Clinical consequences of PKHD1 mutations in 164 patients with autosomal-recessive polycystic kidney disease (ARPKD)

Kidney International, Vol. 67 (2005), pp. 829–848

Clinical consequences of PKHD1 mutations in 164 patients withautosomal-recessive polycystic kidney disease (ARPKD)

CARSTEN BERGMANN, JAN SENDEREK, ELLEN WINDELEN, FABIAN KUPPER, IRIS MIDDELDORF,FRANK SCHNEIDER, CHRISTIAN DORNIA, SABINE RUDNIK-SCHONEBORN, MARTIN KONRAD,CLAUS P. SCHMITT, TOMAS SEEMAN, THOMAS J. NEUHAUS, UDO VESTER, JUTTA KIRFEL,REINHARD BUTTNER, KLAUS ZERRES, and MEMBERS OF THE APN (ARBEITSGEMEINSCHAFT FUR

PADIATRISCHE NEPHROLOGIE)1

Department of Human Genetics, Aachen University, Aachen, Germany; Department of Pediatrics, University of Berne-Inselspital,Berne, Switzerland; Division of Pediatric Nephrology, University Children’s Hospital Heidelberg, Heidelberg, Germany; 1stDepartment of Pediatrics, University Hospital Motol Prague, Prague, Czech Republic; Division of Pediatric Nephrology, UniversityChildren’s Hospital, Zurich, Switzerland; Division of Pediatric Nephrology, University Children’s Hospital Essen, Essen, Germany;and Department of Pathology, University of Bonn, Bonn, Germany

Clinical consequences of PKHD1 mutations in 164 patients withautosomal-recessive polycystic kidney disease (ARPKD).

Background. ARPKD is associated with mutations in thePKHD1 gene on chromosome 6p12. Most cases manifest peri-

1Members of the APN are (listed alphabetically): E. Abel (Greif-swald, Germany), S. Ala-Mello (Helsinki, Finland), B. Ausserer (Dorn-birn, Austria), M. Bald (Stuttgart, Germany), R. Beetz (Mainz, Ger-many), N. Besbas (Ankara, Turkey), M. Brandis (Freiburg, Germany),M. Coulthard (Newcastle-upon-Tyne, United Kingdom), J. Dippel(Frankfurt, Germany), C. Druck Garcia (Porto Alegre, Brazil), M. Fis-chbach (Strasbourg, France), N. Foged (Odense, Denmark), Y. Frish-berg (Jerusalem, Israel), J. Gellermann (Berlin, Germany), N. Gordjani(Wurzburg, Germany), K. Haffner (Freiburg, Germany), R.C. Hen-nekam (Amsterdam, The Netherlands), B. Hoppe (Koln, Germany),P. Hoyer (Essen, Germany), U. John (Jena, Germany), H. Kaariainen(Turku, Finland), M.J. Kemper (Hamburg, Germany), P. Koivisto (Tam-pere, Finland), G. Kruger (Rostock, Germany), E. Kuwertz-Brocking(Munster, Germany), D. Lambert (Montreal, Canada), T. Lennert(Berlin, Germany), S. Li Volti (Catania, Italy), C. Mache (Graz, Aus-tria), G. Matthijs (Leuven, Belgium), O. Mehls (Heidelberg, Ger-many), V. Meiner (Jerusalem, Israel), J. Misselwitz (Jena, Germany),T. Mononen (Kuopio, Finland), D.E. Muller-Wiefel (Hamburg, Ger-many), A. Mustonen (Tampere, Finland), S. Ozen (Ankara, Turkey), J.P.Oliveira (Porto, Portugal), Y. Pirson (Brussels, Belgium), U. Querfeld(Berlin, Germany), W. Rascher (Erlangen, Germany), C. Rudin (Basel,Switzerland), H.G. Santos (Lisboa, Portugal), M. Schroder (Frankfurt,Germany), H.W. Seyberth (Marburg, Germany), S. Shalev (Afula, Is-rael), M. Shohat (Petah Tiqva, Israel), J. Strehlau (Hannover, Ger-many), O. Vierimaa (Oulu, Finland), S. Volpel (Krefeld, Germany), M.Wilson (Sydney, Australia), and B. Zimmerhackl (Innsbruck, Austria).

Key words: autosomal-recessive polycystic kidney disease(ARPKD), polycystic kidney and hepatic disease 1 (PKHD1), fi-brocystin/polyductin, long-term clinical outcome, congenital hepaticfibrosis (CHF), genotype-phenotype correlations.

Received for publication June 23, 2004and in revised form August 13, 2004, and September 13, 2004Accepted for publication September 29, 2004

C© 2005 by the International Society of Nephrology

/neonatally with a high mortality rate in the first month of lifewhile the clinical spectrum of surviving patients is much morevariable than generally perceived.

Methods. We examined the clinical course of 164 neonatalsurvivors (126 unrelated families) over a mean observation pe-riod of 6 years (range 0 to 35 years). PKHD1 mutation screeningwas done by denaturing high-performance liquid chromatogra-phy (DHPLC) for the 66 exons encoding the 4074 aa fibro-cystin/polyductin protein.

Results and Conclusion. This is the first study that reports thelong-term outcome of ARPKD patients with defined PKHD1mutations. The 1- and 10-year survival rates were 85% and 82%,respectively. Chronic renal failure was first detected at a meanage of 4 years. Actuarial renal survival rates [end point definedas start of dialysis/renal transplantation (RTX) or by death dueto end-stage renal disease (ESRD)] were 86% at 5 years, 71%at 10 years, and 42% at 20 years. All but six patients (92%) hada kidney length above or on the 97th centile for age. About75% of the study population developed systemic hypertension.Sequelae of congenital hepatic fibrosis and portal hypertensiondeveloped in 44% of patients and were related with age. Posi-tive correlations could further be demonstrated between renaland hepatobiliary-related morbidity suggesting uniform diseaseprogression rather than organ-specific patterns. PKHD1 mu-tation analysis revealed 193 mutations (70 novel ones; 77%nonconservative missense mutations). No patient carried twotruncating mutations corroborating that one missense muta-tion is indispensable for survival of newborns. We attempted toset up genotype-phenotype correlations and to categorize mis-sense mutations. In 96% of families we identified at least onemutated PKHD1 allele (overall detection rate 76.6%) indicat-ing that PKHD1 mutation screening is a powerful diagnostictool in patients suspected with ARPKD.

Autosomal-recessive polycystic kidney disease(ARPKD) (MIM 263200) occurs with a proposedincidence of 1:20,000 to 1:40,000 [1–3]. Principal his-tologic manifestations involve the fusiform dilation

829

830 Bergmann et al: Clinical consequences of PKHD1 mutations in 164 ARPKD patients

of renal collecting ducts and distal tubuli as well asdysgenesis of the hepatic portal triad known as ductalplate malformation (hyperplastic biliary ducts andcongenital hepatic fibrosis). The majority of cases areidentified late during pregnancy or in the neonatalperiod. However, fetal sonography does not allow anearly and reliable prenatal diagnosis of ARPKD aseven late midtrimester sonography often fails to detectenlargement and increased echogenicity of kidneysor oligohydramnios [4–6]. Severely affected fetusesdisplay a “Potter” phenotype with massively enlargedkidneys. According to present estimates, up to 30% to50% of all patients die shortly after birth from respira-tory insufficiency, while renal failure is rarely a causeof neonatal demise. Advances in renal replacementtherapy have improved the survival rates of ARPKDpatients and some of them reach adulthood. However,life expectancy is still significantly diminished. A widerange of ARPKD-related morbidities frequently evolve,including systemic hypertension, end-stage renal disease(ESRD), and sequelae of congenital hepatic fibrosis(CHF) [7, 8].

The recent characterization of the PKHD1 gene (MIM606702) on chromosome 6p12 provided the basis for di-rect genotyping in ARPKD [9, 10]. PKHD1 is an ex-ceptionally large gene. The longest open reading frame(ORF) transcript comprises 67 exons encoding a pro-tein of 4074 amino acids. Expression analysis suggeststhat PKHD1 and its murine ortholog undergo a com-plex pattern of alternative splicing [10, 11]. The predictedfull-length protein (fibrocystin/polyductin) represents anovel putative integral membrane protein with so far un-known function. Protein expression in kidney and liverhas been shown to be consistent with the sites affectedby the disorder. Recent studies have demonstrated fibro-cystin/polyductin, in common with many other cystopro-teins, to be localized to primary cilia and the basal body[12–15]. This adds to the growing body of data suggestinga causative role for this organelle in cystic renal diseases[16–20].

The large size of PKHD1, its supposed complicatedpattern of splicing and lack of knowledge of the encodedprotein’s function(s) pose significant challenges to DNA-based diagnostic testing. Moreover, the wide variety ofdifferent PKHD1 mutations with most changes unique tosingle families sets further requirements for investigation.In a previous study, preliminary genotype-phenotype cor-relations could be drawn for the type of mutation ratherthan for the site of the individual mutation [21]. All pa-tients carrying two truncating mutations displayed a se-vere phenotype with peri- or neonatal demise. In otherterms, survival of the first month of life obviously requiresthe presence of at least one parental PKHD1 missenseallele.

The present study aimed to (1) define the mutationalspectrum and mutation detection rate in ARPKD fami-lies with at least one neonatal survivor, (2) evaluate thelong-term clinical course in ARPKD patients with de-fined PKHD1 mutations, and (3) extend our knowledgeon clinical consequences of PKHD1 mutations.

METHODS

Selection of the study population

Fourty-two centers provided clinical information for227 patients. Of these, 186 (97 males and 89 females) metour inclusion criteria and had sufficiently documenteddata for analysis. In all but four cases, inclusion of pa-tients was based on the demonstration of typical clinicalfeatures with characteristic ultrasonographic findings [22,23] and absence of renal cysts in parental ultrasounds. Intwo cases, no parental ultrasound data were available,whereas the mother of two affected children exhibitedtwo renal cysts in her left kidney. In those patients fur-ther proof of ARPKD was requested. In one patient thiswas done by parental consanguinity with homozygoushaplotypes for chromosome 6p12 markers in the indexpatient, in the other cases by demonstration of CHF byliver biopsy. Patients were excluded if they had an af-fected parent suggesting autosomal-dominant polycystickidney disease (ADPKD) or additional malformations.In a total of 26 families, the diagnosis of ARPKD wassubstantiated by kidney (N = 7) or liver biopsy (N = 24)(in five cases combined kidney-liver biopsy). In 14 pedi-grees, necropsy was performed in an affected individual.Clinical data (up to 1995) of 29 families (46 patients) inthis series and mutations of 21 pedigrees were publishedpreviously [21, 24] (Table 1). DNA or blood samples wereobtained on the basis of informed consent from theseARPKD patients representing 126 apparently unrelatednuclear families with at least one affected child who sur-vived the neonatal period. The study population repre-sented diverse nationalities from 24, mainly European,countries. Most families have been attended by mem-bers of the German Study Group of Pediatric Nephrol-ogy [Arbeitsgemeinschaft fur Padiatrische Nephrologie,(APN)]. The mean observation period was 6.0 years (0 to35 years). Phenotypes were categorized as severe (peri-natal or neonatal demise) (N = 22) or moderate (eithersurvived complications during the first month of life orbecame first symptomatic beyond the neonatal period)(N = 164).

Elevated serum creatinine levels were defined accord-ing to Wassner [25] as mean values for age and heightgreater than the 97th centile [>2 standard deviations(SD)]. The glomerular filtration rate (GFR) was calcu-lated from serum creatinine levels and height measure-ments based on the estimated method of Schwartz et al

Bergmann et al: Clinical consequences of PKHD1 mutations in 164 ARPKD patients 831

Table 1.

Age at lastPatient Gender Origin examination/death Genotype

F2a,b F Finland Death (perinatal) Ex3: T36M (P)F 20 years Ex3: T36M (M)

225 (C) F Turkey 1 year Ex14: F372L (P)Ex14: F372L (M)

165 (C) M Turkey 7 years Ex16: I473S (P)Ex16: I473S (M)

280 (C)b F Turkey 1 year Ex16: I473S (P)F475 (C) F Lebanon Death (perinatal) Ex16: I473S (M)

M Death (perinatal) Ex32: N1744H (P)M 2 years Ex32: N1744H (M)

153 (C) M Turkey 4 years Ex34: Y1838C (P)Ex34: Y1838C (M)

244 (C) M Turkey 5 years Ex34: Y1838C (P)Ex: Y1838C (M)

162 (C) F Turkey Death (23 years) Ex34: Y1838C (P)Ex34: Y1838C (M)

160 (C)b M Turkey 2 years Ex37: I1998T (P)Ex37: I1998T (M)

F495 M Morocco 1 year Ex53: V2771A (M)Ex53: V2771A (P)

119 (C) M Turkey 11 years Ex54: I2851T (P)Ex54: I2851T (M)

242 (C) F Turkey 6 months Ex58: H3124Y (P)Ex58: H3124Y (M)

F267 (C) M India Death (perinatal) Ex58: G3178C (P)M 4 years Ex58: G3178C (M)

297 F Finland 13 years Ex16: R496X (P)Ex32: C1402F (M)

181 M Finland 5 years Ex16: R496X (M)Ex34: S1862L (P)

F113a,b F Finland 5 years Ex16: R496X (P)F 4 years Ex43: I2331K (M)

284b F Finland 17 years Ex16: R496X (P)Ex61: V3471G (M)

F512 F Finland 2 years Ex16: R496X (P)Ex61: V3471G (M)

223 F Germany 4 months Ex3: T36M (M)Ex36: L1966fs (P)

123 M Germany Death (6 years) Ex3: T36M (M)Ex61: Q3407X (P)

94 F Germany Death (5 months) Ex32: I1687TM 17 years Ex61: Q3407X

F599 M Germany 6 years Ex37: I1998T (M)F Death (3 weeks) Ex61: Q3407X (P)

198 M Germany 3 years Ex43: I2331K (P)Ex61: Q3407X (M)

F3a,b F Finland 22 years IVS13: c.977-1G>A (P)M Death (1 month) Ex43: 12331K (M)

230a M Israel Jew 11 years Ex9: I222V (P)Ex32: A1254fs (M)

255 F The Netherlands 12 years Ex22: P755L (M)Ex32: A1254fs (P)

172 M Israel Jew 12 years Ex32: A1254fs

F105b F Italy Death (perinatal) Ex54: R2840CM Death (perinatal) Ex9: I222V (M)F 8 years Ex58: R3240X (P)

F104 M Italy Death (perinatal) Ex32: C1472Y (M)M 9 years Ex58: R3240X (P)

57a M Germany 19 years Ex5: R124X (P)Ex16: G448R (M)Ex18: L542S (M)

155a,b M Germany 8 years Ex27: A1030E (P)Ex58: L3116fs (M)

161 M Germany 6 years Ex30: G1123S (P)Ex37: Q2010fs (M)


Table 1. continued.


F329 F Czech Republic 5 years Ex32: I1687T (P)F Death (8 months) Ex33: R1775X (M)

149 F France 27 years Ex23: R781XEx34: W1834C

311 M Germany 6 months Ex36: T1960A (P)IVS39: c.6333-2A>G (M)

251 F Italy 5 years Ex37: S2016C (P)Ex61: S3570X (M)

F48a M Belgium 35 years Ex36: N1919fs (P)F 27 years Ex46: I2427T (M)

F578 F Turkey 20 years Ex43: I2331fsM 12 years Ex53: R2795S

240 M Belgium 1 year Ex23: R781X (M)Ex57: 12957T (P)

212 M Germany 1 year Ex58: H3124P (P)Ex59: C3287X (M)

F59a,b M Turkey Death (1 year) Ex58: R3107X (P)M Death (perinatal) Ex59: D3293V (M)M Death (2 years)

F12a F Germany 33 years Ex3: T36M (M)F 31 years Ex9: I222V (P)

F328 F Lebanon/Germany 7 years Ex3: T36M (M)F 5 years Ex21: P676R (P)

262 M Czech Republic 14 years Ex3: T36M (M)Ex26: H916P (P)

254 F Germany 1 year Ex3: T36M (M)Ex26: L926P (P)

F17a M Germany 13 years Ex3: T36M (M)M 10 years Ex30: G1123S (P)M 7 yearsF Germany 4 years Ex3: T36M (P)

238 F 1 year Ex32: C1237Y (M)222 F Germany 6 years Ex3: T36M (P)

F 24 years Ex33: F1785L (M)F313 F Germany Death (perinatal) Ex3: T36M (M)

M 13 years Ex33: V1789L (P)186a M Germany 15 years Ex3: T36M (M)

Ex38: S2080N (P)259 F Denmark 1 year Ex3: T36M (M)

Ex43: I2331K (P)F371 F Austria 10 years Ex3: T36M (M)

M 8 years Ex46: C2422G (P)M 6 years

206a M Germany 15 years Ex3: T36M (M)Ex46: C2422G (P)

218 F Germany 1 year Ex3: T36M (M)Ex56: V2884del (P)

F127b F Australia 7 years Ex3: T36M (M)F Death (perinatal) Ex57: I2957T (P)

F549 F Germany Death (perinatal) Ex3: T36M (M)M 6 years Ex58: A3097T (P)

F242 F Death (12 years) Ex3: T36M (P)M Czech Republic Death (8 years) Ex58: R3240Q (M)F 10 yearsM 8 years

F1a,b F Finland Death (1 month) Ex3: T36M (P)M 15 years Ex61: V3471G (M)

231a F Germany 7 years Ex3:T36M (P)Ex61: L3543S (M)

183 M Germany 8 years Ex5: G112REx16: Q504HEx27: M997K

F507 F Germany 1 year Ex43: I2331K (P)M 4 months Ex55: S2861G (M)

Ex62: P3740S (M)F383 M Bulgaria Death (perinatal) Ex4: R92G (M)

M 6 years Ex39: T2140P (P)


Table 1. continued.


F151a,b M Germany 5 years Ex21: W656C (M)M Death (perinatal) Ex30: G1123S (P)

F4a M Germany 23 years Ex24: P805L/Ex58: I3177T (P)M 20 years Ex50: W2690R (M)

F538 F Australia 1 year Ex24: P805L/ Ex58: I3177T (M)Ex32: C1431S (P)

196a M Germany 9 years Ex27: I948S (M)M 8 years Ex32: l1687T (P)

F236 M Portugal 7 years Ex30: E1124K (M)M 5 years Ex58: R3240Q (P)

234 M Slovenia 1 year Ex32: L1407R (P)Ex58: S3209R (M)

93a M Germany 12 years Ex32: P1486L (M)Ex58: Y3155C (P)

219 M Germany 4 months Ex32: G1746R (P)Ex34: Y1838C (M)

252 F Czech Republic 9 years Ex36: Q1923L (P)Ex61: I3553T (M)

F55a F Germany 15 years Ex43: I2331K (P)M 12 years Ex57: I2957T (M)

45 F Switzerland 4 years Ex57: I2957T (M)Ex61: M3642T (M)

285 F Finland Death (9 months) Ex16: R496X (P)260 F Finland 18 years Ex16: R496X (M)F58 M Israel-Arab 10 years Ex36: L1966fs

F Death (2 months)175 M Germany 6 years Ex36: L1966fs122 M Germany 13 years Ex5: T128fsF262 (C) F Turkey 10 years Ex14: P356fs

F 6 yearsF 1 year

150b M Portugal 12 years Ex14: G369fs (M)258b M Israel-Arab 14 years IVS16: c.1512+1G>A263 F Switzerland 25 years IVS28: c.3228+1G>TF92 M Brasil Death (1 month) Ex58: D3230fs (P)

M 1 year142 F Germany 8 years Ex58: D3230fs (M)F306 F Turkey 12 years Ex60: G3359fs

F 2 years247 F Germany 12 years Ex61: Q3392X195 F Switzerland Death (1 year) Ex61: S3570X (M)F346 M Switzerland Death (perinatal) Ex61: Q3589X

F 12 yearsF375 M Czech Republic Death (perinatal) Ex3: T36M

M 6 yearsF32a,b M Germany Death (22 years) Ex3: T36M

F 17 yearsF27a,b F Germany 13 years Ex3: T36M

F Death (3 years)F9a,b F Germany 18 years Ex3: T36M

M 8 yearsF231a,b M Slovenia 21 years Ex3: T36M

F 4 years224 F Hungary Death (9 years) Ex3: T36M111a,b F Germany 13 years Ex3: T36M (P)117 M Germany 14 years Ex3: T36M30 F Germany 17 years Ex3: T36M202b F Czech Republic 6 years Ex3: T36M (M)286 M Germany 3 years Ex3: T36MF460 M Portugal 2 years Ex3: T36M (M)F19a F Germany 16 years Ex9: I222V

M 15 years10a F Germany 13 years Ex9: I222V (M)236 M Austria 12 years Ex9: I222V164 M Turkey 4 years Ex30: S1156L (M)179 M Germany 2 months Ex30: S1156L


Table 1. continued.


232 F Czech Republic 11 years Ex5: F129SF154 M United Kingdom Death (perinatal) Ex16: G466A (P)

F 6 yearsF559 F French-Canadian 2 years Ex16: G466A (M)65 M Germany 12 years Ex22: P755L

M Death (perinatal)F0a F Germany 16 years Ex25: V865F

M 13 years176 M Switzerland 6 years Ex27: A1030VF608 F Bulgaria 24 years Ex32: S15841F102a F Germany 11 years Ex33: C1787F

M 8 years257 M Germany 6 years Ex43: I2331K (M)5 F Germany 14 years Ex54: R2831KF573 M Czech Republic 6 months Ex57: I2957TF539 F Turkey 3 years Ex57: G2967W

M 1 year81 M Germany 15 years Ex58: H3029P8 F Germany 14 years Ex58: I3051T

M 25 yearsF465 F Portugal 21 years Ex58: R3240Q

F 14 yearsF316b F Germany 24 years Ex61: R3482C

M Death (perinatal)96a F Germany 13 years Ex61: M3642T (M)133 M Germany 8 yearsF271 F Hungary 7 years

M 5 years146 M Germany 3 years249 M Czech Republic 6 years158 M Germany 8 years

Abbreviations are: (C), parental consanguinity known; (P), paternally inherited allele; (M), maternally inherited allele; novel PKHD1 mutations are shown in bold characters.

aClinical data up to 1995 of this patient were included in Zerres et al (Acta Paediatr 85:437–445, 1996).bMutation(s) of this family have been described in Bergmann et al (J Am Soc Nephrol 13:76–89, 2003).Patients with a predominant liver phenotype (N = 15 defined as early esophageal variceal bleeding or ascites and (nearly) normal renal function) are shown on a

light gray background.Patients requiring renal replacement therapy are shown on a gray background.

[26, 27]. In case data on height were missing in patientswith no documented growth retardation, these were ad-justed by assuming an average height for age. Chronicrenal insufficiency (CRI) was defined as GFR <75% nor-mal adjusted for age. Kidney length was measured by ul-trasound in relation to height [28] and expressed as SDscores. The date of diagnosis was requested for ARPKD-related renal morbidities [e.g., systemic hypertension (de-fined as the date antihypertensive therapy was initiated),growth retardation (defined as height < 2 SD for age),and occurrence of urinary tract infections (UTIs) (cys-titis or pyelonephritis)]. The date of diagnosis was alsodocumented for ARPKD-related hepatobiliary morbidi-ties [e.g., portal hypertension (defined by sonographicevidence of hepatosplenomegaly and directional rever-sal of portal vein flow), esophageal variceal bleeding,splenomegaly, hypersplenism, and cholangitis]. When ap-plicable, dates and results of kidney and liver biopsies aswell as transplantations were recorded. Patient data wereentered in a secure database and analyzed with Excel(Microsoft). Comparisons between categorical variables

were analyzed with the v 2 test. Spearman’s rank ordercorrelation coefficient r was used to examine the rela-tionship between age at onset of various comorbidities.Kaplan-Meier curves were constructed for survival ratesand renal function. For all analyses, a P value of <0.05was deemed statistically significant.

Haplotype analysis

At least two informative markers flanking PKHD1were typed in each family with more than one child aspreviously described [21, 29, 30]. All families were com-patible with linkage to the PKHD1 locus on chromosome6p12. In case of sporadic patients, a minimum of three in-tragenic markers (D6S1344, D6S243, and D6S1714) weretested to check for homozygous haplotypes in the indexpatient. After identification of a recurrent PKHD1 mu-tation, allele sharing analysis based on the mentionedintragenic polymorphic sites was done in the respectivefamilies to study a possible founder effect.


Mutation analysis

Mutation screening was done for the 66 exons encodingthe 4074 aa polyductin protein (GenBank NM 138694)by denaturing high-performance liquid chromatogra-phy (DHPLC) on a Wave Fragment Analysis System(Transgenomic, Crewe, UK), as recently described in de-tail [31]. Samples exhibiting an aberrant elution profilewere subjected to direct sequence analysis. Once two mu-tations had been identified in a patient, the altered alleleswere proven to reside on separate chromosomes by di-rect sequencing of the parental DNA probes as far asthese were available (in any but four cases of the 72 pedi-grees in whom both mutations were identified). DNAsamples from 200 apparently unrelated normal controlsubjects were run under appropriate DHPLC conditionsto exclude a possible polymorphism or harmless sequencevariant in case of a missense mutation or subtle in-framechange.

RESULTS

A total of 186 ARPKD patients (97 males and 89females) representing 126 apparently unrelated nuclearfamilies were enrolled in the clinical and molecular di-agnostic studies; 164 of these patients (83 males and 81females) survived the neonatal period. In 48 pedigreesmore than one child was affected (multiplex families).Twenty families were characterized by gross intrafamil-ial phenotypic variability (peri-/neonatal demise in oneand survival into childhood in another affected sibling)(Fig. 1). The mean observation period was 6 years (range0 to 35 years). Of all patients, the mean age at last re-ferral was 10.3 years. Ages at diagnosis of the studycohort and initial clinical features of the 143 children di-agnosed after birth are given in Table 2. Results for renalfunction analysis, kidney lengths, and survival rates aredepicted in Figures 2 to 5 and did not reveal any statis-tically significant gender difference. Correlations amongARPKD comorbidities revealed a significant relationshipbetween the development of chronic renal failure (CRF)and systemic hypertension (r = 0.99). Significant correla-tions could be further demonstrated between renal andhepatobiliary-related morbidity (between CRF and por-tal hypertension, r = 0.84/between serum creatinine >200lmol/L and portal hypertension, r = 0.70).

PKHD1 mutations

We performed a systematic mutation screen byDHPLC of the 66 PKHD1 exons encoding the longestcontinuous ORF. A total of 193 out of expected 252 mu-tations (126 families) was observed corresponding to adetection rate of 76.6%. Seventy of the mutations de-tected were novel ones not reported previously. We wereable to characterize two underlying mutations in 72 fam-

ilies (57%) and one mutation in 49 pedigrees (39%). Nomutation could be detected in five families (4%). Thus,in 96% of families one or two mutations were identi-fied. In all individuals with two mutations, the mutantalleles were proven to reside on separate chromosomes.PKHD1 changes were dispersed over the protein’s entirelength with most mutations being unique to single fam-ilies (“private mutations”). All mutations and polymor-phisms/sequence alterations described so far are listedon the PKHD1 mutation database (URL: http://www.humgen.rwth-aachen.de).

The observed changes fell into two major groups:missense (N = 149) (77%) and truncating mutations(N = 44) (23%). Truncating alterations (23 nonsense, 17frameshifting small deletions/insertions/duplications, andfour splice-site mutations) are expected to produce nullalleles. Canonical splice-site changes c.977 − 1G > A andc.3228 + 1G > T likely lead to skipping of the affectedexon [31] with subsequent out-of-frame fusion of exons,whereas the corresponding transcripts of c.1512 + 1G >

A missing exon 16 and c.6333 − 2A > G missing exon40 may restore the reading frame and actually may notmerit the designation truncating.

About three fourths of the mutations were nonconser-vative missense substitutions that were neither found incontrols nor previously described as polymorphisms [9,10, 21, 31, 33–35]. All but twelve missense changes werefound to replace residues conserved in the murine or-tholog (NM 153179). Missense changes not found to re-place residues conserved in mouse polyductin were stillconsidered pathogenic as further evidence supports theirpathogenicity. First, the nature of each residue in themurine ortholog was different from the amino acid in-corporated by the mutation. Moreover, the mutationsR92G, H916P, I2331K, and R2840C were either con-served in polyductin-like proteins from rat (XM 236979and XM 236984) or chicken. The mutations I2331K,R2840C, and I3177T have been described many times inprevious mutational studies [9, 10, 21, 31, 33, 34]. Finally,the remaining changes each represent nonconservativeamino acid substitutions that can be expected to causeconsiderable changes on protein structure.

Probands carrying putative missense substitutionswere included in the screen of the rest of the gene. Nomore than a maximum of two putative pathogenic muta-tions were observed in all but three cases. In any of thesepatients (57, 183, and F507), both missense changes resid-ing on the same parental chromosome were shown to benonconservative and affected amino acids conserved inthe mouse ortholog. Functional studies will be required toassess the relative pathogenic potential of these changes.Moreover, screening further patients for PKHD1 muta-tions will also contribute to resolve the matter. Enrich-ment of an alteration in the ARPKD population mightbe good evidence that it is disease associated. Notably,


Sibship

2

475

267

113

599

3

105

104

329

625

48

578

59

12

328

17

313

371

127

549

242

1

507

4

383

151

196

236

262

58

92

306

T36MT36M

N1744HN1744HG3178CG3178CR496XI2331KI1998TQ3407T

IVS13-1G>AI2331KI222V

R3240XC1472YR3240XI1687TR1775XI1687TQ3407X

N1919fsXI2427T

I2331fsXR2795SR3107XD3293VT36MI222VT36MP676RT36M

G1123ST36M

V1789LT36M

C2422GT36M

I2957TT36M

A3097TT36M

R3240QT36M

V3471GI2331KS2861GP3740SP805L/I3177T

W2690RR92G

T2140PW656CG1123SI948SI1687TE1124KR3240Q

P356fsX

L1966fsX

D3230fsX

G3359fsX

birth 1y 2y 3y 4y 5y 7y 9y 11y 13y 15y 20y 25y 30y 35yGenotype

Fig. 1. Summary of the clinical outcome of affected siblings of 48 multiplex pedigrees. Genotypes of families are given in the second column.Males: squares; females: circles. Diagonal lines denote deceased patients. Different fillings of symbols are explained below the figure. Abbreviationsare: SCr, serum creatinine; RTX, renal transplantation; CAPD, continuous ambulatory peritoneal dialysis; HD, hemodialysis; portal HTN, portalhypertension; LTX, liver transplantation.


346

375

32

27

9

231

19

154

65

0

102

55

539

465

316

271

Q3589X

T36M

T36M

T36M

T36M

T36M

I222V

G466A

P755L

V865F

C1787F

I2331KI2957T

G2967W

R3240Q

R3482C

Peri-/neonatal demise

SCr >200 µmol/L

RTX/CAPD/HD

Portal HTN/splenomegaly/Caroli’s disease

Esophageal varices/variceal bleeding/ascites/LTX

Fig. 1. (continued.)

Table 2. Ages and clinical features at diagnosis

Ages at diagnosis of the study cohort (N = 186)Prenatal 23%1st month 31%2nd to 12th month 16%2nd to 5th year 19%>5th year 11%

Initial clinical features in childrendiagnosed after birth (N = 143)(more than one category possible)

Palpable abdominal masses 68%Systemic hypertension 43%Hepatosplenomegaly 21%Urinary tract infection 12%Respiratory distress 9%Esophageal varices 5%

the substitutions G448R, M997K, and S2861G have beendescribed in other ARPKD patients further strengthen-ing their pathogenic potential. In two other families (F4and F538) the missense changes P805L and I3177T weretandemly inherited on one parental allele [35].

PKHD1 sequence variants/polymorphisms

In addition to the putative pathogenic changes de-scribed, a vast number of probably harmless polymor-phisms and sequence variants were encountered. Alter-ations were considered as neutral in case they were (1)intronic changes beyond the consensus splice sites, (2) lo-cated in the 5′ or 3′ UTR, (3) silent exonic changes (notaltering an amino acid), (4) a substitution which did notsegregate with the disease or was found in a set of 200normal controls, or (5) already been described in previ-ous studies [9, 10, 21, 31, 33–35]. Overall, we detected 20amino acid substitutions, several of them of nonconser-vative nature that were also present in normal chromo-somes (Table 3). Notably, nine of those missense polymor-phisms (45%) affect residues not conserved in the mouseprotein. Furthermore, we observed 28 silent exonic nu-cleotide changes not causing an amino acid substitution(Table 4) and 30 adjacent intronic sequence variations(Table 5).


0

10

20

30

40

50

60

70

80

90

100

Pro

port

ion

of p

atie

nts,

%

0 5 10 15 20 25

Age, years

X/M(5x)M/M(2x)M/-(2x)X/-

X/MM/- M/- M/M M/M(2x)

-/- M/-

M/-

X/MM/MX/- M/M(2x) X/M

X/-M/-(2x)

X/MM/-

X/M(2x)M/-

X/M

M/M

X/MM/M(2x)

Fig. 2. Kaplan-Meier curves were constructed for renal function analysis of the 164 neonatal survivors included in the present study. The lowercurve depicts the proportion of patients with serum creatinine levels below 100 lmol/L, the middle curve those of individuals with serum creatininevalues below 200 lmol/L, and the upper one indicates the proportion of patients without end-stage renal disease (ESRD) at a given age. Accordingly,our study cohort revealed an actuarial renal survival rate (end point defined as start of dialysis/RTX or by death due to ESRD) of 86% at 5 years,71% at 10 years, 66% at 15 years, and 42% at 20 years. At the end of the study, 38 patients received dialysis treatment for ESRD, were transplanted,or died due to ESRD. Mutation types identified in these 38 patients are shown above the ESRD curve indicating the age when the respectiveindividual first received renal replacement therapy (X is truncating mutation, M is missense mutation, and − is an unidentified allele). The meanage ESRD occurred in patients with a truncating mutation on one parental allele was 7.2 years (N = 16), in patients carrying missense changes onboth alleles 10.1 years (N = 11).

To determine whether some of these changes could dis-rupt predicted splice sites and may, thus, have a potentialpathogenic impact, we generated splice-site scores (pre-dicting the theoretic impact upon donor and acceptor sitestrength and the probability that sequence changes cre-ated novel splice sites) for the wild-type and variant se-quences of those changes that were not identified amongnormal controls. For this purpose we used the splice-siteprediction algorithm at http://www.fruitfly.org/seq tools/splice.html. In cases where this algorithm did not predictan acceptor/donor splice site neither for the wild-type norfor the variant sequence, a further splice site prediction al-gorithm at http://www.cbs.dtu.dk/services/NetGene2/ wasused. However, all but four variants did not result in anychange of the predicted splice-site score (maximum al-teration from 86% to 50% for donor site in case of c.51A> G/A17A). In principle, a pathogenic potential will ul-timately need to be assessed by functional studies.

As the PKHD1 gene is predicted to comprise a min-imum of 86 exons [10], we screened by DHPLC allpatients of the current study with one or no PKHD1mutation in the longest ORF for changes in the pre-

dicted alternatively spliced exons (AY129465). Single nu-cleotide changes involving these exons were observedin several patients. However, most of these alterationswere also found among normal controls with compara-ble allele distributions among patients and controls. Thechanges g.157552C > T, g.175388A > C, g.344048A > T,g.406265C > T have been identified in one patient each,however, were also detected in one out of 400 controlchromosomes. Principally, a potential pathogenic impactof these sequence variations has to await the definitionof transcripts containing alternative exons and their pre-dicted reading frames.

Genotype-phenotype correlations

As previously described [21], all patients carrying atermination-type mutation on both parental alleles diedshortly after birth, regardless of the site of the prema-ture stop of translation. This observation has been subse-quently confirmed by other studies on PKHD1 mutations[31, 33, 34] and could also be corroborated by the presentsurvey in which no family with two truncating mutationshas been identified.


0

100

200

300

400

500

600

SC

r, µm

ol/L

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Age, years

Fig. 3. The development of serum creatinine (SCr) values of 22 patients (13 males and 9 females) with at least one serum creatinine level above200 lmol/L during the observation period is depicted. The color of the curves indicates the mutation status of patients (red is the compoundof truncating and missense mutation; green is the missense mutation on both alleles; blue is the single truncating or missense change, secondmutation not identified). Patients whose first available serum creatinine value was already above 200 lmol/L (N = 16) are not included in this figure.Deterioration of renal function rapidly set in above a serum creatinine level of about 200 lmol/L. Almost every patient who reached that pointdeveloped end-stage renal disease (ESRD) within a period of less than 5 years. The mean age patients with a truncating mutation on one allelereached a serum creatinine value of about 200 lmol/L was 4.1 years (N = 7), patients carrying missense changes on both alleles had a mean age of7.2 years (N = 10).

0

2

4

6

8

10

12

14

16

18

Sta

ndar

d de

viat

ion

of n

orm

al k

idne

y le

ngth

, SD

0 5 10 15 20 25Age, years


X/M(3x)X/-(3x) X/M

X/MM/-

M/M

M/MM/-

M/-

M/M

70

75

80

85

90

95

100

Pro

port

ion

of s

urvi

vors

, %

0 5 10 15 20 25

Age, years

Fig. 5. Kaplan-Meier survival curve for all 186 individuals of the current series (overall survival rate at 1 year of age of 85% (81% males and 89%females) (NS) with most cases of death in the peri-/neonatal period related to respiratory insufficiency). Of those patients who survived the firstmonth of life, 95% were alive at 1 year, 94% at 5 years, and 92% at 10 years. In total, a number of 13 neonatal survivors (8%) (6 males and 7 females)died during the observation period at a median age of 5.6 years. Mutation types identified in these 13 patients are shown below the curve when therespective individual passed away (X is truncating mutation; M is missense mutation; − is unidentified allele). In six patients death occurred duringthe first year of life with predominant extrarenal causes [cardiorespiratory insufficiency (N = 3), end-stage renal disease (ESRD) with sepsis (N =2), pneumonia (N = 1)]. ESRD with sepsis was also causative for demise in six patients at later ages of life (1 to 23 1

2 years). One male died at theage of 22 years due to ESRD complicated by pneumonia and esophageal variceal bleeding. Perpendicular bars indicate the age of surviving patientsat last examination (censored patients).

Except for genotype-phenotype correlations estab-lished for the type of mutations, there have been madeno efforts so far to determine the genetic contributions ofsingle missense alterations to the phenotypes observed.Missense mutations may range in their effects from hy-pomorphic alleles that allow for a milder clinical courseto complete loss of function variants resembling trun-cating alterations. However, such correlations are ham-pered given the lack of mutational hotspots and markedallelic heterogeneity at PKHD1 with the majority of mu-tations unique to a single family. Nevertheless, the bulkof mutational data that has been identified since cloning

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Fig. 4. Kidney lengths measured by ultrasound related to age of 78 patients (40 male and 38 females) of the present study. In most individualsmultiple values at different ages were encountered, however, for clarity only initial and final values are presented. All but six patients (92%) hada kidney length above or on the 97th centile for age ( = normal kidney length for age +2 SD scores). In males (squares), kidney lengths rangedbetween +1 and +17 SD, in females (circles) between 0 and +12 SD. The color indicates the mutation status of patients (red is the compound oftruncating and missense mutation; green is the missense mutation on both alleles; blue is the single truncating or missense change, second mutationnot identified). Correlations were neither observed between kidney length and serum creatinine levels, between kidney length and duration of thedisease nor between kidney length and mutation type.

of PKHD1 should allow the start of categorizing mis-sense mutations. Due to multiple allelism and the highrate of different compound heterozygotes, our approachwas based on the premise that termination-type muta-tions likely all have a uniform and devastating effect withloss of function. Thus, one would expect the type andposition of missense changes to determine the clinicalcourse in compound heterozygotes with one truncatingand one missense mutation. In line with this hypothesis,all missense mutations encountered in the present seriesthat occur in compound with a chain-terminating muta-tion may rather be associated with a moderate phenotype.


Table 3. PKHD1 amino acid substitutions also found among normal controls

cDNA/nucleotide Allele frequencyLocation change aa change Mouse aa among controls References

Ex16 c.1463G > C R488P L486 1/200 21, 35, Current studyEx19 c.1736C > T T579M T577 2/200 21, 35, Current studya

Ex22 c.2170C > T P724S P722 1/200 21, 35, Current studyEx22 c.2278C > T R760C C758 46/200 9, 10, 21, 33–35, Current studya

Ex24 c.2489A > G N830S N828 3/200 21, 33, Current studya

Ex30 c.3407A > G Y1136C Y1134 1/200 21, 35, Current studyEx32 c.3785C > T A1262V V1260 9/200 9, 21, 33, 35, Current studya

Ex53 c.8345G > C G2782A C2774 2/200 Current studya

Ex55 c.8606C > A T2869K V2863 10/200 21, 34, 35, Current Studya

Ex58 c.9215C > T A3072V A3066 1/200 21, 35, Current studyEx58 c.9262G > A D3088N D3082 1/200 21, 35, Current studyEx58 c.9320G > C R3107P G3101 1/200 21, 35, Current studyEx58 c.9415G > T D3139Y N3132 4/200 9, 10, 21, 33–55, Current studya

Ex59 c.9577G > A V3193I V3186 2/200 Current studya

Ex61 c.10515C > T S3505R S3498 6/200 9, 33, Current studya

Ex65 c.11525G > T R3842L R3835 8/200 21, 35, Current studya

Ex66 c.11696A > G Q3899R Q392 50/200 9, 10, 21, 33–35, Current studya

Ex66 c.11714T > A I3905N K3898 1/200 Current studya

Ex67 c.11878G > A V3960I V3954 4/200 9, 10, 21, 35–35, Current studya

Ex67 c.12143A > G Q4048R Q4037 60/200 9, 21, 33, 35, Current studya

9Ward et al, 2002; 10Onuchic et al, 2002; 21Bergmann et al, 2003; 33Rossetti et al, 2003; 34Furu et al, 2003; 35Bergmann et al, 2004.aSharp et al, database submission.GenBank NT 007592.13; NM 138694.2; nucleotide +1 is the A of the ATG-translation initiation codon.

Albeit the dataset is still limited and this concept repre-sents a simplified view, it might constitute a reasonableapproach to start classifying missense mutations into se-vere and moderate/mild changes (Table 6).

The assessment of the character of a single missensechange may in particular come from patients homozy-gous for a given amino acid substitution. F372L, N1744H,V2771A, I2851T, H3124Y, and G3178C were observedhomozygously within the current study. A common an-cestral origin in the Turkish population seems most likelyfor the missense changes I473S and Y1838C.

Renal phenotypes of neonatal survivors of the presentand previous studies [9, 21, 33] who carry the sameset of PKHD1 mutations are compared in Figure 6.Y1838C/Y1838C, I1687T/X, and V3471G/R496X re-sulted in early ESRD, whereas T36M/C2422G andI2331K/X were associated with minor renal disease.

DISCUSSION

Since the PKHD1 gene was only recently character-ized, previous studies reporting ARPKD-related mor-bidities merely described the clinical course of patientswithout providing mutational data. On the other hand,the initial studies describing PKHD1 mutations clearlyfocused on molecular genetic aspects. We performedPKHD1 mutation analysis in a cohort of 186 ARPKDpatients and examined the clinical course over a meanobservation period of 6 years. This is the first study thatcombines PKHD1 mutations with long-term follow-updata considerably extending the knowledge on genetic

contributions to phenotypes observed in ARPKD. Thisinformation should provide a rational starting point tofurther investigate genotype-phenotype correlations.

Clinical studies on ARPKD differed widely by their se-lection criteria of patients and their mode of analysis ofdata (Table 7). Patients of this and our previous study [24]as well as of the survey by Guay-Woodford and Desmond[7] were mostly recruited from departments of pediatricnephrology. As a consequence, individuals with an earlylethal form of ARPKD are underrepresented. Capisondaet al [37] and Gagnadoux et al [38] reported clinical out-comes of patients from specialized single centers. Mostof the individuals in the study by Roy et al [8] had previ-ously been reported by Kaplan et al [39] who exclusivelyincluded patients with pathoanatomically proven CHF.Kaariainen, Koskimies, and Norio [40] analyzed mainlydata obtained from Finnish death registers. Inclusion cri-teria of the study by Cole, Conley, and Stapleton [41] werediagnosis within the first year of life and survival of theneonatal period.

Overall, our data suggest renal morbidities to be posi-tively related with hepatobiliary morbidities rather thanorgan-specific progression of ARPKD [7]. Thus, singleARPKD patients with an organ-specific phenotype maybe regarded as exceptions to the rule. It might be promis-ing to explore genotype-phenotype correlations and(genetic) modifiers in these patients.

Mutation data

Overall, the diagnosis was substantiated in 96% ofpedigrees by identifying one or two PKHD1 mutations.


Table 4. PKHD1 silent changes

AllelecDNA frequencynucleotide among

Exon change aa change controls SS WT SS variant Comment Reference

Ex2 c.51 > G A17A 0/200 D: 0.86 D: 0.50 Decreased strength of Noveldonor site

Ex4 c.214C > T L72L 25/200 21, 33–35, Current studya

Ex4 c.234C > T D78D 15/200 9, 21, Current studya

Ex14 c.999A > G E333E 6/200 21, 35, Current studyEx15 c.1122A > C A374A 0/200 A: 0.96 A: 0.97 Probably no effect NovelEx15 c.1185T > C D395D 4/200 33, Current studya

Ex17 c.1587T > C N529N 10/200 9, 10, 21, 33–35, Current studya

Ex20 c.1950G > A R650R 0/200 D: 0.67 D: 0.67 Probably no effect NovelEx21 c.2046A > C P682P 20/200 21, 33–35, Current studya

Ex27 c.2853C > T T951T 2/200 9, 35, Current studyEx30 c.3393G > A A1131A 1/200 21, 34, 35, Current studyEx30 c.3537T > C N1179N 2/200 21, 35, Current studya

Ex32 c.3756G > C L1252L 7/200 9, 21, 33, Current Pagea

Ex32 c.4035C > A G1345G 0/200 A: 0.77 A: 0.77 Probably no effect Current studya

Ex32 c.4920A > C V1640V 1/200 9, 10, 21, 33–35, Current studya

Ex36 c.5896C > T L1966L 5/200 21, 34, 35, Current studya

Ex39 c.6462T > C V2154V 0/200 D: 0.78 D: 0.78 Probably no effect 21, Current studyEx40 c.6636C > T A2212A 0/200 D: 1.00 D: 1.00 Probably no effect 21, Current studyEx41 c.6777C > T F2259F 1/200 NovelEx43 c.6900C > T N2300N 1/200 21, 35, Current studya

Ex48 c.7587A > G G2529G 65/200 9, 10, 21, 33–35, Current studya

Ex49 c.7764A > G L2588L 20/200 9, 10, 21, 33–35, Current studya

Ex56 c.8673C > G R2891R 2/200 Current studya

Ex58 c.9237G > A A3079A 20/200 9, 10, 21, 33–35, Current studya

Ex58 c.9492C > T S3164S 1/200 9, 10, Current studyEx61 c.10434T > A P3478P 4/200 9, 35, Current studyEx61 c.10521C > T H3507H 5/200 9, 10, 21, 33–35, Current studya

Ex63 c.11340T > C P3780P 10/200 9, 10, 21, 33–35,Current studya

Abbreviations are: SS, Splice-site score using the splice-site prediction algorithm at http://www.fruitfly.org/seq tools/splice.html; A, splice acceptor site; D, donor splicesite; WT, wild-type.

9Ward et al, 2002; 10Onuchic et al, 2002; 21Bergmann et al, 2003; 33Rossetti et al, 2003; 34Furu et al, 2003; 35Bergmann et al, 2004.aSharp et al, database submission. GenBank NT 007592.13; NM 138694.2; nucleotide + 1 is the A of the ATG-translation initiation codon.

In total, 193 of 252 expected mutations (76.6%) couldbe identified. Thus, our data suggest that the previousrelatively low mutation detection rates obtained amongpatients with a milder clinical course were most likelydue to methodologic issues. Although the current rate islower than the 85% obtained in a recent study of ARPKDpatients with peri-/neonatal demise [31], this detectionrate is acceptable and makes PKHD1 mutation screen-ing both robust and efficient for clinical use to establishthe diagnosis of ARPKD.

As described, most PKHD1 mutations are private andrecurrence of changes is often based on a common ances-tral origin [35]. The most common mutation, c.107C > T(T36M), has been described in each PKHD1 mutationstudy reported to date [9, 10, 21, 31, 33–35] and accountsfor approximately 15% to 20% of mutated alleles. Thereis conflicting data on whether it is an ancestral change oroccurs due to a frequent mutational event. Ultimately, itcannot be excluded that some of the mutated c.107C >

T alleles represent a founder effect in the Central Euro-pean population where it is particularly frequent. How-ever, there is compelling evidence that T36M constitutes

a mutational “hotspot,” most likely due to methylation-induced deamination of the mutagenic CpG dinucleotide[36]. Within the present study, T36M was identified in amultitude of obviously unrelated families of different eth-nic origins on various haplotypes. For instance, patientsof nonconsanguineous Finnish family F2 homozygous forc.107C > T were shown to harbor differing haplotypes.Further evidence of recurrence and against a common an-cestral origin of mutated c.107C > T alleles were demon-strated by diverse haplotypes among German-Austrianpedigrees F371 and ID206 carrying the same set of mis-sense mutations (c.107C > T + c.7264T > G).

Given multiple allelism and a proposed carrier fre-quency of approximately 1:70 in nonisolated populations[2], the prevalence of any individual PKHD1 mutationis expected to be low. Therefore, despite screening ofa large number of normal chromosomes, evidence fromsegregation analysis, lack of further changes in the rest ofthe gene, and nonconservative nature of amino acid sub-stitutions, it cannot be definitely ruled out that some ofthe PKHD1 missense changes will finally be revealed asnon-pathogenic. Enrichment of a certain PKHD1 change


Table 5. Intronic PKHD1 polymorphisms/sequence variants

Allelefrequency

amongIVS Location controls SS WT SS variant Comment Reference

IVS2 c.53 − 24A > G 0/200 A: 0.82 A: 0.82 Probably no effect NovelIVS7 c.527 + 19T > C 52/200 21, 39, 35, Current studya

IVS9 c.667 + 18C > T 0/200 D: ND (ND) D: ND (ND) Probably no effect NovelIVS9 c.668 − 30T > C 1/200 21, 35, Current studyIVS15 c.1234 − 10T > A 6/200 34, 35, Current studya

IVS15 c.1234 − 5C > T 1/200 Current studya

IVS18 c.1694 − 33C > G 4/200 34, 35, Current studya

IVS19 c.1837 − 23A > G 0/200 A: 0.63 A: 0.63 Probably no effect NovelIVS22 c.2279 + 13T > G 23/200 21, 33, 35, Current studya

IVS22 c.2280 − 35C > A 0/200 A: 0.77 A: 0.77 Probably no effect NovelIVS23 c.2407 + 50C > T 82/200 21, 33, 35IVS29 c.3365 − 6A > G 0/200 A: ND (0.95) A: ND (0.95) Probably no effect NovelIVS32 c.5236 + 14A > G 11/200 21, 34, 35, Current studya

IVS35 c.5752 − 18del C 0/200 A: 0.97 A: 0.97 Probably no effect NovelIVS37 c.6122 − 9C > G 0/200 A: 0.90 A: 0.83 Decreased strength of Novel

acceptor siteIVS46 c.7351 − 8A > T 0/200 A: 0.88 A: 0.97 Increased strength of Novel

acceptor siteIVS48 c.7734 − 19ins TT 0/200 A: ND (ND) A: ND (ND) Extension of Novel

polypyrimidine stretchIVS48 c.7734 − 4T > C 35/200 35, Current studya

IVS49 c.7912 − 5T > G 0/200 A: ND (ND) A: ND (ND) Possibly decreased Novelstrength of acc. site

IVS52 c.8302 + 12T > A 15/200 21, 34, 35, Current studya

IVS53 c.8441 − 32G > C 42/200 21, 33–35, Current studya

IVS54 c.8643 − 73C > T 27/200 34, Current studya

IVS54 c.8643 − 24A > G 10/200 21, 34, 35, Current studya

IVS55 c.8798 − 19A > C 9/200 21, 34, 35, Current studya

IVS57 c.8797 + 19A > C 0/200 D: 0.98 D: 0.98 Probably no effect NovelIVS58 c.9829 + 70T > G 0/200 D: 0.93 D: 0.93 Probably no effect NovelIVS61 c.11174 + 10A > G 0/200 D: 1.00 D: 1.00 Probably no effect NovelIVS61 c.11174 + 11A > G 5/200 21, 33, 35, Current studya

IVS61 c.11174 + 26A > G 0/200 D: 1.00 D: 1.00 Probably no effect NovelIVS62 c.11310 + 19C > T 0/200 D: 0.89 D: 0.89 Probably no effect Novel

21Bergmann et al, 2003.33Rossetti et al, 2003.34Furu et al, 2003.35Bergmann et al, 2004.aSharp et al, database submission.GenBank NT 007592.13; NM 138694.2; nucleotide +1 is the A of the ATG-translation initiation codon.Abbreviations are: SS, splice-site score using the splice-site prediction algorithm at http://www.fruitfly.org/seq tools/splice.html; A; splice acceptor site; D, donor

splice site; WT, wild-type; ND; not detected (in cases where the above mentioned algorithm did not predict an acceptor/donor splice site neither for the WT nor for thevariant sequence, a further splice site prediction algorithm at http://www.cbs.dtu.dk/services/NetGene2/ was used, results are given in parentheses).

in the ARPKD population and segregation with the dis-ease might finally be the best evidence that a change isin fact disease associated. Difficulties in evaluating thepathogenicity of some mutations emphasize the need tocatalogue all PKHD1 changes in a locus specific database(http://www.humgen.rwth-aachen.de).

The major cause of missing mutations in a consider-able proportion of alleles may be limited sensitivity ofthe screening method even though DHPLC has beenshown to be efficient and effective in analyzing largegenes [45–47]. Further reasons for failure to detect mu-tations may be changes residing in regulatory elementsor introns distant from the splice donor and acceptorsites that have not been screened, but which may leadto aberrant splicing [32, 48, 49]. Moreover, some of the

changes currently categorized as nonpathogenic may ul-timately turn out to be disease associated. It might behypothesized that some silent exonic changes and a sub-set of adjacent intronic sequence variations do also havean effect on PKHD1 splicing [e.g., by affecting splice en-hancer or silencer sites (ESE/ISE or ESS/ISS)]. However,at present stage of knowledge we categorized them asnonpathogenic changes. Functional and mRNA studiesare needed to prove any possible pathogenic effect. Inpatients without a detectable PKHD1 mutation misdiag-nosis of ARPKD has to be considered. However, normalparental ultrasound scans and typical clinical and sono-graphic features were documented in any of the five casesof the present study. In one patient, CHF has even beenconfirmed histologically. Moreover, although the disease


Table 6. Categorization of recurrent PKHD1 missense mutations

Pedigrees Peri-/neonatal Gross intrafamilial ModerateMissense mutation (patients) demise variation phenotype Implication

I473S/I473S 2 (2) 2 (2) ModerateI1687T/X 2 (3) 2 (3) ModerateI2957T/X 2 (2) 2 (2) ModerateI3468V/X 2 (2) 2 (2) ModerateI222V/X 6 (10) 2 (3/2) 4 (5) Rather moderateY1838C/Y1838C 3 (3) 3 (3) Rather moderateY1838C/X 1 (1) 1 (1)I1998T/I1998T 1 (1) 1 (1) Rather moderateI1998T/X 1 (2) 1 (1/1)I2331K/X 4 (6) 1 (1/1) 3 (4) Rather moderateT36M/T36M 2 (3) 1 (1) 1 (1/1) Rather severeT36M/X 12 (15) 11 (13) 1 (2)R760H/R760H 2 (4) 1 (1) 1 (2/1) Rather severeH3124Y/H3124Y 2 (2) 1 (1) 1 (1)H3124Y/X 1 (1) 1 (1) Rather severeV3471G/X 5 (5) 3 (3) 2 (2) Rather severeP805L/X 2 (2) 2 (2) SevereG1122S/X 2 (2) 2 (2) SevereI2303F/X 3 (3) 3 (3) SevereI3177T/X 2 (2) 2 (2) SevereR3482C/R3482C 2 (5) 2 (5) Severe

Recurrent PKHD1 missense mutations (occurring in at least two unrelated families) have been classified in case of a truncating mutation on the other parentalallele or in case of a homozygous amino acid substitution. The number of patients carrying the respective change is given in parentheses. In case of gross intrafamilialvariability among siblings, the number of severely and moderately affected patients is separated by a slash.

0

50

100

150

200

250

300

350

400

ESRD

SC

r, µm

ol/L

0 2 4 6 8 10 12 14 16 18 20 Years

I473S/I473S

Y1838C/Y1838C

I222V/X

I1687T/X

I2331K/X

V3471G/R496X

T36M/C2422G

Fig. 6. Renal disease variability among patients of the present and two previous studies [9, 33] who survived the neonatal period and carry thesame set of PKHD1 mutations. Different symbols represent different mutation sets (depicted below). In case consecutive measurements of serumcreatinine (SCr) levels were available, these are linked by lines.


Tabl

e7.

Sum

mar

yof

findi

ngs

obta

ined

inou

rse

ries

com

pare

dw

ith

prev

ious

stud

ies

Gua

y-W

oodf

ord

and

Des

mon

dC

apis

onda

etal

Gag

nado

uxet

alK

aari

aine

nC

urre

ntst

udy

(200

3)(2

003)

Roy

etal

(199

7)Z

erre

set

al(1

996)

(198

9)K

apla

net

al(1

989)

etal

(198

8)C

ole

etal

(198

7)

Pat

ient

s(n

umbe

r)18

6(1

64)

166

3152

115

3355

73(1

8ne

onat

alsu

rviv

ors)

17

Age

atdi

agno

sis

23%

pren

atal

46%

pren

atal

32%

pren

atal

10%

pren

atal

31%

<1

mon

th27

%<

1m

onth

23%

<1

mon

th41

%<

1m

onth

33%

<1

mon

th42

%<

1m

onth

72%

<1

mon

th16

%1

to12

mon

ths

30%

>1

year

11%

1to

12m

onth

s16

%>

1ye

ar19

%1

to12

mon

ths

26%

>1

year

85%

<1

year

15%

>1

year

23%

1to

12m

onth

s26

%>

1ye

ar55

%1

to18

mon

ths

12%

6to

11ye

ars

42%

1to

12m

onth

s16

%<

1ye

ar6%

<1

year

22%

>1

year

100%

1to

12m

onth

s(i

nclu

sion

crit

eria

)R

enal

func

tion

86%

GF

R<

3rd

cent

ilefo

rag

em

edia

nag

eC

RF

4.0

year

s29

%E

SRD

(by

10ye

ars)

42%

GF

R<

3rd

cent

ilefo

rag

e13

%E

SRD

51%

GF

R<

80m

L/m

in/1

.73

m2

16%

ESR

D

33%

ESR

D(b

y15

year

s)72

%G

FR

<3r

dce

ntile

for

age

10%

ESR

D

42%

GF

R<

80m

L/m

in/1

.73

m2

21%

ESR

D

58%

SC>

100

lmol

/mL

82%

GF

R<

90m

L/m

in/1

.73

m2

35%

GF

R<

40m

L/m

in/1

.73

m2

29%

ESR

D

Kid

ney

leng

th92

%>

2SD

NA

NA

NA

68%

>2

SD10

0%>

2SD

NA

NA

NA

Hyp

erte

nsio

n(%

ondr

ugtr

eatm

ent)

76%

(80%

mal

ean

d72

%fe

mal

e)m

edic

atio

nst

arte

dat

med

ian

age

of3

year

s(5

3%du

ring

first

6m

onth

s)

65%

55%

60%

(by

15ye

ars)

70%

76%

65%

61%

100%

(dru

gtr

eatm

ento

rB

P>

95th

cent

ile)

Gro

wth

reta

rdat

ion

16%

<2

SD(2

3%m

ales

and

10%

fem

ales

)

24%

<2

SDN

AN

A25

%<

2SD

18%

<4

SDN

A6%

<2.

5SD

NA

Ane

mia

14%

(9%

mal

esan

d19

%fe

mal

esN

AN

AN

AN

AN

AN

AN

AN

A

Evi

denc

eof

port

alH

TN

a44

%(4

1%m

ales

and

47%

fem

ales

)38

%sp

leno

meg

aly

15%

37%

23%

(8/3

5)46

%39

%47

%50

% (hep

atom

egal

y)35

%15

%es

opha

geal

vari

ces

2%as

cite

sSu

rviv

alra

te1

year

85%

1ye

ar79

%5

year

s84

%1

year

79%

1ye

ar87

%1

year

89%

10ye

ars

51%

10ye

ars

82%

5ye

ars

75%

9ye

ars

80%

NA

3ye

ars

88%

1ye

ar91

%15

year

s46

%1

year

19%

1ye

ar88

%D

eath

rate

inth

efir

stye

arof

life

15%

8%13

%26

%9%

9%24

%22

%12

%

Abb

reva

tion

sar

e:G

FR

,glo

mer

ular

filtr

atio

nra

te;C

RF,

chro

nic

rena

lfai

lure

;ESR

D,e

nd-s

tage

rena

ldis

ease

;SD

,sta

ndar

dde

viat

ion;

NA

,not

avai

labl

e;B

P,bl

ood

pres

sure

;HT

N,h

yper

tens

ion.

aB

ased

onso

nogr

aphi

cev

iden

ceof

hepa

tom

egal

y,sp

leno

meg

aly,

and

dire

ctio

nalr

ever

salo

fpor

talv

ein

flow

,or

clin

ical

,rad

iolo

gic

oren

dosc

opic

evid

ence

ofes

opha

geal

vari

ces

oras

cite

s.M

anif

esta

tion

ofcl

inic

alsi

gns

ofco

ngen

ital

hepa

tic

fibro

sis

was

posi

tive

lyco

rrel

ated

wit

hag

e.In

87%

incr

ease

dec

hoge

nici

tyof

the

liver

has

been

repo

rted

(89%

mal

ean

d84

%fe

mal

e).

Cys

tic

chan

ges

ofth

eliv

erpr

obab

lyre

pres

enti

ngC

arol

i’sdi

seas

ew

ith

dila

ted

larg

erin

trah

epat

icbi

ledu

cts

have

been

note

din

27in

divi

dual

seq

ualin

g16

%of

the

tota

lcoh

ort(

17%

mal

ean

d15

%fe

mal

e).W

ithi

nth

issu

rvey

only

two

boys

exhi

bite

dim

pair

edhe

pato

cellu

lar

func

tion

(1%

)un

ders

cori

ngth

atliv

erfu

ncti

onis

usua

llyre

tain

edin

(aut

osom

al-r

eces

sive

poly

cyst

icki

dney

dise

ase)

.In

six

pati

ents

(fou

rm

ale

and

two

fem

ale)

liver

tran

spla

ntat

ion

(LT

X)

was

perf

orm

ed(m

ean

age

13.8

year

s).I

nfo

urca

ses

itw

asdo

nein

para

llelw

ith

rena

ltra

nspl

anta

tion

(com

bine

dliv

er-k

idne

ytr

ansp

lant

aion

).


is linked to PKHD1 in the vast majority of ARPKD fam-ilies, locus heterogeneity has to be considered in caseswithout PKHD1 mutation.

Genotype-phenotype correlations

This study corroborates earlier genotype-phenotypecorrelations for the type of PKHD1 mutation [21]. Allpatients carrying two truncating mutations displayed asevere phenotype with peri-/neonatal demise, whereaspatients surviving the neonatal period bear at least onemissense mutation. Although the converse did not applyand some missense changes are obviously as devastat-ing as truncating mutations, missense changes are morefrequently observed among patients with a milder clini-cal course, while chain-terminating mutations are morecommonly associated with a severe phenotype.

Loss of function probably explains the uniformly earlydemise of patients carrying two truncating alleles. In con-trast, missense mutations may have more variable effectson protein’s function. Some might result in hypomor-phic alleles with reduced function allowing for a clini-cally milder course, whereas others may represent loss-of-function variants. As depicted by discordant siblings,phenotypes cannot be simply explained on the basis ofthe genotype, but likely depend on the background ofother genes [50, 51], epigenetic factors (e.g., alternativesplicing [32, 49, 52, 53]), and environmental influencesas well. Such modifiers will probably have their great-est impact on the phenotype in the setting of hypomor-phic missense changes and may explain, at least in part,the highly variable clinical course resulting from missensemutations, whereas they are less likely to be relevant inloss-of-function alleles.

Phenotypes may further be influenced by the locationand character of amino acid substitutions. It might thusbe reasonable to categorize PKHD1 missense mutationsinto severe and moderate/mild changes as done for var-ious other disorders [54, 55]. However, multiple allelismand the high rate of different compound heterozygoteshamper genotype-phenotype correlations. In a simplifiedapproach to classify mutations, termination-type muta-tions represent loss-of-function alleles with a uniform ef-fect on phenotype. Thus, one would expect the positionand character of the missense change to determine theclinical course in compound heterozygotes with one trun-cating and one missense mutation. As a matter of course,conclusions should be dealt tentatively given the still lim-ited dataset, even if PKHD1 missense mutations haveonly been classified in case of recurrence in at least twounrelated families either homozygously or in compoundwith a truncating mutation on the other parental allele(Table 6).

We tested the hypothesis that differences in the clin-ical course observed among the study cohort resulted

from the type of mutations. We focused on those patientsin whom we had identified two putative mutations andsubdivided this cohort into patients carrying missense al-terations on both parental alleles and a group charac-terized by the presence of a truncating change. The lat-ter cohort was supplemented by patients in whom onlya chain-terminating mutation is known while the sec-ond pathogenic allele remains to be identified. It is mostlikely that the unidentified mutation in these patients isan amino acid substitution given the strong correlationthat survival past the neonatal period invariably requiresthe presence of at least one missense mutation.

We first compared the nature of mutations among pa-tients who required renal replacement therapy. One mayhypothesize that ESRD occurred earlier in individualswith a chain-terminating change than in patients withtwo amino acid substitution mutations. In fact, the meanage when renal replacement therapy was required wasconsiderably lower in the former group (7.2 versus 10.1years). However, the small number of patients included ineach subgroup does not allow meaningful statistical com-parisons. The same applies for correlations regarding thedevelopment of renal function depicted in Figure 3. Themean age at which patients with a truncating mutationon one parental allele reached a serum creatinine valueof about 200 lmol/L was earlier than in patients carryingtwo missense changes (4.1 versus 7.2 years), however, theclinical events examined were too few in number.

We further compared the mutation spectrum of pa-tients requiring renal replacement therapy (N = 38) withthose of individuals with a minimum age of 15 years anda serum creatinine <200 lmol/L (N = 20). There was aslight overpresentation of missense changes in the mildlyaffected cohort (71% versus 80%) (NS). We checkedthe sites of missense mutations to look for protein re-gions associated with a specific phenotype. One muta-tion, Y1838C (exon 34), was homozygously observed inthree Turkish patients with early ESRD. This suggests thisportion of PKHD1 to be particularly important for renalintegrity. This hypothesis is further corroborated by RTXpatient ID149 who carries the adjacent change W1834Cin compound with a nonsense mutation.

Furthermore, we questioned if the mutational spec-trum among patients with a predominant liver pheno-type [early esophageal variceal bleeding or ascites and(nearly) normal renal function] (N = 15) differed fromthat of patients without clinical hepatobiliary manifesta-tion at age 15 (N = 17). While missense changes werecomparably distributed among the two groups (70% ver-sus 78%) (NS), mutations of patients with a predominantliver phenotype clustered around amino acids 2831-2840(exon 54) and 3051-3209 (exon 58). However, given thelimited number of patients further functional and muta-tional studies in PKHD1 will have to be awaited beforeprofound conclusions can be drawn from these data.


While the majority of sibships display comparable clin-ical courses, about 20% of ARPKD multiplex pedigreesexhibit gross intrafamilial phenotypic variability (own un-published data of more than 100 multiplex pedigrees)[56]. The even higher proportion of 20 out of 48 sibships(42%) in the current survey is biased due to the studydesign with at least one neonatal survivor per family.Adjusted for differing family sizes the risk for perinataldemise of a further affected child was 37% (22 perinatallydeceased children from a total of 59 patients excludingthe moderately affected index cases). For genetic coun-seling this rate is alarming given that our study cohort isrepresentative for the spectrum of patients followed bydepartments of pediatric nephrology. Of course, pheno-type categorization into severe and moderate is a sim-plified view given the considerably better prognosis forpatients surviving the most critical neonatal period [7].For instance, survival of an individual might depend onavailable intensive care facilities or birth order whetherthe parents are aware of ARPKD risk. In 14 of the over-all 20 sibships with gross phenotypic variability includedin the present study the peri-/neonatally deceased sib-ling was the first born affected child. Nevertheless, somealertness should be warranted in predicting the clinicaloutcome of a further affected child.

ACKNOWLEDGMENTS

The authors would like to thank the patients and families involvedin these studies for their cooperation. The technical assistance ofEdith von Heel, Edith Bunger, Elvira Golz-Staggemeyer, and ClaudiaCanales is gratefully acknowledged. This work was supported by theDeutsche Forschungsgemeinschaft (DFG), the German-Israeli Foun-dation (GIF), and the START program of the medical faculty of AachenUniversity.

Reprint requests to Carsten Bergmann, M.D., or Klaus Zerres, M.D.,Department of Human Genetics, Aachen University, Pauwelsstraße 30,D-52074 Aachen, Germany.E-mail: [email protected] or [email protected]

REFERENCES

1. GUAY-WOODFORD LM: Autosomal recessive polycystic kidney dis-ease: clinical and genetic profiles, in Polycystic kidney disease, editedbyWatson ML, Torres VE, Oxford, Oxford University Press, 1996,pp 237–266

2. ZERRES K, RUDNIK-SCHONEBORN S, STEINKAMM C, et al: Autosomalrecessive polycystic kidney disease. J Mol Med 76:303–309, 1998

3. ZERRES K, RUDNIK-SCHONEBORN S, SENDEREK J, et al: Autosomalrecessive polycystic kidney disease (ARPKD). J Nephrol 16:453–458, 2003

4. ZERRES K, HANSMANN M, MALLMANN R, GEMBRUCH U: Autosomalrecessive polycystic kidney disease. Problems of prenatal diagnosis.Prenat Diagn 8:215–229, 1988

5. ZERRES K, MUCHER G, BECKER J, et al: Prenatal diagnosis of au-tosomal recessive polycystic kidney disease (ARPKD): Moleculargenetics, clinical experience, and fetal morphology. Am J Med Genet76:137–144, 1998

6. REUSS A, WLADIMIROFF JW, STEWART PA, NIERMEIJER MF: Prenataldiagnosis by ultrasound in pregnancies at risk for autosomal re-cessive polycystic kidney disease. Ultrasound Med Biol 16:355–359,1990

7. GUAY-WOODFORD LM, DESMOND RA: Autosomal recessive poly-cystic kidney disease: The clinical experience in North America.Pediatrics 111:1072–1080, 2003

8. ROY S, DILLON MJ, TROMPETER RS, BARRATT TM: Autosomal re-cessive polycystic kidney disease: Long-term outcome of neonatalsurvivors. Pediatr Nephrol 11:302–306, 1997

9. WARD CJ, HOGAN MC, ROSSETTI S, et al: The gene mutated in auto-somal recessive polycystic kidney disease encodes a large, receptor-like protein. Nat Genet 30:259–269, 2002

10. ONUCHIC LF, FURU L, NAGASAWA Y, et al: PKHD1, the polycys-tic kidney and hepatic disease 1 gene, encodes a novel large pro-tein containing multiple immunoglobulin-like plexin-transcription-factor domains and parallel beta-helix 1 repeats. Am J Hum Genet70:1305–1317, 2002

11. NAGASAWA Y, MATTHIESEN S, ONUCHIC L, et al: Identification andcharacterization of Pkhd1, the mouse orthologue of the humanARPKD gene. J Am Soc Nephrol 13:2246–2258, 2002

12. MASYUK TV, HUANG BQ, WARD CJ, et al: Defects in cholangiocytefibrocystin expression and ciliary structure in the PCK rat. Gas-troenterology 125:1303–1310, 2003

13. WARD CJ, YUAN D, MASYUK TV, et al: Cellular and subcellular local-ization of the ARPKD protein; fibrocystin is expressed on primarycilia. Hum Mol Genet 12:2703–2710, 2003

14. ZHANG MZ, MAI W, LI C, et al: PKHD1 protein encoded by thegene for autosomal recessive polycystic kidney disease associateswith basal bodies and primary cilia in renal epithelial cells. ProcNatl Acad Sci USA 101:2311–2316, 2004

15. WANG S, LUO Y, WILSON PD, et al: The autosomal recessive poly-cystic kidney disease protein is localized to primary cilia, with con-centration in the basal body area. J Am Soc Nephrol 15:592–602,2004

16. WATNICK T, GERMINO G: From cilia to cyst. Nat Genet 34:355–356,2003

17. OLBRICH H, FLIEGAUF M, HOEFELE J, et al: Mutations in a novel gene,NPHP3, cause adolescent nephronophthisis, tapeto-retinal degen-eration and hepatic fibrosis. Nat Genet 34:455–459, 2003

18. OTTO EA, SCHERMER B, OBARA T, et al: Mutations in INVS encodinginversin cause nephronophthisis type 2, linking renal cystic diseaseto the function of primary cilia and left-right axis determination.Nat Genet 34:413–420, 2003

19. NAULI SM, ALENGHAT FJ, LUO Y, et al: Polycystins 1 and 2 mediatemechanosensation in the primary cilium of kidney cells. Nat Genet33:129–137, 2003

20. SUN Z, AMSTERDAM A, PAZOUR GJ, et al: A genetic screen in ze-brafish identifies cilia genes as a principal cause of cystic kidney.Development 2004 [electronic publication ahead of print]

21. BERGMANN C, SENDEREK J, SEDLACEK B, et al: Spectrum of muta-tions in the gene for autosomal recessive polycystic kidney disease(ARPKD/PKHD1). J Am Soc Nephrol 13:76–89, 2003

22. BLICKMAN JG, BRAMSON RT, HERRIN JT: Autosomal recessive poly-cystic kidney disease: Long-term sonographic findings in patientssurviving the neonatal period. Am J Roentgenol 164:1247–1250,1995

23. GAREL L: Sonography of renal cystic disease and dysplasia in infantsand children, in Pediatric Nephrology, edited byBrodehl J, EhrichJHH, Berlin, Springer, 1984, pp 359–362

24. ZERRES K, RUDNIK-SCHONEBORN S, DEGET F, et al: Autosomal reces-sive polycystic kidney disease in 115 children: Clinical presentation,course and influence of gender. Acta Paediatr 85:437–445, 1996

25. WASSNER SJ: Conservative management of chronic renal insuffi-ciency, in Pediatric Nephrology (3rd ed), edited byHolliday MA,Barrat TM, Vernier RL, Baltimore, Williams and Wilkins, 1994, pp1314–1337

26. SCHWARTZ GJ, HAYCOCK GB, EDELMANN CM Jr, SPITZER A: A simpleestimate of glomerular filtration rate in children derived from bodylength and plasma creatinine. Pediatrics 58:259–263, 1976

27. SCHWARTZ GJ, FELD LG, LANGFORD DJ: A simple estimate ofglomerular filtration rate in full-term infants during the first yearof life. J Pediatr 104:849–854, 1984

28. DINKEL E, ERTEL M, DITTRICH M, et al: Kidney size in childhood.Sonographical growth charts for kidney length and volume. PediatrRadiol 15:38–43, 1985

29. MUCHER G, BECKER J, KNAPP M, et al: Fine mapping of the autosomal


recessive polycystic kidney disease locus (PKHD1) and the genesMUT, RDS, CSNK2 beta, and GSTA1 at 6p21.1-p12. Genomics48:40–45, 1998

30. ONUCHIC LF, MRUG M, HOU X, et al: Refinement of the autosomalrecessive polycystic kidney disease (PKHD1) interval and exclusionof an EF hand-containing gene as PKHD1 candidate gene. Am JMed Genet 110:346–352, 2002

31. BERGMANN C, SENDEREK J, SCHNEIDER F, et al: PKHD1 mutationsin families requesting prenatal diagnosis for autosomal recessivepolycystic kidney disease (ARPKD). Hum Mutat 23:487–495, 2004

32. MODREK B, LEE C: A genomic view of alternative splicing. Nat Genet30:13–19, 2002

33. ROSSETTI S, TORRA R, COTO E, et al: A complete mutationscreen of PKHD1 in autosomal-recessive polycystic kidney disease(ARPKD) pedigrees. Kidney Int 64:391–403, 2003

34. FURU L, ONUCHIC LF, GHARAVI A, et al: Milder presentation ofrecessive polycystic kidney disease requires presence of amino acidsubstitution mutations. J Am Soc Nephrol 14:2004–2014, 2003

35. BERGMANN C, SENDEREK J, KUPPER F, et al: PKHD1 mutations in au-tosomal recessive polycystic kidney disease (ARPKD). Hum Mutat23:453–463, 2004

36. Cooper DN, Krawczak M (editors): Human Gene Mutation, Oxford,BIOS Scientific Publishers Limited, 1993

37. CAPISONDA R, PHAN V, TRAUBUCI J, et al: Autosomal recessive poly-cystic kidney disease: outcomes from a single-center experience.Pediatr Nephrol 18:119–126, 2003

38. GAGNADOUX MF, HABIB R, LEVY M, et al: Cystic renal diseases inchildren. Adv Nephrol Necker Hosp 18:33–57, 1989

39. KAPLAN BS, FAY J, SHAH V, et al: Autosomal recessive polycystickidney disease. Pediatr Nephrol 3:43–49, 1989

40. KAARIAINEN H, KOSKIMIES O, NORIO R: Dominant and recessivepolycystic kidney disease in children: Evaluation of clinical featuresand laboratory data. Pediatr Nephrol 2:296–302, 1988

41. COLE BR, CONLEY SB, STAPLETON FB: Polycystic kidney disease inthe first year of life. J Pediatr 111:693–699, 1987

42. NICOLAU C, TORRA R, BADENAS C, et al: Sonographic pattern of re-cessive polycystic kidney disease in young adults. Differences fromthe dominant form. Nephrol Dial Transplant 15:1373–1378, 2000

43. AVNI FE, GUISSARD G, HALL M, et al: Hereditary polycystic kid-ney diseases in children: Changing sonographic patterns throughchildhood. Pediatr Radiol 32:169–174, 2002

44. NAHM AM, HENRIQUEZ DE, RITZ E: Renal cystic disease (ADPKDand ARPKD). Nephrol Dial Transplant 17:311–314, 2002

45. WAGNER T, STOPPA-LYONNET D, FLEISCHMANN E, et al: Denaturinghigh-performance liquid chromatography detects reliably BRCA1and BRCA2 mutations. Genomics 15:369–376, 1999

46. ROSSETTI S, CHAUVEAU D, WALKER D, et al: A complete mutationscreen of the ADPKD genes by DHPLC. Kidney Int 61:1588–1599,2002

47. BRAUTIGAM S, KUJAT A, KIRST P, et al: DHPLC mutation analysisof phenylketonuria. Mol Genet Metab 78:205–210, 2003

48. KING K, FLINTER FA, NIHALANI V, GREEN PM: Unusual deep intronicmutations in the COL4A5 gene cause X linked Alport syndrome.Hum Genet 111:548–554, 2002

49. MODREK B, LEE CJ. Alternative splicing in the human, mouse and ratgenomes is associated with an increased frequency of exon creationand/or loss. Nat Genet 34:177–180, 2003

50. GUAY-WOODFORD LM, WRIGHT CJ, WALZ G, CHURCHILL GA. Quan-titative trait loci modulate renal cystic disease severity in the mousebpk model. J Am Soc Nephrol 11:1253–60, 2000

51. SOMMARDAHL C, COTTRELL M, WILKINSON JE, et al: Phenotypic vari-ations of orpk mutation and chromosomal localization of modifiersinfluencing kidney phenotype. Physiol Genomics 21:127–134, 2001

52. NISSIM-RAFINIA M, KEREM B. Splicing regulation as a potential ge-netic modifier. Trends Genet 18:123–127, 2002

53. CACERES JF, KORNBLIHTT AR. Alternative splicing: multiple con-trol mechanisms and involvement in human disease. Trends Genet18:186–193, 2002

54. WHITE PC, SPEISER PW. Congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Endocr Rev 21:245–291, 2000

55. SALVATORE F, SCUDIERO O, CASTALDO G. Genotype-phenotype cor-relation in cystic fibrosis: the role of modifier genes. Am J Med Genet111:88–95, 2002

56. DEGET F, RUDNIK-SCHONEBORN S, ZERRES K. Course of autosomalrecessive polycystic kidney disease (ARPKD) in siblings: a clinicalcomparison of 20 sibships. Clin Genet 47:248–53, 1995

Clinical consequences of PKHD1 mutations in 164 patients with autosomal-recessive polycystic kidney disease (ARPKD)

Documents