Kidney International, Vol. 56 (1999), pp. 1324–1333

Polycystin-1 expression in PKD1, early-onset PKD1,and TSC2/PKD1 cystic tissue

ALBERT C.M. ONG, PETER C. HARRIS, DAVID R. DAVIES, LYNN PRITCHARD, SANDRO ROSSETTI,SIMON BIDDOLPH, DAVID J.T. VAUX, NICOLA MIGONE, and CHRISTOPHER J. WARD

MRC Molecular Haematology Unit, Institute of Molecular Medicine, University of Oxford, Oxford, Department of CellularPathology and Department of Paediatric Pathology, John Radcliffe Hospital, Headington; Sir William Dunn School ofPathology, University of Oxford, Oxford, United Kingdom; and Dipartimento di Genetica, Biologia e Biochimica,Universita di Torino and CNR-CIOS, Torino, Italy

Polycystin-1 expression in PKD1, early-onset PKD1, and tations in one of three genes: PKD1, the major geneTSC2/PKD1 cystic tissue. accounting for 85 to 90% of cases, and PKD2 or PKD3,

Background. The mutational mechanism responsible for accounting for the remaining 10 to 15%. The identifica-cyst formation in polycystic kidney disease 1 gene (PKD1)tion of PKD1 in 1994 [1] and PKD2 in 1996 [2] hasremains controversial, with data indicating a two-hit mecha-provided new tools to study the biology of cyst formationnism, but also evidence of polycystin-1 expression in cystic

tissue. in this disease. The PKD1 protein, polycystin-1, isMethods. To investigate this apparent paradox, we analyzed thought to be a membrane glycoprotein (approximately

polycystin-1 expression in cystic renal or liver tissue from 10 460 kDa) involved in cell adhesion [3, 4], whereas thepatients with truncating PKD1 mutations (including one early-PKD2 protein has homology to a family of voltage-gatedonset case) and 2 patients with severe disease associated withsodium/calcium channels [2]. Sequence similarities alsocontiguous deletions of TSC2 and PKD1, using monoclonal

antibodies (mAbs) to both extreme N-(7e12) and C-terminal indicate that polycystin-1 may play a role in ion transport(PKS-A) regions of the protein. Truncation of the C-terminal [5]. Recent studies showing that these two proteins haveepitope from the putative mutant proteins in each case allowed a similar cellular distribution [6] and can interact in ex-exclusive assessment of the nontruncated protein with PKS-A.

perimental systems [7, 8] are consistent with the pheno-Results. In adult PKD1 tissue, the majority of cysts (approxi-typic similarity of both conditions.mately 80%) showed polycystin-1 expression, although staining

was absent in a variable but significant minority (approximately In spite of the cloning of PKD1, the mutational mecha-20%), in spite of the normal expression of marker proteins. nism underlying cyst initiation remains controversial. Ev-Unlike adult PKD1, however, negative cysts were rarely found idence is accumulating that germline PKD1 mutationsin infantile PKD1 or TSC2/PKD1 deletion cases.

are probably inactivating. First, although relatively fewConclusions. If a two-hit mutational mechanism is opera-mutations have been described, those characterized aretional, these results suggest that the majority of somatic muta-

tions in adult PKD1 are likely to be missense changes. The mainly predicted to terminate the protein prematurelylow level of polycystin-1–negative cysts in the three “early- because of the introduction of a stop or frameshiftingonset” cases, however, suggests that a somatic PKD1 mutation change [9–11]. Second, an analysis of a group of patientsmay not always be required for cyst formation.

with deletions disrupting PKD1 and the adjacent tuber-ous sclerosis gene TSC2 [12, 13] shows that PKD1 isoften completely deleted and a phenotype of tuberousAutosomal dominant polycystic kidney diseasesclerosis with severe, early-onset polycystic kidney dis-(ADPKD) is the most common genetic disease affectingease results. Although the contribution that disruptionthe kidney and accounts for up to 10% of patients onof a single TSC2 allele makes to the cystic phenotype isrenal replacement therapy. ADPKD can result from mu-unclear, these cases show that a null PKD1 allele canresult in cyst formation. Finally, a severe cystic pheno-type manifesting in late fetal life has been observed inKey words: polycystic kidney disease, ADPKD, TSC2, gene mutations,

cystogenesis. the Pkd1del34/del34 knockout mouse [14], indicating that cystformation is associated with the complete absence ofReceived for publication December 15, 1998polycystin-1.and in revised form March 12, 1999

Accepted for publication April 28, 1999 Further insight into the mutational mechanism inPKD1 has also come from studies that show that epithe- 1999 by the International Society of Nephrology

1324

Ong et al: Polycystin-1 expression in PKD1 cystic tissue 1325

lial cells derived from single ADPKD cysts are clonal in size the N-terminal mAb, 7e12, 471 bp of PKD1 (281-origin [15] and that between 17 and 24% of cysts exhibit 751nt) was cloned into the pET15b expression vectorloss of heterozygosity (LOH) at PKD1 [15, 16]. Recently, to generate a fusion protein. Hybridomas were initiallyevidence of inactivating point mutations in liver cystic screened against an Fc chimeric protein containing thetissue has also been described [17]. These results suggest same region of the protein (synthesized in COS-1 cells)that an inactivating germline mutation only results in by enzyme-linked immunosorbent assay (ELISA). Posi-cyst formation once the normal allele is disrupted by tive clones were then screened for their ability to detectsomatic mutation. A “two-hit” model is an attractive Fc chimeric proteins either containing the 7e12 epitope,explanation for the focal nature of cyst formation and LRR (212-802nt) and Nter (90-2747nt), or excluding thatthe phenotypic variability of cystic disease in PKD1. region, lectin (1396-1823nt) by immunoblotting.Additional support for this hypothesis has come from The 7e12 epitope is encoded by a region of the PKD1mice carrying a targeted disruption of PKD2 [18]. Immu- gene duplicated in the homologous genes (HGs) [1].nohistological analysis shows that unlike the surrounding Recent sequence data of two HGs (AC002039), how-normal noncystic tubules, cysts formed in Pkd21/2 ani- ever, indicate that these loci could only encode smallmals do not express polycystin-2, indicating that PKD2 truncated products (approximately 50 kDa; unpublishedcysts only develop in the absence of the protein. In addi- observation), and Western analysis with 7e12 has failedtion, a hypermutable allele with tandemly inserted exon 1 to reveal any such proteins. It therefore seems likely thatsequences (WS-25) was found to develop cysts at a higher any such HG products are unstable and that 7e12 israte than the corresponding heterozygote (Pkd2WS25/2 de- polycystin-1 specific.velops more cysts than Pkd21/2), suggesting that cysto-genesis is associated with somatic mutation. Recently, Plasma membrane preparation from tissuesmore detailed descriptions of mice carrying a targeted Plasma membrane fractions of normal human adultdisruption of Pkd1 have shown that a variable number kidney cortex (obtained at nephrectomy) were preparedof cysts are found in older Pkd11/2 mice, a finding similar by a modification of established methods [26]. Pieces ofto that seen in Pkd21/2 mice. Interestingly, however, in tissue were first thawed, cut into fine pieces, and homoge-this case, not all cysts in Pkd11/2 mice lacked polycystin-1 nized in buffer 1 (0.01 m Tris-HCl, pH 7.4, 0.25 m sucrose,expression, as assessed with a monoclonal antibody 0.2 mm CaCl2) using a polytron homogenizer. The ho-(mAb) to PKD1 (PKS-A) [19]. mogenate was diluted with ice-cold buffer 2 [0.01 m Tris-

At first glance, a two-hit mechanism appears incom- HCl, pH 7.4, 0.025 m sucrose, 1 mm ethylenediaminetet-patible with results obtained from studies of polycystin-1 raacetic acid (EDTA)] and filtered through surgicalexpression in ADPKD cystic renal tissue [20–24]. All of

gauze, and the suspension was centrifuged at 2000 g. Thethe published immunohistological studies (using differ-resulting supernatant was then loaded onto a 35% (wt/ent antibodies) have found continued polycystin-1 ex-vol) sucrose (buffer 3) cushion in SW28 tubes and centri-pression—often at a high level—in cystic epithelia [25],fuged again at 30,000 g at 48C for 30 minutes. The inter-although in all but one case [20] the germline mutationface was collected, diluted with buffer 4 (0.01 m Tris-of the tissue examined has not been identified.HCl, pH 7.4, 0.25 m sucrose), and centrifuged finally atIn an effort to resolve this apparent conflict, we have100,000 g at 48C for 45 minutes. The final pellet, a plasmaidentified and examined cystic renal tissue obtained frommembrane-enriched fraction, was resuspended in 0.5 tonine patients with defined truncating PKD1 mutations,1.0 ml of buffer 4 or 1% sodium dodecyl sulfate (SDS)eight with adult-onset PKD1, one with early-onset PKD1,and aliquots frozen at 2708C until analysis. Proteaseand two patients with contiguous deletions of TSC2 andinhibitors (aprotinin 2 mg/ml, benzamide 1 mm, EDTAPKD1. Cystic PKD1 liver was obtained from one patient1 mm, phenylmethylsulfonyl fluoride (PMSF) 1 mm,with a defined PKD1 mutation. In each case, the germ-pepstatin 4 mg/ml, leupeptin 1 mg/ml, and pefablockline mutation truncates the C-terminal region of the mu-1 mg/ml; Boehringer Mannheim, Germany) were addedtant protein; hence, by using a mAb raised to a C-terminalto the buffers at all stages of the procedure.epitope of polycystin-1, we reasoned that the presence

or absence of the PKD1 protein encoded by the nontrun-Western blottingcated allele could be assayed in individual cysts.

Cell pellets (2 3 107 cells) were harvested by scrapingin ice-cold PBS in the presence of protease inhibitors.

METHODS Total cell lysates were prepared from these pellets underProduction and characterization of monoclonal denaturing (boiling 1% SDS, 10 mm Tris-HCl, pH 7.4)antibodies to epitopes in the duplicated and unique or native (RIPA buffer: 150 mm NaCl, 1% NP-40, 0.5%regions of polycystin-1 DOC, 0.1% SDS, 50 mm Tris-HCl, pH 7.5) conditions.

The protein content of each sample was measured usingThe synthesis and screening of the C-terminal mAbPKS-A have been previously described [20]. To synthe- a microtiter Lowry assay (Bio-Rad, Richmond, CA,

Ong et al: Polycystin-1 expression in PKD1 cystic tissue1326

USA). Each sample (total cell lysate or plasma mem-brane fraction) was resuspended in an equal volume of2 3 Laemmli sample buffer (100 mm Tris-HCl, pH 6.8,4% SDS, 20% glycerol, 0.1% bromophenol blue, 5%b-mercaptoethanol) and incubated at 378C for 30 min-utes or 958C for five minutes prior to loading. SDS-polyacrylamide gel electrophoresis was performed on7.5% separating gels using a Bio-Rad Mini Protean IIapparatus using molecular weight standards (Rainbow;Amersham, Bucks, UK) with a range of 14.3 to 220kDa. Separated proteins were then transferred at 30 Vovernight or 100 V for one hour onto a polyvinylidinedifluoride (PVDF) membrane (Millipore, Bedford, MA,USA). In some experiments, gels were stained with Coo-masie, either in parallel to assess migration or after trans-fer to assess transfer efficiency.

Filters were blocked with 5 to 10% nonfat milk inTris-buffered saline (TBS, pH 7.6) overnight, probedwith the primary antibody for one hour, and a horserad-ish peroxidase (HRP)-conjugated secondary antibody(goat antimouse or goat antirabbit) for one hour. Filterswere washed extensively with TBS (with added Tween0.05%) after detection with each antibody. Bound sec-ondary antibody was detected by enhanced chemilumi-nescence (ECL; Boehringer) using Kodak film.

Clinical details of PKD1 and TSC2/PKD1 patients

Renal tissue was obtained from eight PKD1 patientswith typical adult-onset disease and cystic liver from 1PKD1 patient, in whom truncating PKD1 mutationshave been identified (Fig. 1 and Table 1). The mutationsin OX18, OX40, and OX1330 have been described pre-viously [1, 10, 20]. Full characterization of the othermutations is described elsewhere (Rossetti, manuscriptin preparation).

A case of early-onset PKD1 was also identified inOxford (OX1005), the fetus being born at term (40 weekfemale) but dying two hours after birth with massivelyenlarged kidneys (courtesy of Dr. S. Gould, John Rad-

Fig. 1. Diagram of the polycystic kidney disease-1 (PKD1) protein,cliffe Hospital, Oxford, UK). We have subsequentlypolycystin-1, showing the position of defined PKD1 mutations in 12

identified the mutation in this PKD1 family (Fig. 1 and patients, in whom renal or liver tissue was available for analysis. (Openarrows) In these patients, the regions C-terminal of the arrow areTable 1).removed because of DNA deletion: TSC2/PKD1 (16 and III1) or typicalRenal tissue from two patients with known deletionsPKD1 (OX1330). The III1 breakpoint has been localized to the area

of TSC2 and PKD1 were also obtained: 16 (courtesy of bracketed. (Filled arrows) The point at which polycystin-1 is predictedto be terminated in typical PKD1 patients; OX1005 had early-onsetProfessor M. Barrett and Professor R.A. Risdon, Insti-PKD1. The monoclonal antibodies (7e12, PKS-A) were raised to thetute of Child Health, London, UK) and III1 (courtesyregions of the protein highlighted in gray.

of Dr. R. Tardanico and Dr. F. Scolari, Spedali Civili,Brescia, Italy). The mutation in both of these patientshas been described previously [13, 27]. The position ofthe proximal breakpoint in III1 has been refined to lie Supplies, Aylesbury, UK) and a graded series of alcohol

solutions. For immunoperoxidase detection, endogenousbetween exons 18 to 24.peroxidase was blocked by a 30-minute incubation in a

Immunohistochemistry 0.5% hydrogen peroxide/methanol solution. Nonspecificbinding was blocked with 10% goat serum, and tissuesParaffin sections (5 mm) were cut onto Vectabond-

coated slides and deparaffinized through Citroclear (HD were incubated sequentially with primary and HRP-con-

Ong et al: Polycystin-1 expression in PKD1 cystic tissue 1327

Table 1. Polycystin-1 immunoreactivity in PKD1, TSC2/PKD1 and early-onset PKD1 kidneys

1 Cysts 2 Cysts 6 CystsTotal

Patient Pedigree Mutation cysts N % N % N %

Adult onset PKD1 kidney OX161 P18 E1537X 19 13 2 4OX31 P314 4784delG 19 12 3 4OX35 P9 6645del28 24 13 11 0OX3 P17 Q2243X 23 21 2 0OX4 P7 C2370X 33 25 7 1OX1330 P95 10708del5.5kb 18 17 0 1OX40 P187 Q3513X 33 21 7 5OX18 P225 12739delA 22 11 7 4Totals 191a 133 69.6 39 20.4 19 10

Adult onset PKD1 liver OX1162 P164 6356insG 24 14 58.1 5 20.8 5 20.8Adult onset PKD1 Totals 215 147 68.3 44 20.5 24 11.2Early onset PKD1 OX1005 P118 8126dup20 11b 11 100 0 0 0TSC-2/PKD1 III1 FAM1 del175kb 16 15 1 0

16 2895del26kb 10 10 0 0Totals 26 25 96.2 1 3.8 0

a 119 cysts (62.3%) of the cysts counted were of . 2 mm in diameterb No cysts . 2 mm were found

jugated secondary antibodies for 30 to 60 minutes. Bound Materialssecondary antibody was detected with diaminobenzid- A mAb to utrophin (MANCHO7) was the kind giftine (Sigma, St. Louis, MO, USA), and the nuclei were of Professor Glenn Morris (NE Wales Institute, Clwyd,counterstained with hematoxylin. Sections were then Wales, UK). Mouse mAbs employed were to the a1mounted in Depex mountant (BDH) after dehydration subunit of rabbit Na1-K1-ATPase, cross-reactive withthrough a series of ethanol and xylene solutions. As the human protein (Upstate Biotechnology, Lake Placid,negative controls tissue sections were incubated without NY, USA) and to Epithelial Membrane Antigen (EMA;the primary antibody or with an aliquot of antibody that Sigma, Poole, UK). TRITC-conjugated PNA was pur-had been preincubated overnight with 5 to 10 mg of the chased from Vector Labs (Burlingame, CA, USA).fusion protein containing the epitope to which it hadbeen raised.

RESULTSTo determine the pattern of staining in each patient,all cysts of more than 1 mm in diameter were counted Production and characterization of a monoclonalon all tissue available and were graded for staining inten- antibody to the LRR domain of polycystin-1sity and epithelial morphology in a single blinded fashion

To complement the existing mAb to the C terminusby an experienced renal pathologist (D.R.D.). A cyst(PKS-A) [20], a mAb to the flank-leucine rich repeat-(cystically dilated tubule or Bowman’s capsule) was de-flank region (24-180aa) was produced (Fig. 1). The speci-fined as an abnormal epithelial lined cavity with dimen-ficity of this antibody (7e12) was initially tested for itssions wider than 1 mm (that is, greater than five timesability to recognize Fc fusion proteins containing differ-the normal tubular diameter). For the purposes of quan-ent regions of polycystin-1 that either included or ex-titation, cysts were excluded if no epithelial lining wascluded the LRR domain by immunoblotting (Fig. 2A).observed or if the lumen appeared to communicate withSecond, we examined the affinity of 7e12 for nativetransitional epithelium. Cyst lining epithelium was de-polycystin-1 in a range of renal cell lines. The highestscribed morphologically as attenuated, bosselated, papil-level of expression was found in COS-1 cells, with alary, cuboidal, or columnar.band of approximately 400 kDa (similar to the predictedImmunofluorescence was performed as detailed forpolycystin-1 molecular weight) detected (Fig. 2A). Simi-the immunoperoxidase method except that a FITC-con-larly sized proteins were detected in membrane fractionsjugated goat antimouse secondary antibody (Cappel,of normal human kidney (Fig. 2B). The approximateDurham, NC, USA) was used. Where dual staining withsize of polycystin-1 was confirmed by detection of a simi-a fluorescent-conjugated lectin (for example, peanut ag-lar-sized endogenous protein, utrophin (400 kDa), inglutinin; PNA) and PKD1 mAb was performed, lectinCOS-1 cells by immunoblotting (data not shown).binding preceded the blocking step. Stained sections

Finally, to see if 7e12 would identify a novel band inwere mounted in Moviol and imaged using a CCD cam-Western analysis of cells containing a PKD1 mutation,era on an epifluorescence microscope with a mercuryfibroblasts from the 77 family, in which PKD1 is dis-lamp and appropriate filter sets for fluorescein and rho-

damine. rupted in exon 15 by a chromosomal translocation [1],

Ong et al: Polycystin-1 expression in PKD1 cystic tissue1328

Fig. 2. Characterization of the N-terminal polycystin-1 monoclonal antibody (mAb) 7e12 by immunoblotting. (A) Recombinant proteins (10 ng)containing regions of polycystin-1: C-type lectin (398-539aa; 32 kDa), leucine-rich repeats (LRR; 1-197aa; 35 kDa) and N-terminal region (Nter;1-845aa; 103 kDa), and COS-1 cell lysate (50 mg) detected with 7e12. 7e12 identified the approximately 400 kDa endogenous protein in COS-1cells (and other renal cell lines; data not shown) and recombinant molecules containing the LRR domain. The proteins detected in the LRR andNter samples are larger than predicted from the amino acid composition, probably because of glycosylation. A larger protein detected in the LRRsample may reflect oligomerization of the LRR Fc protein. (B) A doublet approximately 400 kDa polycystin-1 product is detected with 7e12 inplasma membrane-enriched fractions isolated from two normal kidneys (NK1 and NK2; 50 mg). The a1 subunit of Na1,K1-ATPase (100 kDa)was also strongly detected in these membrane fractions, showing enrichment for plasma membrane proteins (data not shown). (C) Cell lysates ofCOS-1 cells, skin fibroblasts of a normal control (Fib) and from two patients with a chromosomal translocation disrupting PKD1 (77-2 and 77-4[1]; 50 mg) detected with 7e12. The endogenous polycystin-1 is detected in COS-1, but not normal fibroblasts. A mutant polycystin-1 moleculeapproximately 300 kDa is detected in the translocation containing cells.

were examined (Fig. 2C). Although polycystin-1 is not Polycystin-1 expression in normal adult human kidneynormally detected in fibroblasts, the same novel sized Analysis of normal human adult kidney with PKS-Aband (approximately 300 kDa) was detected in two mem- (and 7e12) revealed polycystin-1 expression localizingbers of this family with the translocation, consistent with predominantly, although not exclusively, to a subset ofthe size predicted from the transcript (approximately 8 cortical tubules and medullary collecting ducts, whichkb) detected in these individuals [1]. It is not clear why are also PNA and EMA positive [6]. This distal pattern

of expression in the adult kidney is distinct from thea novel polycystin-1 molecule is expressed in these cellsmore generalized expression of polycystin-1 by bothwhen the normal protein is not. However, previous stud-proximal and distal nephron segments in fetal kidneyies have shown that this abnormal transcript is very sig-that we have previously reported [20].nificantly up-regulated [1] and may reflect the different

chromosomal environment, deletion of the normal 39 endPolycystin-1 expression assessed with PKS-A and 7e12

of the gene (which encodes the transmembrane region ofin adult PKD1 kidneys and liver with characterized

the protein) and/or the presence of a novel C-terminal truncating deletionsend. Nevertheless, this analysis clearly illustrates the af-

Previous analysis of ADPKD epithelia with the mAbfinity of 7e12 for polycystin-1. Preliminary attempts toPKS-A showed staining of cystic epithelia, including indetect the presence of other mutant polycystin-1 proteinsone patient with a known deletional mutation (OX1330),from PKD1 cystic tissue with 7e12 have thus far beenwhich removed the PKS-A epitope [20]. We have fol-

unsuccessful.lowed up this study by analyzing renal tissue from seven

PKS-A does not detect the native protein in COS-1 other patients with known truncating mutations: OX3,cells by Western blotting under all conditions attempted, OX4, OX18, OX31, OX35, OX40, and OX161 (Table 1but does detect a recombinant protein containing the and Fig. 1). Each mutation is predicted to generate aterminal 1225aa of polycystin-1 in transfected COS-1 mutant protein lacking the PKS-A epitope; thus, onlycells, as previously described [20]. The reason for this protein encoded by the other allele would be detecteddifference is unclear but may represent the lower affinity with this antibody (the mutation in OX18 removes onlyof PKS-A for native polycystin-1, as compared with 7e12 part of the region used to synthesize PKS-A; Fig. 1). A(and thus the inability to detect the much lower levels systematic analysis of these cases is shown in Table 1,normally seen), or be due to a conformation-restricted which illustrates that the majority of cysts (approxi-

mately 70%) show clear immunoreactivity for the PKD1recognition of the epitope.

Ong et al: Polycystin-1 expression in PKD1 cystic tissue 1329

protein (Figs. 3 and 4). Nevertheless, some cysts (approx- was rarely observed in these two cases (Fig. 4 andTable 1).imately 20%) were negative for polycystin-1: The distri-

bution of these negative cysts was nonuniform and variedPolycystin-1 expression in early-onset PKD1 kidneyfrom patient to patient (Table 1). No consistent correla-

A case of early-onset PKD1 with a characterized trun-tion with a specific epithelial cell morphology was de-cating mutation (OX1005; Fig. 1) was also analyzed. Oftected, although, in general, negative staining for poly-interest, both glomerular and tubular cysts were presentcystin-1 was more frequently observed in cysts with anin this patient, although no cysts of more than 2 mmattenuated cell lining; less commonly, some cysts with awere identified (Fig. 4). Again, unlike adult-onset PKD1,cuboidal or bosselated lining were also negative (Figs.no negative cysts could be identified (Fig. 4 and Table3 and 4). In a minority of cysts (approximately 10%),1). In four other cases of early-onset PKD1 (data notheterogeneous polycystin-1 expression was detectedshown), negative cysts were also rarely observed, but in(Fig. 3). In approximately half of these cysts, negativethese cases, the germline mutation could not be identi-staining in an otherwise positive cyst lining correlatedfied, so they have not been included in this analysis.with a region of the cyst where the epithelial morphology

of the cell lining changed abruptly to an attenuated form.In other cysts, positive cells could be observed in an DISCUSSIONotherwise negative cyst lining and vice versa. A similar This study of polycystin-1 expression in PKD1 cysticexpression pattern was found in the one PKD1 liver tissue was initially prompted by uncertainty about theavailable for analysis (OX1162), except that the propor- mutational mechanism in PKD1 and, in particular, thetion of heterogenous cysts detected was higher (approxi- apparent inconsistency between observations of LOH inmately 20%) and that of positive cysts lower (approxi- isolated renal cyst epithelia and immunohistochemicalmately 60%) than in the cases of cystic kidney (Fig. 4 studies of polycystin-1 in cystic tissue [25]. To addressand Table 1). specifically the question as to whether polycystin-1 pro-

To analyze the pattern of staining more precisely, sec- tein encoded by the “normal” PKD1 allele was presenttions were either dually stained with the lectin PNA and in individual cysts, we limited our analysis to PKD1 cystica polycystin-1 mAb or serial sections were stained with tissue obtained from patients in whom germline muta-a polycystin-1 mAb and a mAb to EMA. PNA and EMA tions truncating the C-terminal region of polycystin-1are segment-specific markers for distal cortical tubules had been identified. Using a mAb raised to the C-termi-and collecting ducts in normal human kidney [28, 29], nal region of polycystin-1, PKS-A, we found that overall,although EMA expression may be induced in injured 70% of cysts were positive in adult-onset cases, indicatingproximal tubules [30]. Not all cysts were EMA or PNA the presence of protein encoded by the “normal” allele.positive, confirming previous observations that cysts may These results are consistent with previous immunohisto-arise from different nephron segments [28]. Figure 3 chemical studies of cystic epithelia that showed thatA–C shows an example of a cyst lining that is not stained polycystin-1 immunoreactivity can be detected in thewith PKS-A but is PNA positive; normal tubules are majority of cysts [20–24], although in these cases, linkagestained with both. Heterogenous polycystin-1 expression to PKD1 was not always established, and moreover, thewas observed in small and large cysts. An example of a germline mutation was unknown.cyst that binds PNA uniformly but that shows differential In this study, we also observed a proportion of cystsexpression of polycystin-1 can be seen in Figure 3 G–I. (approximately 20%) that are negative for polycystin-1Cysts or cystic regions found to be negative with PKS-A when stained with PKS-A. The absence of polycystin-1were also negative with 7e12, confirming the absence in these cysts could be due to somatic mutations thatof the entire protein (Fig. 4). Staining of all cysts was delete or truncate the “normal” PKD1 allele, consistentabolished by preincubation of the polycystin-1 mAb with the results obtained from the analysis of isolated(PKS-A, 7e12) with the relevant fusion protein (data not cyst epithelia [15–17]. Interestingly, we have also foundshown). that cysts negative for PKS-A were also nearly always

negative with 7e12, a mAb generated to the N-terminusPolycystin-1 expression in TSC2/PKD1 cystic epithelia of polycystin-1. Because an N-terminal antibody should

A rare subgroup of patients providing a unique insight in theory detect mutant polycystin-1 proteins in mostinto the pathophysiology of cyst formation are cases with cases analyzed (Fig. 1), this result indicates that thesedeletions disrupting PKD1 and TSC2, resulting in early- proteins may become unstable when truncated (dis-onset PKD [13]. We have studied two such TSC2/PKD1 cussed later in this section).deletion cases in which much of the PKD1 gene is deleted No consistent relationship between epithelial mor-(Fig. 1 and Table 1; discussed in the Methods section). phology and polycystin-1 expression was observed, al-

though cysts with an attenuated cell lining were moreUnlike adult PKD1, negative polycystin-1 expression

Ong et al: Polycystin-1 expression in PKD1 cystic tissue 1331

Fig. 3. Dual fluorescent staining of PKD1 kidneys with a mAb to the C-terminus of PKD1, PKS-A (FITC, green), and the lectin PNA (TRITC,red). Renal tissue was obtained from three patients with mutations truncating the C-terminus of polycystin-1: OX3 (A-C), OX161 (D-F), andOX40 (G-I). Polycystin-1 staining is shown in (A, D, and G), PNA reactivity in (B, E, and H), and dual color images of each section in (C, F,and I). (A–C) Note the polycystin-1–negative cuboidal cyst lining (arrows) that is PNA positive in OX3. Some PNA and polycystin-1 positivetubules are also seen. (D–F) Note the expression of polycystin-1 in two adjoining large PNA positive cysts (arrows) in OX161. Both cysts havean attenuated epithelial lining. (G–I) In contrast to uniform PNA reactivity, heterogenous staining for polycystin-1 is seen in a finger-like projectionin OX40. These heterogenous cysts comprised approximately 10% of all cysts counted (magnification 3200).b

often negative, and those with a cuboidal lining were rupting the DNA-binding domain [31], and may, in addi-tion, have a dominant negative effect by binding to themore often positive. However, many examples of posi-

tive attenuated cysts and negative cuboidal cysts were normal protein [34]. Thus, it is possible that the highlevel of polycystin-1 we see in the majority of PKD1observed (Figs. 3 and 4). Polycystin-1 negativity could

not be explained entirely by “de-differentiation” because cysts may be due to an accumulation of inactive protein.In the light of this, it is of interest that using thenegative cysts often, although not always, express other

proteins such as EMA or PNA, which are expressed by N-terminal mAb 7e12, we were able to detect a mutantpolycystin-1 protein of predicted size in skin fibroblastsmature distal tubules and collecting ducts. Polycystin-1

expression was also noted in EMA- or PNA-negative of two individuals with PKD from a family in whichPKD1 is disrupted in exon 15 by a chromosome translo-cysts, implying that cysts arising from other nephron

segments also express polycystin-1. Thus, it is unlikely cation. Unfortunately, we do not know if this protein isexpressed in the kidney because no renal tissue or cellsthat polycystin-1–negative cysts have arisen from poly-

cystin-1–negative tubular segments. This is consistent were available from this family for further study. Thisis the first time a mutant polycystin-1 protein has beenwith previous studies showing that in ADPKD, renal

cysts may arise from proximal, distal, and other nephron demonstrated, and although interesting, it is probablethat this is a unique situation because the mutant PKD1segments [28].

Our results could be reconciled with a two-hit mecha- transcript is highly overexpressed compared with thelevels for the normal allele [1]. Moreover, in preliminarynism initiating cystogenesis if the majority of somatic

mutations are missense, with a variable but smaller pro- studies, we have not been able to detect mutantpolycystin-1 proteins in other adult-onset PKD1 patientsportion of stop, frame-shifting changes or deletions (in-

cluding LOH). Missense mutations could functionally with defined germline mutations.Like others, we have observed a proportion of cystsinactivate the “normal” polycystin-1 protein while allow-

ing its expression and detection. Although the majority (approximately 10%) with partial staining for poly-cystin-1. In half of these cysts, it was evident that aof germline PKD1 mutations that have been identified

so far are stop or frame-shifting changes (in part due to positive region abuts onto a negative region (Fig. 3G),often where the morphology of the cyst epithelium hadthe methods used to specifically detect such changes),

the spectrum of somatic changes may be quite different. also changed abruptly (for example, from columnar orbosselated to attenuated). The most likely explanationThe situation may be analogous to that found with muta-

tions of the tumor suppressor gene p53 in many tumors. for this is that two different cysts have fused. In othercysts, positive and negative cells were found adjacent toMost p53 mutations changes are missense [31] and are

associated with a high level of p53 when monitored by each other, a preliminary finding also reported by others(abstract; Droz et al, J Am Soc Nephrol 8:371, 1997). Itimmunohistochemistry [32] caused by the increased sta-

bility of the mutant protein [33]. Nevertheless, these is unlikely that this type of patchy staining representsLOH and may be due to changes in polycystin-1 expres-missense changes inactivate the protein, often by dis-

b

Fig. 4. Immunoperoxidase staining for polycystin-1 and EMA in PKD1 kidney; OX40 (A–F), PKD1 liver; OX1162 (G), TSC2/PKD1 kidney; 16(H); III1 (I) and early-onset PKD1; OX1005 (J–L). Sequential sections from OX40 stained with PKS-A (A and D), 7e12 (B and E), or EMA (Cand F) showing the presence (A and B) or absence (D and E) of polycystin-1 staining with both mAbs in EMA positive cysts. Note the cuboidalepithelium in both positive (A–C) and negative (D–F) cysts and the expression of polycystin-1 by both EMA-positive and -negative tubules inA–C. The liver cysts displayed from OX1162 (G) shows polycystin-1 expression (PKS-A) in several biliary microhamartoma and part of anattenuated negative cyst lining (arrows). Prominent expression of polycystin-1, as detected by PKS-A, in cystic and normal tubular epithelium ofboth patients with a TSC2/PKD1 deletion 16 (H) and III1 (I) was observed. An identical expression pattern was seen in both patients with 7e12(data not shown). Note the increased cytoplasmic to nuclear ratio typically seen in these cystic cells involving TSC2 disruption [13] (arrows) andthe area of papillary outgrowth in the cyst lining illustrated (I). In OX1005, both glomerular (J) and tubular cysts (K and L) express polycystin-1as detected by PKS-A (J and K) or 7e12 (L; magnification 3100 for all except 3200; I). Glomeruli (g), cysts (c), and tubules (t) are indicated.

Ong et al: Polycystin-1 expression in PKD1 cystic tissue1332

sion, perhaps dependent on the stage of cell division or PKD1, however, points to the possibility of a differentmutational mechanism.other intracellular factors. Similar changes have been

observed for p53 in tumors and may be more evident inthe hyperproliferative cells of cystic epithelia than in ACKNOWLEDGMENTSnormal tubules. This work was funded by the Medical Research Council, National

Kidney Research Fund, Telethon-Italy (S.R.) and the Wellcome Trust.A particularly interesting group of patients for a studyA portion of this work was presented at the 30th American Societyof the mutational mechanism are those with early-onsetof Nephrology meeting (November, 2–5, 1997, San Antonio, TX, USA)

PKD1. In these cases, the two-hit hypothesis appears and has been published in abstract form (abstract; Ong et al; J AmSoc Nephrol 8:378, 1997). We thank S. Gould, R. Doyonnes, M. Hollins-particularly difficult because of the many cysts, and hencehead, G.E. Morris, and H. Turley for their generous gifts of advicesecond hits required, in the few weeks of renal develop-and materials, and Professor Sir D.J. Weatherall for encouragement

ment in utero. Interestingly, the two groups with early- and support. A.C.M.O. is a National Kidney Research Fund SeniorFellow, and C.J.W. holds a Wellcome Career Development Fellowship.onset disease analyzed in this study had a very low num-

ber of polycystin-1–negative cysts, suggesting that thereReprint requests to Dr. Albert C.M. Ong, MRC Molecular Haematol-

may be an unusual disease mechanism in these cases ogy Unit, Institute of Molecular Medicine, Headington, Oxford OX39DS, United Kingdom.that does not require a second hit at PKD1. The firstE-mail: a.ong@Sheffield.ac.ukgroup with an early-onset disease also has disruption of

the adjacent TSC2 gene, and germline deletion of bothREFERENCESof these genes is associated with severe childhood onset

1. European Polycystic Kidney Disease Consortium: The polycys-PKD and cystic epithelia with a characteristic morphol-tic kidney disease 1 gene encodes a 14 kb transcript and lies withinogy [13]. Although it is possible that gross mutation ata duplicated region on chromosome 16. Cell 77:881–894, 1994

PKD1/TSC2 promotes a higher level of somatic muta- 2. Mochizuki T, Wu G, Hayashi T, Xenophontos SL, VeldhusienB, Saris JJ, Reynolds DM, Cai Y, Gabow PA, Pierides A, Kimber-tion or more rapid cyst development (and hence, early-ling WJ, Breuning MH, Deltas CC, Peters DJM, Somlo S:onset disease), the staining results suggest that dosagePKD2, a gene for polycystic kidney disease that encodes an integral

reduction of the PKD1 and TSC2 proteins during renal membrane protein. Science 272:1339–1342, 19963. Hughes J, Ward CJ, Peral B, Aspinwall R, Clark K, San Mil-development may be sufficient to initiate cystogenesis,

lan JL, Gamble V, Harris PC: The polycystic kidney disease 1possibly indicating that these proteins may act in the(PKD1) gene encodes a novel protein with multiple cell recognition

same or complementary pathways. domains. Nature Genet 10:151–160, 19954. International Polycystic Kidney Disease Consortium: Polycys-The second group of early-onset PKD1 cases is those

tic kidney disease: The complete structure of the PKD1 gene andin which severe disease arises in families with otherwiseits protein. Cell 81:289–298, 1995

typical adult-onset disease. Analysis of the recurrence 5. Moy GW, Mendoza LM, Schulz JR, Swanson WJ, Glabe CG,Vacquier VD: The sea urchin sperm receptor for egg jelly is arisk for early-onset disease in these families [35] andmodular protein with extensive homology to the human polycysticmolecular studies [36] indicate that the germline muta-kidney disease protein, PKD1. J Cell Biol 133:809–817, 1996

tion does not dictate the severity of disease but that 6. Ong ACM, Ward CJ, Butler RJ, Biddolph S, Bowker C, TorraR, Harris PC: Coordinate expression of the ADPKD proteins,other modifying loci (similar to those found in rodentpolycystin-2 and polycystin-1, in normal and cystic tissue. Am Jmodels of PKD [37, 38]) determine the disease course.Pathol 154:1721–1729, 1999

Analogous to the contiguous gene deletion cases, our 7. Qian F, Germino FJ, Cai Y, Zhang X, Somlo S, Germino GG:PKD1 interacts with PKD2 through a probable coiled-coil domain.staining results suggest that cysts could develop in theseNature Genet 16:179–183, 1997cases with haploinsufficiency at PKD1 and mutation at 8. Tsiokas L, Kim E, Arnould T, Sukhatme VP, Walz G: Homo-

a modifying locus (whose products may be part of the and heterodimeric interactions between the gene products ofPKD1 and PKD2. Proc Natl Acad Sci USA 94:6965–6970, 1997same complex or pathway), without the requirement for

9. Peral B, San Millan JL, Ong ACM, Gamble V, Ward CJ, Strongsomatic change at PKD1. Alternatively, modifier genes C, Harris PC: Screening the 39 region of the polycystic kidneycould influence the rate of somatic mutation at PKD1, disease 1 (PKD1) gene reveals six novel mutations. Am J Hum

Genet 58:86–96, 1996but a two-hit disease mechanism in these cases would10. Peral B, Gamble V, Strong C, Ong ACM, Sloane-Stanley J,require a high rate of somatic missense changes at PKD1 Zerres K, Winearls CG, Harris PC: Identification of mutations

in utero [39, 40]. in the duplicated region of the polycystic kidney disease 1 (PKD1)gene by a novel approach. Am J Hum Genet 60:1399–1410, 1997In summary, we have found evidence in nine adult-

11. Roelfsema JH, Spruit L, Saris JJ, Chang P, Pirson Y, van Ommenonset PKD1 patients with defined truncating mutations G-JB, Peters DJM, Breuning MH: Mutation detection in theof continued expression of the “normal” polycystin-1 repeated part of the PKD1 gene. Am J Hum Genet 61:1044–1052,

1997protein in the majority of cysts and the absence of12. Brook-Carter PT, Peral B, Ward CJ, Thompson P, Hughes J,

polycystin-1 expression in a significant minority. These Maheshwar MM, Nellist M, Gamble V, Harris PC, Sampsonresults may be reconciled with a “two-hit” model of JR: Deletion of the TSC2 and PKD1 genes associated with severe

infantile polycystic kidney disease: A contiguous gene syndrome.cystogenesis if the majority of somatic PKD1 mutationsNat Genet 8:328–332, 1994

are missense or “in frame.” The different expression 13. Sampson JR, Maheshwar MM, Aspinwall R, Thompson P,Cheadle JP, Ravine D, Roy S, Haan E, Bernstein J, Harris PC:pattern of polycystin-1 in the three early-onset cases of

Ong et al: Polycystin-1 expression in PKD1 cystic tissue 1333

Renal cystic disease in tuberous sclerosis: The role of the polycystic cles from simian virus 40-transformed mouse fibroblasts. J BiolChem 252:1990–1997, 1977kidney disease 1 gene. Am J Hum Genet 61:843–851, 1997

27. Longa L, Scolari F, Brusco A, Carbonara C, Polidoro S, Val-14. Lu W, Peissel B, Babakhanlou H, Pavlova A, Geng L, Fan X,zorio B, Riegler P, Migone N, Maiorca R: A large TSC2 andLarson C, Brent G, Zhou J: Perinatal lethality with kidney andPKD1 gene deletion is associated with renal and extra-renal signspancreas defects in mice with a targeted Pkd1 mutation. Natureof the autosomal dominant polycystic kidney disease. Nephrol DialGenet 17:179–181, 1997Transplant 12:1900–1907, 199715. Qian F, Watnick TJ, Onuchic LF, Germino GG: The molecular

28. Faraggiana T, Bernstein J, Strauss L, Churg J: Use of lectinsbasis of focal cyst formation in human autosomal dominant polycys-in the study of histogenesis of renal cysts. Lab Invest 53:575–579,tic kidney disease type 1. Cell 87:979–987, 1996198516. Brasier JL, Henske EP: Loss of the polycystic kidney disease

29. Holthofer H, Virtanen I, Pettersson E, Tornroth T, Alfthan(PKD1) region of chromosome 16p13 in renal cyst cells supports O, Linder E, Miettinen A: Lectins as fluorescence microscopica loss-of-function model for cyst pathogenesis. J Clin Invest 99: markers for saccharides in the human kidney. Lab Invest 45:391–194–199, 1997 399, 1981

17. Watnick TJ, Torres VE, Gandolph MA, Qian F, Onuchic LF, 30. Howie AJ: Epithelial membrane antigen in normal and proteinuricKlinger KW, Landes G, Germino GG: Somatic mutation in indi- glomeruli and in damaged proximal tubules. J Pathol 148:55–60,vidual liver cysts supports a two-hit model of cystogenesis in autoso- 1986mal dominant polycystic kidney disease. Mol Cell 2:247–251, 1998 31. Greenblatt MS, Bennett WP, Hollstein M, Harris CC: Muta-

18. Wu G, D’Agati V, Cai Y, Markowitz G, Park JH, Reynolds tions in the p53 tumor suppressor gene: Clues to cancer etiologyDM, Maeda Y, Le TC, Hou J Jr, Kucherlapati R, Edelmann and molecular pathogenesis. Cancer Res 54:4855–4878, 1994

32. Baas IO, Mulder J-WR, Offerhaus GJA, Vogelstein B, Hamil-W, Somlo S: Somatic inactivation of Pkd2 results in polycysticton SR: An evaluation of six antibodies for immunohistochemistrykidney disease. Cell 93:177–188, 1998of mutant p53 gene product in archival colorectal neoplasms.19. Lu W, Fan X, Basora N, Babakhanlou H, Law T, Rifai N,J Pathol 172:5–12, 1994Harris PC, Perez-Atayde AR, Rennke HG, Zhou J: Late onset

33. Finlay CA, Hinds PW, Tan TH, Eliyahu D, Oren M, Levineof renal and hepatic cysts in Pkd1-targeted heterozygotes. NatureAJ: Activating mutations for transformation by p53 produce aGenet 21:160–161, 1999gene product that forms an hsc700-p53 complex with an altered20. Ward CJ, Turley H, Ong ACM, Comley M, Biddolph S, Chettyhalf-life. Mol Cell Biol 8:531–539, 1988R, Ratcliffe PJ, Gatter K, Harris PC: Polycystin, the polycystic 34. Harvey M, Vogel H, Morris D, Bradley A, Bernstein A, Done-

kidney disease 1 protein, is expressed by epithelial cells in fetal, hower LA: A mutant p53 transgene accelerates tumour develop-adult and polycystic kidney. Proc Natl Acad Sci USA 93:1524–1528, ment in heterozygous but not nullizygous p53-deficient mice. Na-1996 ture Genet 9:305–311, 1995

21. Peters DJM, Spruit L, Klingel R, Prins F, Baelde HJJ, Gior- 35. Zerres K, Rudnik-Schoneborn S, Deget F: German workingdano PC, Bernini LF, de Heer E, Breuning MH, Bruijn JA: group on paediatric nephrology: Childhood onset autosomal domi-Adult, fetal and polycystic kidney expression of polycystin, the nant polycystic kidney disease in sibs: Clinical picture and recur-polycystic kidney disease-1 gene product. Lab Invest 75:221–230, rence risk. J Med Genet 30:583–588, 1993

36. Peral B, Ong ACM, San Millan JL, Gamble V, Rees L, Harris1996PC: A stable, nonsense mutation associated with a case of infantile22. Palsson R, Sharma CP, Kim K, McLaughlin M, Brown D, Ar-onset polycystic kidney disease 1 (PKD1). Hum Mol Genet 5:539–naout MA: Characterization and cell distribution of polycystin,542, 1996the product of autosomal dominant polycystic kidney disease gene

37. Iakoubova OA, Dushkin H, Beier DR: Localization of a murine1. Mol Med 2:702–711, 1996recessive polycystic kidney disease mutation and modifying loci23. Griffin MD, Torres VE, Grande JP, Kumar R: Immunolocaliza-that affect disease severity. Genomics 26:107–114, 1995tion of polycystin in human tissues and cultured cells. Proc Assoc

38. Woo DDL, Nguyen DKP, Khatibi N, Olsen P: Genetic identifica-Am Phys 108:185–197, 1996 tion of two major modifier loci of polycystic kidney disease progres-24. Geng L, Segal Y, Peissel B, Deng N, Pei Y, Carone F, Rennke sion in PCY mice. J Clin Invest 100:1934–1940, 1997

HG, Glucksmann-Kuis AM, Schneider MC, Ericsson M, Reed- 39. Reeders ST, Zerres K, Gal A, Hogenkamp T, Propping P,ers ST, Zhou J: Identification and localization of polycystin, the Schmidt W, Waldherr R, Dolata MM, Davies KE, WeatherallPKD1 gene product. J Clin Invest 98:2674–2682, 1996 DJ: Prenatal diagnosis of autosomal dominant polycystic kidney

25. Ong ACM, Harris PC: Molecular basis of renal cyst formation: disease with a DNA probe. Lancet 2:6–8, 1986One hit or two? Lancet 349:1039–1040, 1997 40. Waldherr R, Zerres K, Gall A, Enders H: Polycystic kidney

disease in the fetus. Lancet 2:274–275, 198926. Lever JE: Active amino acid transport in plasma membrane vesi-