 121 (15) 1567 � 154 (15) 0.003Renal cAMP (pmol/mg protein) 3.8 � 0.2 (8) 4.1 � 0.5 (8) NSUrine cAMP (nmol/24 h) 11.6 � 2.2 (5) 12.8 � 2.9 (5) NSPlasma NOx (�M) 96 � 12 (9) 90 � 12 (9) NSUrine NOx (�M/24 h) 4.2 � 0.4 (9) 1.8 � 0.2 (9) 0.0004Urine PGE2 (�g/24 h) 7.1 � 1.1 (13) 4.6 � 0.8 (13) 0.07Urine calcium (�mol/24 h) 32 � 5 (15) 19 � 2 (15) 0.02Urine calcium (�M) 30 � 5 (15) 48 � 5 (15) 0.01

aData are means � SEM. Numbers in parentheses refer to number of mice. All urinary data are expressed in cumulativevalues per 24 h, obtained in metabolic cages. The urinary concentrations of AVP and calcium are also indicated since theymay interact with luminal receptors. AVP, Arginine vasopressin; NOx, nitric oxide metabolites; PGE2, prostaglandin E2.


the binding parameters (Kd and Bmax) were similar in bothgroups. We next tested the aquaretic response of Pkd1 mice tothe V2R antagonist SR121463B. Using incremental dosages, weshowed a dosage-dependent increase in the aquaretic effect inWT mice, whereas Pkd1�/� mice showed a decreased sensitiv-ity to the V2R antagonist (lower diuresis during the 6-h period)at dosages that ranged from 0.1 to 10 mg/kg (Figure 2B). Thedecreased response to SR121463B was confirmed by a rightshift of the dosage-response curve (Figure 2C), with Pkd1�/�

mice showing a significantly higher ED50 in comparison withPkd1�/� mice (Log ED50: 0.723 � 0.03 in Pkd1�/� versus 0.507 �

0.07 in Pkd1�/�; n � 5 pairs; P � 0.02). These data demonstratethat, despite similar distribution and binding parameters ofV2R, the Pkd1�/� mice have a decreased sensitivity to a V2Rantagonist.

Mechanism of Antidiuresis in Pkd1�/� Mice: Real-TimeRT-PCR Analyses

The potential mechanism for the inappropriate antidiuresisin the Pkd1�/� mice was further investigated. We used real-time RT-PCR to test for the differential expression of transcriptsthat primarily are involved in the AVP signaling pathway,including the V1a and V2 receptors, the calcium-sensing recep-tor (CaR), endothelin-1 (ET1), AQP2, the cAMP-responsive el-ement binding protein (CREB, a mediator in gene transcriptionof AQP2) in the kidney, and AVP in the brain. In addition,mRNA levels of other cAMP-dependent molecules, such as

AQP3, the epithelial sodium channel (ENaC) �-subunit, and theurea transporter 1 (UTA1) were measured (Figure 3). The ex-pression levels of these mediators were similar in both groups,except for a slight but significant decrease (average �21%) inthe expression of endothelin 1 (ET1). The expression levels oftranscripts related to intracellular Ca2�, such as adenylate cy-clase isoforms 3 and 6 (AC3, AC6), calmodulin (CaM), parval-bumin, and calcineurin A� or � (PPP3CA, PP3CB) were similar.The mRNA levels of endothelial nitric oxide synthase (eNOS)and neuronal nitric oxide synthase (nNOS), two mediators incAMP-independent cell-surface expression of AQP2, were alsosimilar in both groups. There was no upregulation of Pkd2expression in kidney and brain. Importantly, the brain AVPexpression was unchanged in the Pkd1�/� mice (average 108%)despite chronic hypo-osmolality.

Mechanism of Antidiuresis: Phosphorylation andRecruitment of AQP2

We next investigated whether the antidiuresis that was ob-served in the heterozygous Pkd1 mice could be related to mod-ifications in the AQP2 trafficking in the CD (Figure 4). Immu-noblotting showed a significant upregulation of AQP2 andp-AQP2 in membrane fractions from Pkd1�/� kidneys, contrast-ing with stable AQP1 levels (Figure 4A). Densitometry analysisconfirmed the significant increase in AQP2 (approximately 0.6-fold) and p-AQP2 (approximately 1.15-fold) in the membranefractions that were obtained from Pkd1�/� kidneys (Figure 4B).

Figure 1. Response to water deprivation and water loading in Pkd1 mice. (A) Urine output was measured during 24-h baseline (BL)and 24-h water deprivation (WD) in four pairs of Pkd1�/� and Pkd1�/� mice. The Pkd1�/� mice had a significantly lower urineoutput at BL and after 24-h WD. The WD resulted in a similar weight loss in both groups. (B) Urine osmolality (Uosm) was higherat BL and after WD in the Pkd1�/� group versus Pkd1�/� group. *P � 0.05, Pkd1�/� versus Pkd1�/�. (C) Nine pairs of mice wereadministered an intraperitoneal injection of 2 ml of sterile water. In comparison with wild-type (WT) littermates, Pkd1�/� miceshowed a significantly delayed ability to excrete water up to 3 h after water load. *P � 0.035, #P � 0.005, Pkd1�/� versus Pkd1�/�.


The recruitment of AQP2 to the apical membrane of CD cellswas also reflected by its increased excretion in urine (Figure4C).

In strictly controlled conditions of incubation (Figure 5), thestaining for both AQP2 and p-AQP2 was upregulated in theapical membrane of the principal CD cells in the medulla ofkidneys from Pkd1�/� mice (Figure 5A). In contrast, the AQP3labeling was restricted to the basolateral plasma membrane,with no difference in staining intensity or distribution. Immu-nogold staining at the EM level (Figure 5B) showed a significantlabeling for AQP2 at the apical plasma membrane in most CDprincipal cells in Pkd1�/� mice (Figure 5B, top) but an evenmore abundant apical membrane labeling in the cells ofPkd1�/� mice (Figure 5B, bottom). Of interest, the principal cellsin the CD of Pkd1�/� mice often exhibited extensive infoldings(small microvilli or more probably microplicae as a result ofapical plasma membrane infoldings), which were not observedas often in the Pkd1�/� cells (Figure 5B, top versus bottom).Morphometry analysis of the gold particles that localized at the

plasma membrane demonstrated a two-fold increase in thedensity of AQP2 labeling in the Pkd1�/� mice (Figure 5C). Ofnote, there were no modifications in the structure or diameterof the vasa recta in the Pkd1�/� mice (data not shown).

In view of the role of Rho signaling in regulating the cy-toskeletal dynamics and AQP2 translocation in CD cells, weinvestigated the expression and phosphorylation of RhoA (Fig-ure 6A) and activity level of RhoA (Figure 6B) in the Pkd1kidneys. Western blotting demonstrated a significant upregu-lation of p-RhoA, with downregulation of RhoA in kidneyextracts from Pkd1�/� mice. Furthermore, affinity precipitationof GTP-Rho followed by Western blotting confirmed that theamount of active RhoA was significantly decreased in thePkd1�/� kidney extracts. The ERK1/2 signaling, another regu-lator of Rho activity, was also downregulated as evidenced bythe significant decrease of p-ERK1/2 over total ERK1/2 ratio inhomogenates from Pkd1�/� kidneys (Figure 6C). These datademonstrate that the inappropriate water reabsorption that wasobserved in Pkd1�/� mice reflects an increased phosphorylation

Figure 2. Autoradiography and binding properties of renal V2 receptors and efficacy of the vasopressin (AVP) V2 receptor (V2R)antagonist SR121463B in Pkd1 mice. (A) Autoradiography of Pkd1�/� and Pkd1�/� kidneys that were incubated with the highlyselective V2R ligand [3H]SR121463 alone (1.5 nM, total binding; a and b) and in the presence of 1 �M unlabeled SR121463B (c andd) or AVP (e and f; nonspecific binding). The obtained autoradiograms show a dense specific labeling confined in the inner/outermedulla and papilla area, corresponding to the main localization of the V2R in rodent collecting ducts (CD). Binding studiesshowed that [3H]SR121463 binds with high affinity to renal V2R. The distribution pattern is similar in Pkd1�/� and Pkd1�/�

kidneys. The saturation binding experiments revealed similar binding parameters of [3H]SR121463 for renal V2R in Pkd1 mice (NS;i.e., apparent equilibrium dissociation constant [Kd] and maximum binding density [Bmax]). (B) Hourly urine output after injectionof various dosages (0.1 to 30 mg/kg) of SR121463B in five pairs of mice. In comparison with WT littermates (solid trait), Pkd1�/�

mice (dashed trait) showed a systematically lower urine excretion during the 6-h period for dosages that ranged from 0.1 to 10mg/kg. Each point is the mean of five mice in each group. (C) Dosage-response curves and cumulative ED50 determination. Asignificant higher ED50 is observed in Pkd1�/� mice versus WT littermates (Log ED50 mean 0.507 � 0.07 in Pkd1�/�, 0.723 � 0.03in Pkd1�/�; n � 5 pairs). The Log ED50 value corresponds to dosage concentrations of 3.2 mg/kg in Pkd1�/� and 5.3 mg/kg inPkd1�/� mice. Diuresis values (�l per 6-h period) were significantly lower in Pkd1�/� versus Pkd1�/� at dosage concentrations of1, 3, and 10 mg/kg. *P � 0.05.


of AQP2 and RhoA and decreased activity of RhoA, promotingthe recruitment of AQP2 at the apical plasma membrane of theprincipal cells.

[Ca2�]i in IMCD from Pkd1 MiceFor investigation of whether the heterozygous loss of Pkd1 is

sufficient to alter resting [Ca2�]i in the principal cells of the CD,isolated IMCD that were dissected from three pairs of age- andgender-matched Pkd1 mice were loaded with Fura-2 to measure[Ca2�]i levels. As shown in Figure 7, [Ca2�]i values were sig-nificantly lower in Pkd1�/� versus Pkd1�/� cells (146 � 3.0versus 186 � 3.5 nM, respectively; P � 0.0001).

DiscussionIn this study, we show that reduced Pkd1 gene dosage in

mouse leads to a syndrome of inappropriate antidiuresis(SIAD), in the absence of cystic changes and renal failure. Theheterozygous Pkd1�/� mice are characterized by the inappro-priate expression of AVP in brain and the recruitment of AQP2in the apical plasma membrane of the CD principal cells, re-flecting decreased [Ca2�]i levels and decreased activity ofRhoA in these cells. These data, the first to document functional

modifications in heterozygous Pkd1 mice, emphasize the im-portance of abnormal AVP and Ca2� signaling in ADPKD andgive insights in the potential roles of PKD1. Also, the Pkd1�/�

mice represent a model of inappropriate antidiuresis that maybe useful to decipher the mechanisms that are involved inAQP2 trafficking.

In contrast with the Pkd1-null mouse models, which areembryonically lethal, heterozygous Pkd1 mice have a normalgrowth and no detectable abnormalities at birth and duringadulthood (17–21). With the exception of a limited number ofrenal and liver cysts in a minority of old mice (17,21,22), nodetailed functional phenotype has been associated with Pkd1haploinsufficiency. Recent studies in cystic mouse (Pkd2�/tm1Som

and pcy) and rat (PCK) models with various degree of renalfailure pointed out that increased cAMP levels, secondary toabnormal V2R signaling in CD cells, could play a role in cystprogression (8–10). However, the effects of such an abnormalsignaling could be masked by nonspecific structural changesthat alter the osmotic water handling by the CD (16). Thus,adult Pkd1�/� mice offer the opportunity to test the functionalconsequences of a Pkd1 haploinsufficient state on AVP signal-ing and water handling in the absence of intercurrent mecha-nisms.

Several lines of evidence show that Pkd1�/� mice have aSIAD. At baseline, the Pkd1�/� mice have a decreased urinaryoutput with higher urine sodium/osmolality and lower plasmasodium/osmolality. After water deprivation, the Pkd1�/� miceare able to concentrate urine to a greater extent than WT litter-mates. Conversely, they have an impaired ability to excrete awater load. All of these elements indicate that haploinsuffi-ciency in Pkd1 is associated with abnormal osmoregulation anda positive water balance. That both the water intake and theexpression of AVP in the brain are unchanged, irrespective ofthe chronic hypo-osmolality, does suggest a central defect inPkd1�/� mice. Such a central defect could reflect high expres-sion levels of polycystins in the brain (20) and a potential rolein the pathways that regulate AVP secretion. Of note, twopotential mechanisms may explain the significantly lower val-ues of urine AVP excretion in the Pkd1�/� mice: (1) The waterretention, causing a decrease in the urinary output, and (2) thaturinary AVP excretion is influenced by the osmolar clearance(Cosm), as a result of interference with the reabsorption/deg-radation of filtered AVP in the proximal tubule (31). Accord-ingly, the lower urinary AVP excretion could reflect the signif-icant decrease in Cosm in Pkd1�/� versus WT mice (3.9 � 0.2versus 6.2 � 0.3 �l/min; n � 15 pairs; P � 0.0001), leading toaccelerated tubular degradation of AVP. Recent studies haveshown an increased reactivity of the aortic and renal vascula-ture in Pkd1�/� mice (32), and a reduced renal blood flow couldexplain such a reduced Cosm. At any rate, the positive waterbalance of noncystic Pkd1�/� mice contrasts with the mildconcentrating defect reported in patients with ADPKD (14). Theexistence of nephrogenic diabetes insipidus in conditions thatare associated with structural changes in the medulla (16) andthe correlation between the number of renal cysts and theextent of the concentrating defect (33) suggest that the latter

Figure 3. Real-time reverse transcription–PCR quantification of themRNA expression of mediators that are involved in AVP signaling inthe kidney and brain of Pkd1 mice. The mRNA levels were firstadjusted to glyceraldehyde-3-phosphate dehydrogenase (GAPDH),then normalized to the WT level set at 100% using the following for-mula: Ratio � 2 [�Ct(GAPDH � Target�/�) � �Ct(GAPDH � Target�/�).These normalized values (mean � SEM) are shown in the rightcolumn. In addition to the haploinsufficiency in Pkd1, there was asignificant decrease in the expression of endothelin-1 (ET1) inPkd1�/� kidneys. Nine pairs of kidneys and 11 pairs of brainswere analyzed.


primarily reflects cystic changes in the medulla of patients withADPKD (11).

Kidney-specific mechanisms are also involved in the inap-propriate antidiuresis phenotype of the Pkd1�/� mice (Figure8). In normal conditions, the binding of AVP to V2R at thebasolateral pole of CD principal cells triggers a heterotrimericG-protein–coupled cascade, activating AC6 and increasingcAMP levels, which leads to the phosphorylation of AQP2 atSer256 by protein kinase A (PKA), followed by the trafficking ofp-AQP2 to the apical plasma membrane and the increase in theosmotic water permeability of the cells. The cAMP-inducedtranslocation of AQP2 is facilitated by PKA-mediated phos-phorylation (Ser188) of the small GTP-binding protein RhoA,causing Rho inactivation and depolymerization of F-actin (28).In some conditions, the V2R-mediated antidiuretic actions ofAVP may be balanced by the apical V1aR and CaR in theprincipal cells (31,34,35). Binding of luminal AVP to V1aRstimulates phospholipase C (PLC), leading to inositol trisphos-phate receptor (IP3R)-mediated release of Ca2� from the endo-plasmic reticulum. In turn, increased [Ca2�]i activates phos-phodiesterase-1 (PDE1) and inhibits AC6, leading to decreasedcAMP. However, activation of CaR by high luminal calciumconcentrations leads to (1) increased [Ca2�]i via PLC and IP3Rand (2) activation of protein kinase C (PKC) and phosphoryla-tion of ERK1/2, followed by activation of phospholipase A2(PLA2), release of PGE2, and prostaglandin EP3 receptor

(EP3R)-mediated activation of RhoA, resulting in F-actin forma-tion and reduced insertion of AQP2 into the apical plasmamembrane (28,36,37). The stimulation of CaR may also activatePKC isoforms that mediate AQP2 endocytosis (34,38). Severalabnormalities in these signaling pathways could potentiallylead to the inappropriate water retention in the Pkd1�/� mice,as discussed next.

First, there is a consistent and highly significant decrease in[Ca2�]i levels in isolated CD from Pkd1�/� mice. It is increas-ingly recognized that the functional interaction between poly-cystins 1 and 2 in primary cilia plays an important role inluminal flow sensing and regulation of [Ca2�]i homeostasis inresponse to mechanosensation in tubular cells (11,39–41). Sev-eral lines of evidence suggest that disruption of the polycystinspathway leads to reduced [Ca2�]i. For instance, decreased rest-ing [Ca2�]i levels have been observed in cultured cells thatwere derived from human ADPKD cysts (42) and vascularsmooth muscle cells from heterozygous Pkd2�/� mice (43). Asignificant decrease in [Ca2�]i was also observed in vascularsmooth muscle cells from the Pkd1�/� mice that were used inthis study (32). By analogy, the lower [Ca2�]i levels in IMCDthat were isolated from Pkd1�/� mice could reflect the reducedPkd1 dosage. Alternatively, the lower [Ca2�]i may reflect adecreased activity of the apical V1aR and/or CaR signaling(31,34,35). However, the Pkd1�/� mice showed significantly

Figure 4. Expression of aquaporin 1 (AQP1), AQP2, and phosphorylated AQP2 (p-AQP2): Immunoblotting. (A) Representativeimmunoblots for AQP1, AQP2, and p-AQP2 in homogenate (H) and membrane (M) fractions that were prepared from Pkd1 mousekidneys. Equal loads (20 �g) were compared, as verified by similar �-actin expression. Although there is no difference in AQP2expression in the H fractions, there is an upregulation of AQP2 and p-AQP2 in the M fractions from Pkd1�/� kidneys. There isno difference in AQP1 expression in these very M fractions. (B) Densitometry analysis (core and glycosylated bands) confirms thatthere is a significant increase of AQP2 (relative OD 161 � 2%; P � 0.01) and p-AQP2 (relative OD 214 � 6%; P � 0.0002) in theM fractions of Pkd1�/� kidneys. (C) Representative immunoblot for AQP2 and p-AQP2 in Pkd1 mouse urine. Samples (15 �l) wereloaded and analyzed by Western blot under nonreducing condition. *P � 0.05, Pkd1�/� versus Pkd1�/�.


higher urinary concentration of calcium and AVP, with un-changed expression of both V1aR and CaR in the kidney.

Second, we documented an increased p-AQP2 and recruit-ment of AQP2 in the apical plasma membrane of CD cells andincreased p-RhoA coupled to decreased activity of RhoA in thePkd1�/� kidneys. These data suggest that inactivation of RhoA,causing the depolymerization of F-actin, facilitates the AVP-elicited insertion of AQP2 into the apical plasma membrane ofthe Pkd1�/� mice. Several factors could contribute to thesemodifications, including the decreased [Ca2�]i ; the increasedPKA-mediated phosphorylation of AQP2 and RhoA (28); andthe less active ERK-PLA2 pathway, as indicated by the lowerp-ERK1/2 over total ERK1/2 ratio and a trend for lower uri-nary PGE2 excretion (28,36). These modifications of the ERKpathway are different from observations in cultured renal cells.Yamaguchi et al. (44) showed that the cAMP-dependent prolif-eration of cultured human ADPKD cyst-lining cells is mediatedthrough phosphorylation/activation of ERK. By contrast,cAMP inhibits ERK activity and slows proliferation in normalepithelial cells from human kidney cortex. However, whenimmortalized mouse M1 cortical CD cells were treated withcalcium channel blockers or EGTA to lower [Ca2�]i, the cells

converted to a cystic-like phenotype, with cAMP-dependentactivation of ERK and proliferation. Of note, lowering [Ca2�]i

alone (a situation that is similar to that observed in Pkd1�/�

kidneys) was not sufficient to activate ERK and proliferation inthese cells (45). Many elements contribute to the differencesbetween the native, noncystic Pkd1�/� kidneys and culturedcells, including time course and magnitude of [Ca2�]i modifi-cations, residual levels of polycystin-1, and adaptation mecha-nisms, yet the lower [Ca2�]i levels that were observed in thePkd1�/� CD cells may represent an intermediate state, in whicha further loss of polycystin-1 or changes in cAMP levels maylead to a proliferative or cystic phenotype. Conversely, a low[Ca2�]i could decrease the activity of Ca2�-dependent proteinphosphatases and calcineurin-A� and �, resulting in higherlevels of phosphorylated AQP2 and a reduced recycling fromthe apical plasma membrane (46).

Third, the apical V1aR and CaR pathways seem to be inactivein Pkd1�/� mice, as evidenced by lower [Ca2�]i in isolated CD,decreased activity of the ERK-PLA2 pathway, and decreasedRhoA activity. As discussed, the apical CaR may sense urinarycalcium levels and influence AQP2 targeting to adjust waterhomeostasis. The experiments of Sands et al. (38), performed on

Figure 5. Immunostaining and electron microscopy immunogold labeling for AQP2 in the kidneys of Pkd1 mice. (A) In comparisonwith Pkd1�/�, there is a strong increase in the apical signal for AQP2 (a and b) and p-AQP2 (c and d) in the kidneys of Pkd1�/�

mice. The typical staining for AQP3 (e and f) in the basolateral membrane of the principal cells (PC) is similar in both groups.Several unlabeled CD cells correspond to intercalated cells. Bar � 50 �m. (B) Representative electron microscopy gold labeling ofAQP2 in the outer medullary CD of Pkd1�/� (a) and Pkd1�/� (b) mouse kidney. The labeling was restricted to PC and particularlyabundant at the plasma membrane. However, AQP2 labeling was remarkably more intense in Pkd1�/� than in Pkd1�/� kidneys.IC, intercalated cell; N, nucleus; J, tight junction; L, collecting duct lumen. Bar � 0.5 �m. (C) Morphometric analysis of the densityof gold particles at the apical plasma membrane of Pkd1�/� and Pkd1�/� PC confirmed an approximately two-fold increase (*P �1.10�6) in the apical AQP2 in the Pkd1�/� versus Pkd1�/� kidney (number of gold particles per millimeter of membrane length:11.44 � 0.87 versus 5.18 � 0.58). The total number of particles and total membrane length were 725 particles over 140 mm (29 cellsfrom 3 Pkd1�/� mice) and 1633 particles over 142 mm (27 cells from 3 Pkd1�/� mice).


isolated rat IMCD, showed that increasing luminal calciumfrom 1 to 5 mM causes a 30% decrease in the AVP-elicitedosmotic water permeability but no change in the basal perme-ability. Similar calcium concentrations were used to show thelink between CaR activation and AQP2 trafficking in culturedcells (34). These calcium concentrations are 100-fold higher thanthose observed in Pkd1�/� mice (50 �M, only 1.5-fold higherthan in WT mice). A significant polyuria has also been observedin hypercalciuric mouse models in vivo. For instance, the diure-sis is increased two-fold in the Trpv5-null mice, characterizedby very high calciuresis (averaging 250 �mol/24 h, six-foldhigher than in WT littermates) (47). Again, these conditions arevery distinct from the Pkd1�/� mice, which have no real hyper-calciuria (average 20 �mol/24 h). Regarding the luminal V1areceptors, their antagonistic action has been demonstrated onlyin rabbit CD, with no evidence in rat or mouse kidney (31).Therefore, the increased luminal concentrations of calcium andAVP in Pkd1�/� mice versus WT probably have limited biologicrelevance.

Fourth, the unchanged cAMP levels in kidney and urine andthe stable expression of AQP2 and V2R mRNA expression inthe kidney of Pkd1�/� mice argue against a direct role ofchronically increased cAMP levels in this model. Furthermore,

the mRNA level of targets of the V2R signaling pathway (in-cluding urea transporter 1 and the � subunit of epithelial so-dium channel) were similar in Pkd1�/� and Pkd1�/� mice.These data contrast with the elevation of cAMP, paralleled bythe upregulation of AQP2 and V2R mRNA that was observedin other PKD mouse models (8,9). However, all of these modelsshow renal cysts, with a positive correlation between cAMPlevel and the magnitude of cystic changes, suggesting a causallink between the two (10). Although we did not detect anincrease in cAMP in whole tissue, we cannot exclude that alower [Ca2�]i could favor a local increase in cAMP by a dualeffect on AC6 and phosphodiesterase E1 in the principal cells ofthe Pkd1�/� mice. Indeed, recent studies suggested that com-partmentalization of cAMP signaling in microdomains mayparticipate in the regulation of AQP2 trafficking (48).

Fifth, in addition to the aforementioned mechanisms, thetrafficking of AQP2 in CD cells can be stimulated by cGMP viaactivation of NOS (49). The Pkd1�/� mice showed no differencein the renal expression of eNOS and nNOS isoforms but asignificant decrease in the urinary excretion of NO metabolites.These data, which confirm previous reports of impaired NOsynthesis in this mouse model (17) and patients with ADPKD(50,51), suggest that the cGMP-mediated cascade is probably

Figure 6. Activity of RhoA and extracellular signal–regulated kinase (ERK) signaling in Pkd1 kidneys: Immunoblotting and affinityprecipitation. (A) Representative immunoblots for p-RhoA and RhoA in homogenates from Pkd1 mouse kidneys. Equal loads (20�g per lane) were compared, as verified by similar �-actin expression. After probing for p-RhoA, the membrane was stripped andreprobed for RhoA. The upregulation of p-RhoA in the Pkd1�/� group is reflected by the downregulation of RhoA. Densitometryanalysis confirms the significant increase of the p-RhoA over RhoA ratio in Pkd1�/� kidneys versus 100% of the Pkd1�/� group(749 � 165%; P � 0.001). (B) Representative immunoblot for quantification of active RhoA (GTP-bound) in Pkd1 mouse kidneys.Equal loads (20 �l per lane) were compared. Affinity precipitation followed by Western blotting confirmed that the amount ofactive RhoA was decreased in the Pkd1�/� kidneys. Densitometry analysis confirms that there is a significantly decreased amountof active RhoA in Pkd1�/� kidneys (relative OD 35 � 13%; P � 0.01). (C) Representative immunoblot for phosphorylated and totalERK1/2 in homogenates from Pkd1 mouse kidneys. Equal loads (20 �g per lane) were compared and verified by similar �-actinexpression. There is a significant decrease of the p-ERK1/2 over ERK1/2 ratio in Pkd1�/� kidneys, as confirmed by densitometry(46 � 7% versus Pkd1�/� taken as 100%; P � 0.02).


not involved in the recruitment of AQP2. Furthermore, thedecreased [Ca2�]i in Pkd1�/� CD could participate in the de-creased renal NOS activity, because eNOS and nNOS both areCa2� dependent. Therefore, a reduced generation of NO coulddecrease the medullary blood flow, sensitizing the Pkd1�/�

mice to the sympathetic tone and contributing to the antidiure-sis phenotype by a positive effect on medullary hypertonicity.Another factor is medullary ET1, which antagonizes the AVP-induced cAMP accumulation in CD cells and increases medul-lary blood flow in vivo (52). The ET1 action is mediated by theETB receptor, and it involves PLC, PLA2, and PKC, as well as[Ca2�]i, NO, and PGE2. Mice that lack ET1 in the CD show noabnormalities at baseline but a reduced ability to excrete urineduring acute water loading (53). Together with the reduced[Ca2�]i and decreased NO production, the mild but significantdecrease in the mRNA expression of ET1 evidenced in Pkd1�/�

kidneys could thus contribute both to increased AVP signalingand to decreased medullary blood flow, leading to water re-tention.

The Pkd1�/� mice show a significant resistance to V2R an-tagonism, despite similar distribution and affinity of the V2Rand unchanged mRNA expression of AVP. This observationmay be relevant for the use of V2R antagonists in patients withADPKD, with the aim to interfere with the cystogenic effect ofcAMP (11). In the sole orthologous model of ADPKD that hasbeen investigated thus far (Pkd2�/tm1Som mouse), treatment withthe V2R antagonist OPC31260 has not been associated withsignificant changes in urine output and osmolality, and a com-parison of its efficacy in WT and Pkd2 mice has not beenreported (9). Of interest, it has been shown recently that a

Figure 7. Baseline intracellular Ca2� concentrations ([Ca2�]i) ininner medullary CD (IMCD) tubules of Pkd1 mice. Differentregions (n � 24 per group) from six Pkd1�/� (filled symbols)and six Pkd1�/� (open symbols) IMCD tubules that originatedfrom three pairs of mice were analyzed. [Ca2�]i were lower inthe Pkd1�/� group (open symbols) compared with WT group(filled symbols; 146 � 3.0 versus 186 � 3.5 nM, respectively; P �0.0001).

Figure 8. Model for the effects of PKD1 haploinsufficiency onAVP signaling and AQP2 trafficking in the PC of CD. Repre-sentation of a typical PC of the CD showing the influence of[Ca2�]i levels and various signaling pathways on the traffickingof AQP2 and the modifications in Pkd1�/� mice. In the normalstate, the binding of AVP to the basolateral V2R activatesadenylyl cyclase 6 (AC6), resulting in cAMP-dependent activa-tion of protein kinase A (PKA) and phosphorylation of AQP2(Ser256) and its insertion into the apical membrane. The processis facilitated by PKA-mediated phosphorylation of RhoA(Ser188), which inactivates RhoA and causes the depolymeriza-tion of F-actin. The V2R-mediated effects of AVP could bebalanced by the apical vasopressin-1a (V1aR) and calcium-sensing (CaR) receptors. Luminal AVP activates V1aR, whichincreases [Ca2�]i via phospholipase C (PLC) and inositoltrisphosphate receptor (IP3R). In turn, the increased Ca2� acti-vates phosphodiesterase-1 (PDE1) and inhibits AC6, leading todecreased cAMP. High extracellular, urinary Ca2� levels canactivate CaR, leading to (1) increased [Ca2�]i via PLC and IP3Rand (2) activation of PKC and dual phosphorylation of ERK1/2via PLC and diacylglycerol (DAG), leading to activation of phos-pholipase A2 (PLA2) and release of PGE2, activation of prostaglan-din EP3 receptor, and downstream activation of RhoA with sub-sequent F-actin formation, which reduces the insertion of AQP2into the apical plasma membrane. The polycystin-1 and -2interact in the primary cilium located in the apical plasmamembrane of CD cells. In response to luminal flow, the poly-cystin-1/2 complex regulates [Ca2�]i by mediating a Ca2� entryinto the cell, which releases Ca2� stores via the ryanodinereceptors (RyR) on the endoplasmic reticulum (ER). The hap-loinsufficient Pkd1 state is characterized by the increased re-cruitment of AQP2 into the apical plasma membrane, reflectingdecreased [Ca2�]i, decreased ERK-PLA2 activity, increasedPKA-mediated phosphorylation of AQP2 and RhoA, and de-creased activity of RhoA. The decreased [Ca2�]i may also resultin increased efficiency of the cAMP-mediated signaling in mi-crodomains of the cell. These events increase the efficiency ofV2R-mediated signaling, leading to the recruitment of AQP2 inthe apical plasma membrane and inappropriate reabsorption ofwater by the PC. The large arrows indicate the changes thatwere documented in Pkd1�/� mice. The increased urinary con-centrations of calcium and AVP are between parentheses be-cause there is no evidence of stimulated apical receptors.Adapted from references (31,34–36).


partial loss of Pkd2 attenuated the polyuria and increased theurine osmolality in a mutant V2R mouse model of nephrogenicdiabetes insipidus (54), suggesting that an alteration of thepolycystin pathway may indeed cause water retention by CDprincipal cells. Another interesting observation is that hypona-tremia is frequently observed in neonates with autosomal re-cessive polycystic kidney disease (55).

Finally, it should be pointed out that the Pkd1�/� mice thatwere studied here represent a potentially interesting model toinvestigate water handling in the CD. A new disease entity, thenephrogenic SIAD, has recently been attributed to activating(gain-of-function) missense mutations in AVPR2, the gene thatencodes V2R in humans (56). To the best of our knowledge,there are no genetically modified mice with such an antidiure-sis phenotype at baseline. An impaired ability to lower urineosmolality and increase urinary water excretion was recentlyreported in mice that lack the taurine transporter gene Taut (57),which could play a role in primary cilia (58). Thus, in additionto the interactions between polycystins and calcium signalingpathways, the Pkd1�/� mice could give insights into the mech-anisms that govern osmoregulation, AVP signaling, and traf-ficking of AQP2 in the CD.

AcknowledgmentsThese studies were supported in part by the Fonds National de la

Recherche Scientifique, the Fonds de la Recherche Scientifique Medi-cale, an Action de Recherche Concertee (ARC 05/10-328), an IAP VI,and the EuReGene integrated project (FP6).

Some of these data were presented during the 39th annual meeting ofthe American Society of Nephrology; November 16 through 19, 2006;San Diego, CA; and published in abstract form (J Am Soc Nephrol 17:513A, 2006).

We are grateful to V. Beaujean, Y. Cnops, H. Debaix, F. Jouret, K.Parreira, and L. Wenderickx for excellent assistance and Profs. L.Bankir, D. Bichet, L. Guay-Woodford, J.-C. Henquin, N. Morel, Y.Pirson, A. Robert, R. Sandford, P. Steels, V. Torres, E. Van Kerkhove,and Dr. I. Smets for helpful discussions. We thank Sanofi-Aventis forproviding SR121463B.

DisclosuresNone.

References1. Pirson Y, Chauveau D, Devuyst O: Autosomal dominant

polycystic kidney disease. In: Oxford Textbook of ClinicalNephrology, 3rd Ed., edited by Davison AM, Cameron JS,Grunfeld JP, Ponticelli C, Ritz E, Winearls G, Ypersele CV,Oxford, Oxford University Press, 2005, pp 2304–2324

2. Hateboer N, Dijk MA, Bogdanova N, Coto E, Saggar-MalikAK, San Millan JL, Torra R, Breuning M, Ravine D: Com-parison of phenotypes of polycystic kidney disease types 1and 2. Lancet 353: 103–107, 1999

3. Igarashi P, Somlo S: Genetics and pathogenesis of polycys-tic kidney disease. J Am Soc Nephrol 13: 2384–2398, 2002

4. Verani RR, Silva FG: Histogenesis of the renal cysts inadult (autosomal dominant) polycystic kidney disease: Ahistochemical study. Mod Pathol 1: 457–463, 1988

5. Devuyst O, Burrow CR, Smith BL, Agre P, Knepper MA,

Wilson PD: Expression of aquaporin-1 and -2 duringnephrogenesis and in autosomal dominant polycystic kid-ney disease. Am J Physiol 271: F169–F183, 1996

6. Mangoo-Karim R, Uchic M, Grant M, Shumate WA, CalvetJP, Park CH, Grantham JJ: Renal epithelial fluid secretionand cyst growth: The role of cyclic AMP. FASEB J 3:2629–2632, 1989

7. Belibi FA, Reif G, Wallace DP, Yamaguchi T, Olsen L, Li H,Helmkamp GM Jr, Grantham JJ: Cyclic AMP promotesgrowth and secretion in human polycystic kidney epithe-lial cells. Kidney Int 66: 964–973, 2004

8. Gattone VH, Xiaofang W, Harris PC, Torres VE: Inhibitionof renal cystic disease development and progression by avasopressin V2 receptor antagonist. Nat Med 9: 1323–1326,2003

9. Torres VE, Xiaofang W, Qian Q, Somlo S, Harris PC, Gat-tone VH: Effective treatment of an orthologous model ofautosomal dominant polycystic kidney disease. Nat Med10: 363–364, 2004

10. Wang X, Gattone V 2nd, Harris PC, Torres VE: Effective-ness of vasopressin V2 receptor antagonists OPC-31260and OPC-41061 on polycystic kidney disease developmentin the PCK rat. J Am Soc Nephrol 16: 846–851, 2005

11. Torres VE: Vasopressin antagonists in polycystic kidneydisease. Kidney Int 68: 2405–2418, 2005

12. Christensen BM, Zelenina M, Aperia A, Nielsen S: Local-ization and regulation of PKA-phosphorylated AQP2 inresponse to V2-receptor agonist/antagonist treatment.Am J Physiol Renal Physiol 278: F29–F42, 2000

13. Nielsen S, Frokiaer J, Marples D, Kwon TH, Agre P, Knep-per MA: Aquaporins in the kidney: From molecules tomedicine. Physiol Rev 82: 205–244, 2002

14. Gabow PA, Kaehny WD, Johnson AM, Duley IT, Manco-Johnson M, Lezotte DC, Schrier RW: The clinical utility ofrenal concentrating capacity in polycystic kidney disease.Kidney Int 35: 675–680, 1989

15. Danielsen H, Pedersen EB, Nielsen AH, Herlevsen P, Kor-nerup HJ, Posborg V: Expansion of extracellular volume inearly polycystic kidney disease. Acta Med Scand 219: 399–405, 1986

16. Bichet D: Polyuria and diabetes insipidus. In: The Kidney:Physiology and Pathophysiology, edited by Seldin DW, Giebi-sch G, Philadelphia, Lippincott Williams & Wilkins, 2000,pp 1261–1285

17. Muto S, Aiba A, Saito Y, Nakao K, Nakamura K, Tomita K,Kitamura T, Kurabayashi M, Nagai R, Higashihara E, Har-ris PC, Katsuki M, Horie S: Pioglitazone improves thephenotype and molecular defects of a targeted Pkd1 mu-tant. Hum Mol Genet 11: 1731–1742, 2002

18. Lu W, Peissel B, Babakhanlou H, Pavlova A, Geng L, FanX, Larson C, Brent G, Zhou J: Perinatal lethality withkidney and pancreas defects in mice with a targetted Pkd1mutation. Nat Genet 17: 179–181, 1997

19. Kim K, Drummond K, Ibraghimov-Beskrovnaya O,Klinger K, Arnaout MA: Polycystin 1 is required for thestructural integrity of blood vessels. Proc Natl Acad SciU S A 97: 1731–1736, 2000

20. Boulter C, Mulroy S, Webb S, Fleming S, Brindle K, Sand-ford R: Cardiovascular, skeletal, and renal defects in micewith a targeted disruption of the Pkd1gene. Proc Natl AcadSci U S A 98: 12174–12179, 2001

21. Lu W, Fan X, Basora N, Babakhanlou H, Law T, Rifai N,


Harris PC, Perez-Atayde AR, Rennke HG, Zhou J: Lateonset of renal and hepatic cysts in Pkd1-targeted heterozy-gotes. Nat Genet 21: 160–161, 1999

22. Piontek KB, Huso DL, Grinberg A, Liu L, Bedja D, ZhaoH, Gabrielson K, Qian F, Mei C, Westphal H, GerminoGG: A functional floxed allele of Pkd1 that can be con-ditionally inactivated in vivo. J Am Soc Nephrol 15: 3035–3043, 2004

23. Serradeil-Le Gal C, Lacour C, Valette G, Garcia G, FoulonL, Galindo G, Bankir L, Pouzet B, Guillon G, Barberis C,Chicot D, Jard S, Vilain P, Garcia C, Marty E, Raufaste D,Brossard G, Nisato D, Maffrand JP, Le Fur G: Character-ization of SR 121463A, a highly potent and selective, orallyactive vasopressin V2 receptor antagonist. J Clin Invest 98:2729–2738, 1996

24. Serradeil-Le Gal C, Raufaste D, Double-Cazanave E, Guil-lon G, Garcia C, Pascal M, Maffrand JP: Binding propertiesof a selective tritiated vasopressin V2 receptor antagonist,[3H]-SR121463. Kidney Int 58: 1613–1622, 2000

25. Munson PV, Rodbard D: Ligand: A versatile computerizedapproach for characterization of ligand-binding systems.Anal Biochem 107: 220–239, 1980

26. Morita M, Kita Y, Morikawa N, Iwami M, Notsu Y: Ex-pression of arginine vasopressin and vasopressin V1a re-ceptor mRNA in diabetic (db/db) mice. Exp Clin EndocrinolDiabetes 109: 261–266, 2001

27. Jouret F, Auzanneau C, Debaix H, Wada GH, Pretto C,Marbaix E, Karet FE, Courtoy PJ, Devuyst O: Ubiquitousand kidney-specific subunits of the vacuolar H�-ATPaseare differentially expressed during nephrogenesis. J AmSoc Nephrol 16: 3235–3246, 2005

28. Tamma G, Klussmann E, Procino G, Svelto M, RosenthalW, Valenti G: cAMP induced AQP2 translocation is asso-ciated with RhoA inhibition through RhoA phosphoryla-tion and interaction with RhoGDI. J Cell Sci 116: 1519–1525,2003

29. Longhurst DM, Watanabe M, Rothstein JD, Jackson M:Interaction of PDZRhoGEF with microtubule-associatedprotein 1 light chains. J Biol Chem 281: 12030–12040, 2006

30. Smets I, Caplanusi A, Despa S, Molnar Z, Radu M, vande-Ven M, Ameloot M, Steels P: Ca2� uptake in mitochondriaoccurs via the reverse action of the Na�/Ca2� exchangerin metabolically inhibited MDCK cells. Am J Physiol RenalPhysiol 286: F784–F794, 2004

31. Bankir L: Antidiuretic action of vasopressin: Quantitativeaspects and interaction between V1a and V2 receptor-me-diated effects. Cardiovasc Res 51: 372–390, 2001

32. Devuyst O, Ahrabi AK, Caron N, Desjardins F, BalligandJL, Horie S, Morel N: Pkd1 haploinsufficiency is associatedwith altered vascular reactivity and abnormal calcium sig-naling in the mouse aorta [Abstract]. J Am Soc Nephrol 17:704A, 2006

33. Seeman T, Dusek J, Vondrak K, Blahova K, Simkova E,Kreisinger J, Dvorak P, Kyncl M, Hribal Z, Janda J: Renalconcentrating defect capacity is linked to blood pressure inchildren with autosomal dominant polycystic kidney dis-ease. Physiol Res 53: 629–634, 2004

34. Procino G, Carmosino M, Tamma G, Gouraud S, Laera A,Riccardi D, Svelto M, Valenti G: Extracellular calcium an-tagonizes forskolin-induced aquaporin 2 trafficking in col-lecting duct cells. Kidney Int 66: 2245–2255, 2004

35. Torres VE, Harris PC: Mechanisms of disease: Autosomal

dominant and recessive polycystic kidney diseases. NatClin Pract Nephrol 2: 40–55, 2006

36. Kifor O, MacLeod RJ, Diaz R, Bai M, Yamaguchi T, Yao T,Kifor I, Brown EM: Regulation of MAP kinase by calcium-sensing receptor in bovine parathyroid and CaR-trans-fected HEK293 cells. Am J Physiol Renal Physiol 280: F291–F302, 2001

37. Davies SL, Gibbons CE, Vizard T, Ward DT: Ca2�-sensingreceptor induces Rho kinase-mediated actin stress fiberassembly and altered cell morphology, but not in responseto aromatic amino acids. Am J Physiol Cell Physiol 290:C1543–C1551, 2006

38. Sands JM, Naruse M, Baum M, Jo I, Hebert SC, Brown EM,Harris HW: Apical extracellular calcium/polyvalent cat-ion-sensing receptor regulates vasopressin-elicited waterpermeability in rat kidney inner medullary collecting duct.J Clin Invest 99: 1399–1405, 1997

39. Delmas P, Padilla F, Osorio N, Coste B, Raoux M, Crest M:Polycystins, calcium signaling, and human diseases. Bio-chem Biophys Res Commun 322: 1374–1383, 2004

40. Hanaoka K, Qian F, Boletta A, Bhunia AK, Piontek K,Tsiokas L, Sukhatme VP, Guggino WB, Germino GG: Coas-sembly of polycystin-1 and -2 produces unique cation-permeable currents. Nature 408: 990–994, 2000

41. Koulen P, Cai Y, Geng L, Maeda Y, Nishimura S, WitzgallR, Ehrlich BE, Somlo S: Polycystin-2 is an intracellularcalcium release channel. Nat Cell Biol 4: 191–197, 2002

42. Yamaguchi T, Hempson SJ, Reif GA, Hedge AM, WallaceDP: Calcium restores a normal proliferation phenotype inhuman polycystic kidney disease epithelial cells. J Am SocNephrol 17: 178–187, 2006

43. Qian Q, Hunter LW, Li M, Marin-Padilla M, Prakash YS,Somlo S, Harris PC, Torres VE, Sieck GC: Pkd2 haplo-insufficiency alters intracellular calcium regulation invascular smooth muscle cells. Hum Mol Genet 12: 1875–1880, 2003

44. Yamaguchi T, Pelling JC, Ramaswamy NT, Eppler JW,Wallace DP, Nagao S, Rome LA, Sullivan LP, Grantham JJ:cAMP stimulates the in vitro proliferation of renal cystepithelial cells by activating the extracellular signal-regu-lated kinase pathway. Kidney Int 57: 1460–1471, 2000

45. Yamaguchi T, Wallace DP, Magenheimer BS, Hempson SJ,Grantham JJ, Calvet JP: Calcium restriction allows cAMPactivation of the B-Raf/ERK pathway, switching cells to acAMP-dependent growth-stimulated phenotype. J BiolChem 279: 40419–40430, 2004

46. Jo I, Ward DT, Baum MA, Scott JD, Coghlan VM, Ham-mond TG, Harris HW: AQP2 is a substrate for endoge-nous PP2B activity within an inner medullary AKAP-signaling complex. Am J Physiol Renal Physiol 281: F958 –F965, 2001

47. Hoenderop JG, van Leeuwen JP, van der Eerden BC,Kersten FF, van der Kemp AW, Merillat AM, WaarsingJH, Rossier BC, Vallon V, Hummler E, Bindels RJ: RenalCa2� wasting, hyperabsorption, and reduced bonethickness in mice lacking TRPV5. J Clin Invest 112: 1906 –1914, 2003

48. Henn V, Stefan E, Baillie GS, Houslay MD, Rosenthal W,Klussmann E: Compartmentalized cAMP signalling regu-lates vasopressin-mediated water reabsorption by control-ling aquaporin-2. Biochem Soc Trans 33: 1316–1318, 2005

49. Bouley R, Breton S, Sun T, McLaughlin M, Nsumu NN, Lin


HY, Ausiello DA, Brown D: Nitric oxide and atrial natri-uretic factor stimulate cGMP-dependent membrane inser-tion of aquaporin 2 in renal epithelial cells. J Clin Invest 106:1115–1126, 2000

50. Wang D, Iversen J, Strandgaard S: Endothelium-dependentrelaxation of small resistance vessels is impaired in pa-tients with autosomal dominant polycystic kidney disease.J Am Soc Nephrol 11: 1371–1376, 2000

51. Persu A, Stoenoiu MS, Messiaen T, Davila S, Robino C,El-Khattabi O, Mourad M, Horie S, Feron O, Balligand JL,Wattiez R, Pirson Y, Chauveau D, Lens XM, Devuyst O:Modifier effect of ENOS in autosomal dominant polycystickidney disease. Hum Mol Genet 11: 229–241, 2002

52. Kohan DE: The renal medullary endothelin system in con-trol of sodium and water excretion and systemic bloodpressure. Curr Opin Nephrol Hypertens 15: 34–40, 2006

53. Ge Y, Ahn D, Stricklett PK, Hughes AK, Yanagisawa M,Verbalis JG, Kohan DE: Collecting duct-specific knockoutof endothelin-1 alters vasopressin regulation of urine os-molality. Am J Physiol Renal Physiol 288: F912–F920, 2005

54. Torres VE, Wang X, Somlo S, Harris PC, Wess J: Polycystickidney disease and nephrogenic diabetes insipidus in dou-ble Pkd2 and V2 vasopressin receptor mutants [Abstract].J Am Soc Nephrol 16: 140A, 2005

55. Kaplan BS, Fay J, Shah V, Dillon MJ, Barratt TM: Autoso-mal recessive polycystic kidney disease. Pediatr Nephrol 3:43–49, 1989

56. Feldman BJ, Rosenthal SM, Vargas GA, Fenwick RG,Huang EA, Matsuda-Abedini M, Lustig RH, Mathias RS,Portale AA, Miller WL, Gitelman SE: Nephrogenic syn-drome of inappropriate antidiuresis. N Engl J Med 352:1884–1890, 2005

57. Huang DY, Boini KM, Lang PA, Grahammer F, DuszenkoM, Heller-Stilb B, Warskulat U, Haussinger D, Lang F,Vallon V: Impaired ability to increase water excretion inmice lacking the taurine transporter gene TAUT. PflugersArch 45: 668–677, 2006

58. Christensen ST, Voss JW, Teilmann SC, Lambert IH: Highexpression of the taurine transporter TauT in primary ciliaof NIH3T3 fibroblasts. Cell Biol Int 29: 347–351, 2005

Supplemental information for this article is available online at http://www.jasn.org/.


pkd1 haploinsufficiency causes a syndrome of inappropriate

Documents