Kidney International, Vol. 40 (1991), pp. 989—996

EDITORIAL REVIEW

Polycystic kidney disease: Clues to pathogenesis

Autosomal-dominant polycystic kidney disease (ADPKD),largely neglected for several decades, has emerged in recentyears as the renal disease most likely to be understood from thegene to the patient. Major breakthroughs have occurred in thegenetics of the disorder and new experimental data is providinginsights in the pathobiology of cyst formation (Table I).

Genetics of ADPKD

Localization of ADPKD gene

Prior to 1985 information on the genetics of ADPKD wasrestricted to the simple concept of transmission of an autosomaldominant disorder. In 1985 Reeders et al applied the techniquesof reverse genetics to this disorder and linkage was detectedbetween the ADPKD gene and the 3' hypervariable region (3'HVR) of the alpha globin locus [I]. This linkage placed theADPKD gene (ADPKDI) on the short arm of chromosome 16.Subsequent investigation has resulted in the development ofclosely-linked flanking probes which more precisely localize thegene to the tip of this chromosome [2]. This discovery whichdelineates the location of the ADPKD gene has enlarged theclinician's diagnostic repertoire for this disease, and providesthe first step to ultimate isolation of the gene and its geneproduct.

Prior to the availability of gene linkage analysis, the disorderand hence the gene carrier state could only be diagnosed afterbilateral renal cysts were identified. This often imposed barriersto effective genetic counselling, as the cysts were frequentlydetected only after the childbearing years. In the last decadethis barrier to early diagnosis, and hence effectively-timedgenetic counselling, has been lowered substantially by theimprovement of imaging techniques with more sensitive ultra-sonography and the use of computed tomography and magneticresonance imaging which permit the reliable detection of verysmall renal cysts [3—5]. For example, in our family studies ofsubjects at risk for ADPKD, 40% of children less than 20 yearsof age who undergo abdominal ultrasonography have detectablerenal cysts, a percentage approaching the predicted frequencyof 50% in an at-risk population. In other words, screeningultrasonography alone appears to detect about 80% of thosewho on a statistical basis are likely to have the disease.Similarly Bear et al estimate that 68% of children between ages11 and 20 who have ADPKD have detectable cysts usingultrasonography [6]. Most recently, Parfrey et al have demon-strated that in individuals who are likely to have the ADPKDgene by linkage techniques (ADPKDI; see below) and who are

Received for publication October 17, 1990and in revised form February 20. 1991Accepted for publication February 22, 1991

© 1991 by the International Society of Nephrology

less than 30 years of age, only 17% did not have diagnosticultrasonography for ADPKD [7].

Nonetheless, despite these improvements in renal imagingwhich permit earlier diagnosis, these methods are still depen-dent upon the presence of detectable renal cysts. Localizationof the gene and the subsequent commercial availability of theDNA probes have made diagnosis possible prior to the detec-tion of any renal cysts. In fact, it is now possible to determinethe gene status of an at-risk fetus in utero [8]. Since the geneitself has not yet been identified, the necessary prerequisites forthe application of gene linkage analysis must be present: thepresence of at least two affected individuals in order to deter-mine the marker type segregating with ADPKD in the family;the presence of informative marker types in the family, that is,different types in affected and unaffected individuals; and afamily with the ADPKD gene located on chromosome 16, asdiscussed later (Fig. 1). Individuals fulfilling all these criteriacan have the presence of the ADPKD gene predicted prior tothe occurrence of any phenotypic manifestation at 99 to 1 oddsrather than the 50/50 odds that would be predicted from theprinciple of inheritance alone [9]. Moreover, as exemplified bythe study of Cobben et al, utilization of linkage analysis haspermitted expansion of the knowledge of phenotypic presenta-tion of ADPKD to include the occurrence of congenital hepaticfibrosis which previously had been believed to be largelyassociated with autosomal-recessive polycystic kidney disease[10].

Although these discoveries have provided new tools fordiagnosis, their role in screening and in prenatal diagnosis hasgenerated some controversies [II]. Resolution of these willlargely depend upon the development of effective early inter-vention which will provide a therapeutic incentive for screen-ing.

Genetic heterogeneity

Another major advance in the genetics of ADPKD was thediscovery that the clinical disorder which was assumed to resultfrom one gene defect could result from different genetic defects[12—15]. All of the initial 100 families examined with the DNAprobes for the alpha globin region of chromosome 16 demon-strated by linkage analysis that the putative gene was located onchromosome 16, suggesting genetic homogeneity—one genedefect for one disorder [2]. However, Kimberling et al thenreported a large Sicilian family whose gene for ADPKD was notlinked to markers for the alpha globin region of chromosome 16(ADPKD2 or n0nADPKDI) [12]. This report and subsequentreports [13—15] confirmed genetic heterogeneity in ADPKD,that is, different gene defects could produce the disorder ofADPKD. Among the Caucasian population it appears that theADPKDI gene (the gene on chromosome 16) accounts for 90%of ADPKD [16]. Although there appears to be no geographicclustering of gene types in the Caucasian population, it remains

989

990 Gabow: Polycysrk kidney disease

Table 1. Advances in ADPKD

Genetic breakthroughs in ADPKDLocalization of an ADPKD gene to chromosome 16Establishment of genetic heterogeneity

Development of models for studying ADPKDCulture of ADPKD renal cyst epitheliumMDCK cell-cyst modelcpk MouseMetanephric organ culture (cpk mouse)

Information on pathobiology of renal cystogenesisIncreased cellular proliferation

Intracyst polypsRenal adenoma in ADPKDIncreased growth potential in tissue culturesPeptide growth factors' alterationsOncogene promoted cystogenesis

Secretion in cyst formationcAMP stimulation of MDCK cell cysts"Reversal of Na-K ATPase polarity

Abnormal extracellular matrixPathologic abnormality of cyst basement membraneAltered basement composition

II

III

aa ab

Fig. 1. Pedigree of an ADPKD family. Clinically affected individualsare shown as solid symbols. The linkage types are displayed below eachsymbol. The affected male in the first generation is dead and his typehas been deduced to be ab. In this family the b type is segregating withthe PKD gene. Thus, in the Ill generation it can be deduced that the aaoffspring (or fetus) has a 95% chance of being unaffected and the aboffspring (or fetus) has a 95% chance of being affected. Families can beuninformative in a variety of ways. For example, if the female ingeneration I or the unaffected spouse in generation II had been ab, thelikelihood of carrying the ADPKD gene could not be established usingonly a single probe. Other marker probes would likely be informative.

to be determined if gene distribution is similar in all racialgroups. The occurrence of genetic heterogeneity has onlymodestly complicated the use of gene linkage techniques infamilies. The information that one gene is on chromosome 16,the availability of flanking probes to that gene, and the substan-tially greater frequency of ADPKD1 versus ADPKD2 permitsassessment of probability of PKD type in a family, and if thefamily is ADPKD1 the carrier status of the individuals in

question can still be determined in a majority of families with anaccuracy of 95% or greater.

With the discovery of genetic heterogeneity, the question ofclinical or phenotypic heterogeneity logically followed.ADPKD, like other autosomal dominant diseases, demon-strates considerable phenotypic variability. Although Dalgaardsuggested similarity in the age of end-stage renal disease infamily members with ADPKD [171, information on interfamilyvariability in ADPKD which would support different genesproducing different clinical syndromes has been lacking. How-ever, information is now being accumulated that supportsphenotypic variability between the two gene types, suggestingthat ADPKD2 is a milder renal form of the disorder [71. Thiscould provide a genetic explanation for Dalgaard's clinicalobservation of clustering of renal disease severity within fami-lies. The array and severity of extrarenal manifestations forADPKD1 and ADPKD2 remain to be examined. It should alsobe emphasized that it is not yet clear that all the so-calledADPKD2 families are similar; they, in fact, could representdifferent genes' defects. Clarification of this point will awaitidentification of the chromosome carrying the ADPKD2 gene.

These important advances in our understanding of the genet-ics of ADPKD have not yet permitted us to connect the geneticmutation to a specific molecular or cellular abnormality whichproduces the systemic disorder of ADPKD. However, muchnew information has been generated about the major pathobio-logical processes which participate in renal cystogenesis: aug-mented cell growth, disordered cellular secretion, and elabora-tion of an abnormal extracellular matrix. To date, informationhas not permitted us to establish which, if any, of these factorsis the primary pathogenetic mechanism in ADPKD. If one ofthese abnormalities is ultimately proven to be primary in thepathogenesis of ADPKD, the genetic defect should be related toit either directly or indirectly and then the other abnormalitieswould be secondary or pan passu events. The potential role ofeach of these factors in cystogenesis has been suggestedthrough pathologic examination of renal tissues, investigationsof animal models of induced and naturally occurring recessivecystic disease, and cell culture studies of human cystic epithe-ha.

Pathobiology of cystogenesis

Altered growth

The first observation of altered growth in both human andanimal models of cystic disease was the electron microscopicfinding of cyst wall epithelial cell hyperplasia manifested bymicropolyp formation within the cyst [18]. Subsequent studieshave demonstrated other hyperphastic abnormalities within cystwalls, including cord-like arrays of hyperplastic epithelium [19].Moreover, mathematical calculations based on cyst size andepithehial cell dimensions also support the need for cells toproliferate in order to form cysts [191. That is, a cyst is notsimply a ballooning of a renal tubule and stretching of cells, butindeed the result of increased cell numbers.

Further characterization of the exaggerated growth potentialof cystic epithelial cells was obtained from tissue culturestudies. In 1986 Wilson Ct al first successfully cultured humanADPKD renal cystic epithehium [20]. Primary cultures ofADPKD cystic epithehia cells showed suggestively increased

ab

Gabow: Polycystic kidney disease 991

cell growth potential compared to age-matched human renalproximal straight tubular, thick ascending limb, and corticalcollecting tubular cells in that the ADPKD culture achievedconfluence more rapidly than did the normal epithelial cells[201. An increased growth capacity was characterized by atwofold increase in the number of rounds of cell division in theproliferative phase and a prolonged stationary phase of nineweeks in ADPKD cyst epithelial cells compared to three weeksin tubules from both age-matched normal kidneys and end-stagenon-ADPKD kidneys [21]. The exaggerated growth potential ofADPKD cyst epithelial cells is further underscored by theability of these cells to attach and proliferate without matrix onuncoated plastic—a prerequisite for individually isolated nor-mal renal tubular cell growth [21]. Of note, other investigatorshave not demonstrated an alteration in the rate of epithelial cellproliferation [22, 23j, an altered life span [23], nor an alteredability to clone on plastic surfaces [23] in ADPKD epithelia.However, those studies which fail to demonstrate alteredgrowth characteristics have used passaged cell cultures ratherthan primary cell, and these passaged cells may not accuratelyexhibit the true in vivo characteristics of ADPKD epithelium.Certainly, other cellular characteristics are altered by passagingof the cells [24]. Moreover, these latter studies have notcompared ADPKD epithelia to age-matched normal tissuewhich could alter interpretation of growth characteristics.Thus, in aggregate, the data suggest that cystogenesis results, atleast in part, from increased cell proliferation. This alteredproliferative capacity should result from a primary or secondaryabnormality in one or more of the factors involved in regulationof epithelial cell growth, including peptide growth factors,oncogenes, or cell matrix. In fact, growth promoting alterationshave been identified in ADPKD in all three factors.

Growth factors in ADPKD

Epidermal growth factor, which is a potent mitogen, has beendemonstrated not only to elicit an exaggerated proliferativeresponse in cultured ADPKD cystic epithelium [22, 25, 26], butalso to be present in ADPKD cyst fluid and cystic epithelium[25, 27]. In contrast, an inhibitor of renal tubular proliferation,TGF/3, does not produce the expected inhibition in culturedADPKD cystic epithelium [25]. This information suggests thatan alteration between peptide stimulators and inhibitors of cellgrowth could contribute to cellular proliferation in ADPKD.

Onco genes in ADPKD

Altered cell growth could be a potential link between anabnormal genetic message and cystogenesis. Oncogenes whichimpart to cells malignant traits, including proliferative capacity,are a logical target for such a link. Initial studies to pursue thepossible role of oncogenes utilized the CPK mouse, which is amodel for recessive PKD. In this mouse strain homozygotesdevelop massively enlarged kidneys riddled with cysts and dieof renal failure in the first month of life [28, 29]. Although thistype of PKD is recessively inherited, relevant informationregarding basic principles of cystogenesis can be derived.Cowley et at demonstrated a two- to sixfold increase in c-,nycRNA at two weeks and an approximately 30-fold increase atthree weeks in the kidneys of the affected CPK mice comparedto normal litter mates [30]. Cowley et al extended these originalobservations demonstrating that the elevated oncogene expres-

sion was not limited to c-myc but that mRNA for c-fos andc-Ki-ras were also increased [31]. As with c-myc mRNA, amodest elevation at two weeks of age was followed by a moremarked increase at three weeks of age. Although it is temptingto tie this elevated expression of several oncogenes to theobserved increased cell proliferation in ADPKD, the cell pro-liferation was not increased to the same degree as was theoncogene expression; moreover, oncogene mRNA was highestwhen proliferation was no longer as marked. Hence thesestudies, albeit provocative, did not prove a definitive casual rolefor oncogenes in cystogenesis.

Subsequently, transgenic mice studies have furthered theconcept of a causal role for gene-mediated proliferation incystogenesis. Studies producing transgenic mice utilizing thesimian virus 40 (SV4O) early region encoding for the either smalland large T antigen or the large T antigen alone manifest renalcystic disease coupled with tubular cell hyperplasia and otherabnormalities [32—34]. Similarly, a fusion gene consisting of theSV4O enhancer, the beta globin promoter, and the codingportion of the murine c-myc gene resulted in transgenic micewhich manifested bilateral renal cystic disease, tubular epithe-hal cell hyperplasia and in occasional kidneys, microadenomas[351. Since 100% of these transgenic mice developed renalcystic disease, the study provides compelling evidence for therole of oncogenes and/or gene mediated proliferation in cysto-genesis. However, further studies will be required to define therelationship of these observations, which substantiate the ca-pacity of oncogenes to produce cystogenesis, to the actual PKDgene(s). In these transgenic mice the putative oncogene is in allcells; however, in at least one study the major defect wasconfined to the kidney [34]. This raises the critical question ofhow specific cell types are targeted in genetic disorders.

In fact, this question has been raised regarding the cellularexpression of the ADPKD gene: If all cells of an affectedindividual contain the defective gene, why is the disorder sosporadic both within epithelial structures and within the af-fected organ? The renal cyst involvement in ADPKD dramati-cally underscores the latter issue. Less than five percent of allnephrons appear to undergo cystic changes despite the pres-ence of the gene in all cells [19]. Although similar carefulquantitative dissection studies have not been performed in theliver, it is likely that a similar distribution of cystic andnon-cystic bile ducts occurs. Moreover, within individual neph-rons only some segments undergo cystogenesis, and addition-ally the cystogenic segments vary from nephron to nephron.Some investigators have observed that this variability in cellu-lar expression is compatible with the need for a sporadicadditional somatic mutation such as occurs in retinoblastoma[21; others have suggested that local microenvironmental fac-tors alter the cellular expression of the gene [23, 36].

Role of secretion in cystogenesisGiven that disordered growth has been demonstrated in some

circumstances in cystic disease it is interesting to note that it isnot confined to cyst formation alone. There is also an increasedfrequency of renal adenoma in both human polycystic disease[37] and the transgenic mouse models [33—35]. The presence ofnon-cystic adenoma in ADPKD kidneys suggests the necessityfor a factor in addition to cell proliferation in order for cysto-genesis to occur. This factor appears to be cellular secretion.

992 Gabow: Polvcystic kidney disease

Cellular proliferation without secretion will produce a tumor oradenoma, whereas increased cellular proliferation accompaniedby cellular secretion which is accompanied by increased inter-cyst pressure [38] produces outward expansion yielding a cyst.The interrelationship of these two processes and the key role ofsecretion in cystogenesis have been demonstrated by studies ofMadin-Darby canine kidney (MDCK) cells in culture [38, 39].McAteer, Evan and Gardner developed a model for cystogen-esis in which MDCK cells were dispersed within a hydratedcollagen gel [40]. Via clonal growth, single MDCK cells formedcystic structures polarized with the basolateral surface in con-tact with the gel and the apical surface facing the cyst lumen[40]. Of note, only about 7% of MDCK cells cultured in thismanner participate in cyst formation, emphasizing yet again theheterogeneity in the cellular capacity to initiate cyst formation[38].

A series of elegant experiments by Mangoo-Karim et al [391,utilizing subcultures from the cells of a single MDCK cystgrown by the method of McAteer et al [40], were performed toidentify chemical mediators of secretion in cystogenesis. Pre-liminary studies lead to a focus on the role of cAMP. In theabsence of stimulation of cAMP or the absence of a cAMPagonist in the medium, these MDCK cells, endowed with thecapacity for cystogenesis, did not form cysts [39]. With theaddition of 3-isobutyl-l-methylxanthine, a phosphodiesteraseinhibitor, low intracellular levels of cAMP presumably occurredand cell proliferation was initiated, resulting in balls of cells ortumors, but not cysts [39]. Compounds such as prostaglandinE1 (PGEJ) and cholera toxin, which increase intracellularcAMP substantially, resulted in large fluid-filled cysts [39].Further evidence for a key role for cAMP was the enhancementof the effect of each of these compounds on both cyst number(cell proliferation) and volume (fluid secretion) by the additionof 3-isobutyl-l-methylxanthine [39].

Additionally, evidence for the role of cAMP in cyst fluidsecretion has been obtained from examination of confluentmonolayers of these MDCK cyst epithelial cells. "Uphill" fluidsecretion against a concentration gradient was demonstratedwhen the basolateral surfaces were exposed to agents whichstimulated cAMP [39].

The potential role of cAMP in human cystogenesis is sup-ported by data from hepatic and renal cyst puncture studies inpatients with ADPKD. In the study by Everson et al threehepatic and one non-gradient renal cysts were punctured andthe response to intravenous secretin was measured. Two of thethree hepatic cysts and the renal cyst demonstrated increasedfluid secretion within eight minutes of secretin administration[41]. Although the process by which secretin stimulates epithe-hal secretion is not entirely delineated, it is known to involvecAMP [42, 43]. Therefore, both in vitro and in vivo data suggesta role for stimulation of cAMP in cyst fluid secretion and hencein cyst growth. This area offers promise for possible modalitiesfor therapeutic intervention in this disorder.

An additional potential mediator for disordered secretionrelates to abnormalities in the quantity of and location of renalepithelial Na-K ATPase. Avner et al have delineated thealterations in renal ATPase in relationship to the stages ofcystogenesis in the CPK mouse [44]. From the beginning of cystformation at fetal day 17 to postnatal day 12 (the period ofproximal tubular cysts), total renal ATPase and Na-K ATPase

were increased in cystic kidneys compared to control kidneys[44]. Furthermore, in a different model, the metanephric organculture of the CPK mouse, Na-K ATPase activity inductionwith triiodothyronine was associated with sustained proximaltubular cystogenesis, and conversely Na-K ATPase inhibitionwith ouabain prevented sustained cyst formation [45].Addition-ally, in the MDCK epithelial cell cyst model, ouabain com-pletely inhibited fluid secretion [39].

Human cystic ADPKD epithelial tissue, both that which hasbeen recently excised from the cyst wall and that in primary cellculture, appears to demonstrate altered cellular polarity withNa-K ATPase located exclusively on the apical rather than thebasolateral cell surface as in normal tubules [46]. Of note, avariety of other basolateral and apical proteins maintained theirappropriate location. However, actin and the ankyrin-fodrincomplex, which are believed to be involved in the anchoring ofNa-K ATPase in the appropriate membrane, colocalized withNa-K ATPase at the apical surface in the ADPKD tissue [46].

In this altered location Na-K ATPase appears to maintain itsfunction, and its reversal in location is accompanied by analtered direction of cellular sodium transport from the basolat-eral to the apical cell surface [46]. One can easily imagine howsuch altered direction of transport within a blind cyst wouldresult in cyst enlargement. However, the relationship of thisabnormality to the primary gene defect remains to be deter-mined. In this regard, although alterations in both proliferationand secretion appear pivotal to cystogenesis, the relationship ofthese abnormalities to the non-cystic systemic manifestations ofADPKD is difficult to construct. Moreover, as pivotal ascellular proliferation and secretion are in renal cystogenesis,there is a caveat in focusing on these processes as primary inADPKD. Many congenital disorders [47J, chemical substances[48], hypokalemia [49] and long-term dialytic therapy [50]all areassociated with renal cysts, suggesting that in this tubular organengaged in reabsorption and secretion, cysts may well be a finalcommon pathway for diverse pathophysiologic insults.

Extra cellular matrix in cysto genesis

The third major abnormality associated with human cystoge-nesis is an altered extracellular matrix. From a clinical perspec-tive, a reason for implicating the extracellular matrix inADPKD is the array of extrarenal abnormalities. These abnor-malities have recently been reviewed [51] and include hepaticcysts [52, 53], cardiac valve abnormalities [54, 55], intracranialaneurysms [56, 57], hernia formation [51], and perhaps diver-ticula of the colon [58]. Although a defect in the extracellularmatrix would offer a reasonable explanation for both thediversity and type of systemic manifestations, the simple pres-ence of these abnormalities provides no direct evidence for anextracellular matrix defect. Data supporting the clinical sugges-tion of a matrix abnormality come from examination of the renalabnormalities.

Milutinovic noted that renal biopsy specimens from subjectsat risk for ADPKD who later developed detectable cystsrevealed splitting and duplication of the glomerular basementmembranes [59]. Wilson, Hreniuk and Gabow [21] extended anearlier pathologic observation of Cuppage [60] demonstratingthat the basement membrane surrounding the cysts in ADPKDkidneys is extraordinarily thickened, up to twenty times thethickness of normal tubular basement membrane, and appears

Gabow: Polycystic kidney disease 993

to consist of interwoven fibrils rather than uniformly densematerial, as can be seen with the mild thickening that occurswith aging.

However, as does the clinical profile, these histologic findingsonly provide indirect evidence for altered extracellular matrixin the pathogenesis of ADPKD. More direct support for a roleof altered extracellular matrix is obtained from ultrastructuratand compositional changes in the basement membrane of ani-mal and experimental models of PKD, and from tissue culturestudies of cystic and normal renal epithelium. Ojeda et al haveelegantly depicted the ultrastructural abnormalities that accom-pany the development and regression of tubular cysts in thecorticosteroid-induced model of cystic disease in newbornrabbits [61]. Cystic tubular basement membrane containedincreased amounts of ruthenium red staining material (whichmay represent proteoglycans), and appeared to have slightlydecreased amounts of immunofluorescent staining for lamininand type IV collagen compared to basement membrane fromnormal control animals. Fibronectin, an extracellular matrixcomponent found in the interstitium, appeared increased in thearea surrounding the cyst [611.

Carone, Makino and Kanwar found somewhat similar resultsutilizing antibodies directed against basement membrane con-stituents [62]. They compared ADPKD kidneys from humans tonon-ADPKD human kidneys and rat kidneys with toxin-in-duced cystic disease to normal kidneys from control animals.Although there were some differences in data from human andrat tissue, there was loss of reactivity to anti-heparan sulfateproteoglycans in cyst basement membrane and no change or anincrease in immunofluorescence with anti-type IV collagen andlaminin in cystic kidneys. Fibronectin appeared to be greatlyincreased in peritubular areas and in the interstitium. More-over, in other similar studies Carone documented that thebasement membrane alteration occurred in lockstep with cystdevelopment, suggesting a causal relationship [63]. However,Wilson and Sherwood have demonstrated that increased typeIV collagen mRNA occurs late in the course of ADPKD [251.Others have found alterations in type IV collagen and laminin inthe recessive murine model of PKD. Increased mRNA levelsfor collagen IV and lamina were found in the analysis of wholekidneys of three-week-old cpk mice [64]. This observation wasextended by Taub et al with the demonstration that primarytissue cultures of kidney epithelial cells of CPK mice displayincreased synthesis of collagen IV and laminin and increasedcellular mRNA for the laminin BI chain [65].

Wilson et a! demonstrated that human ADPKD cells inculture elaborated greater amounts of extraceltular matrix thandid normal human renal proximal and distal tubular epithelia[20]. Under similar conditions of cell culture confluence,ADPKD cyst epithelial monolayers incorporated 10- to 15-foldthe amount of radiolabeled inorganic sulfate than did normaltubules. However, electron microscopy of these cell culturesrevealed that the extracellular matrix had an abnormal appear-ance. Rather than being a thin layer of material as in normalepithelial cell cultures, the matrix contained bundles of bandedcollagen and spherical accumulation of proteinous materialwhich stained with ruthenium red, suggesting that at least onecomponent of the increased matrix was a proteoglycan [201which may not achieve normal structural characteristics [25].

Granot et al utilized a slightly different culture system to

examine both intracellular and extracellular protein profilesusing 35S-methionine protein radiolabeling and SDS-PAGEseparation for ADPKD cyst epithelial cell culture and for cellcultures from normal human kidney cortex [22]. A remarkablesimilarity existed for the 465 radiolabeled proteins identified.However, ADPKD cell cultures underexpressed three intracel-lular proteins and overexpressed six other intracellular pro-teins. Moreover, ADPKD primary cell cultures produced threeextracellular proteins with apparent molecular weights of 45,170 and 220 kDa which were not produced by the normalkidney. Of considerable interest, although primary cultures ofnormal human kidney did not elaborate these proteins, multi-ple-passaged normal cells began to elaborate these same threeproteins, suggesting this altered synthesis in ADPKD epitheliais an inducible defect.

Thus, all the available data from various models and humanADPKD do not clearly establish a single uniform alteration inthe extracellular matrix composition; however, taken together,the data from pathological examination, animal and experimen-tal models and tissue culture support some type of an alterationin extracellular matrix in the process of cystogenesis. Initially,it was hypothesized that this alteration might contribute to cystformation via a simple change in the distensibility of the tubularbasement membrane. However, imaginative experiments byGrantham et al demonstrated unaltered viscoelastic propertiesof the basement membrane in experimental cystic disease [66].It appears more likely that the alteration in extracellular matrixcould contribute to cystogenesis via alterations in cell-matrixinteractions. These interactions are considerable including celladhesion [67], cell growth [68], cytodifferentiation [69], andperhaps gene expression [70]. Thus, if the extracellular matrixwere altered, cell proliferation could be altered as a secondaryevent. Similarly, secondary changes in cytodifferentiation couldaffect cell polarity, which in turn could alter secretory charac-teristics of the cell.

Microen vironmental factors

In addition to these three major aspects of cystogenesis anumber of other events may also be operative. For example, itis obvious that matrix remodelling must occur for cysts to formin the MDCK model of cyst growth [40]. In this regard it is ofnote that the fluid obtained from liver cysts of patients withADPKD appear to contain substances capable of modifyingprotein structure [41]. These proteins may contribute to matrixremodelling which would facilitate outward cyst expansion.Given the paucity of information regarding matrix remodelling,its role in cystogenesis in human disease has yet to be explored.

Both macro- and microenvironmental factors may also becapable of modulating cystogenesis. The potential role of themacroenvironment was demonstrated by studies of Werder,Gardner and others in which the progression of a naturallyoccurring murine model and a toxin-induced cystic diseasecould be radically altered by changing the external environment[71—73]. A germ-free environment significantly ameliorated thedisease, and introduction into the ambient environment orexposure to endotoxin markedly worsened the development ofcysts in the animals. A role for immune modulators in translat-ing this environmental effect to the renal microenvironment hasbeen proposed by Gardner [36]. Sterile cyst fluids from bothhepatic and renal cysts from patients with ADPKD often

994 Gabow: Polycystic kidney disease

contain large numbers of white cells suggesting an inflammatoryresponse. Moreover, Gardner et al have demonstrated that cystfluid contains an array of cytokines and lymphokines. Sixty-nine percent of the 94 cysts fluid contained interleukin-1B; 72%of 75 fluids contained tumor necrosis factor alpha and 30%contained interleukin-2 [74]. Thus, these immune modulatorscould provide macro- and microenvironmetal modulation ofcystogenesis.

However, others have suggested that influence of the mac-roe nvironment supports the interpretation that PKD is a thresh-old trait. With a threshold trait, the abnormality appears whenthe aggregate effect of the genetic and environmental factorsreach a critical or threshold level. McDonald et al have pre-sented evidence supporting this concept of a threshold modelfor PKD based on studies of multiple strains of mice whichmanifested highly variable susceptibility to the cystogenic effectof glucocorticoid administration [75]. Although this conceptmay have relevance to chemically-induced cystic disease, itsrelationship to human ADPKD is less clear. One possiblecorollary to this concept of a threshold phenomenon whichwould be of great interest, if applicable to human disease, is thepossibility of regression of the cystic disease. Kanwar andCarone have demonstrated marked regression of establishedcysts when the toxin is removed in a toxin-induced model inrats [761. This underscores the potential of reversibility of cysticdisease if putative agents are removed.

Thus, in the last decade our understanding of ADPKD hasadvanced substantially in the areas of genetics, clinical mani-festations and in the pathogenesis of the disorder. An ADPKDgene has been located on chromosome 16, genetic heterogene-ity has been established, and a potential role for oncogenes incystogenesis has been deduced from experimental data. Thepathobiology of cystogenesis has been extended from our initialconcept of a ballooning tubule filling with glomerular filtrate tothe demonstration of the linked contributions of increased cellgrowth and secretion. Data have accumulated demonstratingabnormal extracellular matrix in both experimental models andhuman cystic disease. Despite these significant steps forwardwe cannot yet identify the primary defect responsible forADPKD. We must continue our efforts to identify the specificgene abnormalities and the abnormal gene products; to relatethe gene defect to the array of abnormal cellular events incystogenesis in order to determine which events are, in fact,primary and finally we must begin to relate the gene defect notonly to the renal manifestations of this disorder, but also to theother prominent non-renal clinical abnormalities as well. Thisinformation will ultimately permit us to understand this com-mon disorder from the gene to the patient.

Acknowledgments

PATRICIA A. GABOWDenver, Colorado, USA

This research was supported by Grant 5 P01 DK34039, HumanPolycystic Kidney Disease (PKD), awarded by the Department ofHealth and Human Services, Public Health Service, NIDDK and theClinical Research Center, Grant MORR-0005l from the General Clini-cal Research Centers Research Program of the Division of ResearchResources, National Institutes of Health. The author thanks Dr.William Kaehny and Ann Johnson for their helpful editing of themanuscript and Linda Bonenberger for the manuscript preparation.

Reprint requests to Patricia Gabow, M.D., Box C283, University ofColorado, 4200 East Ninth Avenue, Denver, Colorado 80262, USA.

References

I. REEDERS ST. BREUNING MH, DAVIES KE, NICHOLLS RD, JARMANAP, HIGGS DR, PEARSON PL, WEATHERALL DJ: A highly polymor-phic DNA marker linked to adult polycystic kidney disease onchromosome 16. Nature 317:542—544, 1985

2. REEDERS ST, GERMINO GG, GILLESPIE GAJ: Recent advances inthe genetics of renal cystic disease. Mo! Biol Med 6:81—86, 1989

3. LAWSON TL, MCCLENNAN BL, SHIRKHODA A: Adult polycystickidney disease: Ultrasonographic and computed tomographic ap-pearance. JCU 6:295—382, 1978

4. LEVINE E, GRANTHAM JJ: The role of computed tomography in theevaluation of adult polycystic kidney disease. Am J Kidney Dis1:99—105, 1981

5. LEUNG AWL, BYDDER GM, STEINER RE, BRYANTDi, YOUNG IR:Magnetic resonance imaging of the kidneys. AJR 143:1215—1227,1984

6. BEAR JC, MCMANAMON P, MORGAN J, PAYNE RH, LEWIS H,GAULT MH, CHURCHILL DN: Age at clinical onset and at ultraso-nographic detection of adult polycystic kidney disease: Data forgenetic counselling. Am J Med Genet 18:45—53, 1984

7. PARFREY PS, BEAR JC, MORGAN J, CRAMER BC, MCMANAMON PJ,GAULT MH, CHURCHILL DN, SINGH M, HEWITT R, SOMLO S,REEDERS ST: The diagnosis and prognosis of of autosomal domi-nant polycystic kidney disease. N EngI J Med 323:1085—1090, 1990

8. REEDERS ST. ZERRES K, GAL A, HOGENKAMP T, PROPPING P,SCHMIDT W, WALDHERR R, DOLATA MM, DAVIES KE, WEATH-ERALL DJ: Prenatal diagnosis of autosomal dominant polycystickidney disease with a DNA probe. Lancet 2:6—8, 1986

9. BREUNING MH, REEDERS ST, BRUNNER H, IJDO JW, SARIS ii,VERWEST A, VAN OMMEN GJB, PEARSON PL: Improved earlydiagnosis of adult polycystic kidney disease with flanking DNAmarkers. Lancet 2:1359—1361, 1987

10. COBBEN JM, BREUNING MH, SCHOOTS C, TENKATE LP, ZERRESK: Congenital hepatic fibrosis in autosomal dominant polycystickidney disease. Kidney mt 38:880—885, 1990

II. GABOW PA, GRANTHAM ii, BENNETT W, CHILDRESS JF, COLE B,CONNEALLY PM, GARDNER K, KIMBERLING WJ, MARSH F, REED-ERS S: Gene testing in autosomal dominant polycystic kidneydisease: Results of National Kidney Foundation workshop. Am JKidney Dis 13:85—87, 1989

12. KIMBERLING WJ, FAIN PR, KENYON JB, GOLDGAR D, SUJANSKYE, GABOW PA: Linkage heterogeneity of autosomal dominantpolycystic kidney disease. N Engl J Med 319:913—918, 1988

13. RoMEo G, DEVOTO M, COSTA G, RONCUZZI L, CATIZONE L,ZUCCHELLI P, GERMINO GG, KEITH T, WEATI-IERALL DJ, REED-ERS ST: A second genetic locus for autosomal dominant polycystickidney disease. Lancet 2:8—10, 1988

14. DAWSON DB, TORRES yE, CIIARBONEAU JW, THIBODEAU SN:Detection of a family with autosomal dominant polycystic kidneydisease loosely linked to DNA markers from l6p. (abstract) Kidneymt 35:203, 1989

15. BRISSENDEN JE, RoscOF JM, SILVERMAN M: Assessment of alinked DNA marker for presymptomatic diagnosis of adult polycys-tic kidney disease in North America. (abstract) Kidney mt 35:202,1989

16. KIMBERLING WJ, PIEKE SA, KENYON JB, GABOW PA: An estimateof the proportion of families with autosomal dominant potycystickidney disease unlinked to chromosome 16. (abstract) Kidney mt37:249, 1990

17. DALGAARD OZ: Bilateral polycystic disease of the kidneys: Afollow-up of two hundred and eighty-four patients and their fami-lies. Ada Med Scand 328 (Suppl):l—255, 1957

18. EVAN AP, GARDNER KD JR. BERNSTEIN J: Polypoid and papillaryepithelial hyperplasia: A potential cause of ductal obstruction inadult polycystic disease. Kidney mt 16:743—750, 1979

19. GRANTHAM JJ, GEISER JL, EVAN AP: Cyst formation and growth inautosomal dominant polycystic kidney disease. Kidney mt 31:1145—1152. 1987

Gabow: Polycystic kidney disease 995

20. WILsoN PD, SCHRIER RW, BRECKON RD, GABOW PA: A newmethod for studying human polycystic kidney disease epithelia inculture. Kidney mt 30:371—378, 1986

21. WILSON PD, HRENIUK D, GABOW PA: Relationship betweenabnormal extracellular matrix and excessive growth of human adultpolycystic kidney disease epithelia. J Cell Physiol (in press)

22. GRANOT Y, VAN PUTTEN V, PRZEKWAS J, GABOW PA, SCHRIERRW: Intra- and extracellular proteins in human normal and poly-cystic kidney epithelial cells. Kidney fin 37:1301—1309, 1990

23. CARONE FA, NAKAMURA S, SCHUMACHER BS, PUNYARIT P,BAUER KD: Cyst-derived cells do not exhibit accelerated growth orfeatures of transformed cells in vitro. Kidney fin 35:1351—1357,1989

24. FRE5HNEY RI: Culture of Animal Cells. New York, Alan Liss Inc.,1983

25. WILsoN PD, SHERWOOD AC: Tubulocystic epithelium. Kidney mt39:450—463, 1991

26. AVNER ED, SWEENY WE: Epidermal growth factor induces hyper-plastic tubular cysts and increased Na-K-ATPase activity in vitro.(abstract) Am Soc Nephrol 1989, p. 24

27. MosKowlTz D, BONAR SL, MARCUS MD, CLAYMAN RV, AVNERED: Epidermal growth factor content of human and mouse (cpk)renal cyst. (abstract) Clin Res 37:497A, 1989

28. RUSSELL ES, MCFARLAND EC: Cystic kidneys. Mouse Newsleit56:40—43, 1977

29. PREMINGER GM. KOCH WE, FRIED FA, MCFARLAND E, MURPHYED, MANDELL J: Murine congenital polycystic kidney disease: Amodel for studying development of cystic disease. J Urol 127:556—560, 1982

30. COWLEY BD JR, SMARDO FL JR, GRANTHAM JJ, CALVET JPElevated c-myc protooncogene expression in autosomal recessivepolycystic kidney disease. Proc Nail Acad Sd USA 84:8394—8398,1987

31. COWLEY BD, CHADWICK U, GRANTHAM JJ, CALVET JP: Elevatedprotooncogene expression in polycystic kidneys of C57BL/6J (cpk)mouse. JAm Soc Nephrol 1:1048—1053, 1991

32. MACKAY K, STRIKER U, PINKERT CA, BRINSTER RL, STRIKERGE: Glomerulosclerosis and renal cysts in mice transgenic for theearly region of SV4O. Kidney mt 32:827—837, 1987

33. DYER KR: Inducible pathology in liver, pancreas, and kidney oftransgenic mice expressing 5V40 early region. Am J Pathol 401—410, 1989

34. KELLY KA, AGARWAL N, REEDERS 5, HERRUP K: Renal cystformation and multifocal neoplasia in transgenic mice carrying thesimian virus 40 early region. JAm Soc Nephrol 2:84—97, 1991

35. TRUDEL M, D'AGATI V, COSTANTINI F: c-myc Gene as an inducerof polycystic kidney disease in transgenic mice. Kidney fin 39:666—672, 1991

36. GARDNER KD JR: Cysts and cystic kidneys, in The Cystic Kidney,edited by GARDNER KD JR. BERNSTEIN J, Dordrecht, KluwerAcademic Publishers, 1990, pp. 3—17

37. GREGOIRE JR, TORRES VE, HOLLEY KE, FARROW GM: Renalepithelial hyperplastic and neoplastic proliferation in autosomaldominant polycystic kidney disease. Am J Kidney Dis 9:27—38, 1987

38. GRANTHAM JJ, UCHIC M, CRAGOE EJ JR, KORNHAUS J,GRANTHAM JA, DoNoso V. MANGOO-KARIM R, EVAN A,MCATEER J: Chemical modification of cell proliferation and fluidsecretion in renal cysts. Kidney mt 35:1379—1389, 1989

39. MANGOO-KARIM R, UcHic M, LECHENE C, GRANTHAM JJ: Renalepithelial cyst formation and enlargement in vitro: Dependence oncAMP. Proc Nail Acad Sci USA 86:6007—6011, 1989

40. MCATEER JA, EVAN AP, GARDNER KD: Morphogenetic clonalgrowth of kidney epithelial cell line MDCK. Anat Record 217:229—239, 1987

41. EVERSON GT, EMMETT M, BROWN WR, REDMOND P, THICKMAND: Functional similarities of hepatic cystic and biliary epithelium:Studies of fluid constituents and in vivo secretion in response tosecretin. Hepatology 11:557—565, 1990

42. FOLSCH UR, FISCHER H, SOLING H-D, CREUTZFELDT W: Effects

of gastrointestinal hormones and carbamylcholine on cAMP accu-mulation in isolated pancreatic duct fragments from the rat. Diges-tion 20:277—292, 1980

43. KAMINSKI DL, RUWART MJ,DESHPANDE YG: The role of cyclic

AMP in canine secretin-stimulated bile flow. J Surg Res 27:57—61,1979

44. AVNER ED, SWEENEYWE JR, YOUNG MC, ELLIS D: Congenitalmurine polycystic kidney disease. II. Pathogenesis of tubular cystformation. Pediatr Nephrol 2:210-218, 1988

45. AVNER ED, SWEENEY WE JR., ELLIS D: In vitro modulation oftubular cyst regression in murine polycystic kidney disease. Kidneyfnt 36:960—968, 1989

46. WILsoN PD, SHERWOOD AC, PALLA K, Du J, WATSON R, NOR-MAN JT: Reversed polarity of NaK-ATPase: Mislocation to

apical plasma membranes in human polycystic kidney diseaseepithelia. Am J Physiol 260:Fl—Fl 1, 1991

47. BERNSTEIN J: A classification of renal cysts, in The Cystic Kidney,edited by GARDNER KD itt, BERNSTEIN J, Dordrecht, KluwerAcademic Publishers, 1990, pp. 147—170

48. AVNER ED, MCATEER JA, EVAN AP: Models of cysts and cystickidneys, in The Cystic Kidney, edited by GARDNER KD JR, BERN-STEIN J, Dordrecht, Kluwer Academic Publishers, 1990, pp. 55—98

49. TORRES yE, YOUNG WF JR. OFFORD KP, HATTERY RR: Associ-

ation of hypokalemia, aldosteronism, and renal cysts. N EngI JMed 322:345—351, 1990

50. ISHIKAWA I: Acquired cystic disease, in The Cystic Kidney, editedby GARDNER KD JR. BERNSTEIN J, Dordrecht, Kiuwer AcademicPublishers, 1990, pp. 35 1—377

51. GABOW PA: Autosomal dominant polycystic kidney disease—Morethan a renal disease. Am J Kidney Dis I6:403—413, 1990

52. MILUTIN0vIc J, FIALKOW PJ, RUDD TG, AGODOA LY, PHILLIPSLA, BRYANT JI: Liver cysts in patients with autosomal dominantpolycystic kidney disease. Am J Med 68:741—744, 1980

53. GABOW PA, IKLE DW, HOLMES JH: Polycystic kidney disease:Prospective analysis of nonazotemic patients and family members.Ann Intern Med 101:238—247, 1984

54. HOSSACK KF, LEDDY CL, JOHNSON AM, SCHRIER RW, GABOWPA: Echocardiographic findings in autosomal dominant polycystickidney disease. N EngI J Med 3 19:907—912, 1988

55. LEIER CV, BAKER PB, KILMAN JW, WOOLEY CF: Cardiovascularabnormalities associated with adult polycystic kidney disease. AnnIntern Med 100:683—688, 1984

56. WAKABAYASHI T, FUJITA 5, OHBORA Y, SUYAMA T, TAMAKI N,MATSUMOTO 5: Polycystic kidney disease and intracranial aneu-rysms: Early angiographic diagnosis and early operation for theunruptured aneurysm. J Neurosurg 58:488—491, 1983

57. BROWN RAP: Polycystic disease of the kidneys and intracranialaneurysms. The etiology and interrelationship of these conditions:Review of recent literature and report of seven cases in which bothconditions coexisted. Glasgow Med J 32:333—348, 1951

58. SCHEFF RT, ZUCKERMANG, HARTER H, DELMEZ J, KOEHLER R:Diverticular disease in patients with chronic renal failure due topolycystic kidney disease. Ann Intern Med 92:202—204, 1980

59. MILUTINOVIC J, AGODOA UY: Potential causes and pathogenesis inautosomal dominant polycystic kidney disease. Nephron 33:139—144, 1983

60. CUPPAGE FE, HUSEMAN RA, CHAPMAN A, GRANTHAM ii: Ultra-

structure and function of cysts from human adult polycystic kid-neys. Kidney mt 17:372—381, 1980

61. OJEDA JL, Ros MA, ICARD0 JM, GARCIA-PORRERO JA: Basementmembrane alterations during development and regression of tubularcysts. Kidney Int 37:1270—1280, 1990

62. CARONE FA, MAKINO H, KANWAR YS: Basement membraneantigens in renal polycystic disease. Am J Pathol 130:466—471, 1988

63. CARONE FA, HOLLENBERG PF, NAKAMURA 5, PUNYARIT P.GLOGOWSKI W, FLOURET G: Tubular basement membrane changeoccurs pan passu with the development of cyst formation. KidneyInt 35:1034—1040, 1989

64. EBIHARA I, KILLEN PD, LAURIE GW, HUANG T, YAMADA Y,MARTIN GR, BROWN KS: Altered mRNA expression of basementmembrane components in a murine model of polycystic kidneydisease. Lab Invest 58:262—269, 1988

65. TAUB M, LAURIE GW, MARTIN GR, KLEINMAN HK: Alteredbasement membrane protein biosynthesis by primary cultures ofcpk/cpk mouse kidney. Kidney fin 37:1090—1097, 1990

66. GRANTHAM ii, DoNoso VS, EVAN AP, CARONE FA, GARDNER

996 Gabow: Polycystic kidney disease

KD JR: Viscoelastic properties of tubule basement membranes inexperimental renal cystic disease. Kidney mt 32:187—197, 1987

67. TOOLE BP: Glycosaminoglycans in morphogenesis, in Cell Biologyof Extracellular Matrix, edited by HAY ED. New York, PlenumPress, 1981, pp. 259—294

68. HAY ED: Cell Biology of Extracellular Matrix. New York, PlenumPress, 1981

69. LASH JE, VASAN NS: Somite chondrogenesis in vitro. Stimulationby exogenous extracellular matrix components. Dev Biol 66:151—171, 1978

70. BISSELL MJ, HALL HG, PANY G: How does the extracellularmatrix direct gene expression? J Theor Biol 99:31—68, 1982

71. WERDER AA, AMOS MA, NIELSEN AH, WOLFE GH: Comparativeeffects of germfree and ambient environments on the developmentof cystic kidney disease in CFWWJ mice. fLab Clin Med 103:399—407, 1984

72. GARDNER KD JR, REED WP, EVAN AP, ZEDALIS J, HYLARIDESMD, LEON AA: Endotoxin provocation of experimental renalcystic disease. Kidney mt 32:329—334, 1987

73. GARDNER KD JR, EVAN AP, REED WP: Accelerated renal cystdevelopment in deconditioned germ-free rats. Kidney mt 29:1116—1123. 1986

74. GARDNER KD JR, BURNSIDE JS, ELZINGA LW, LOCKSLEY RM:Cytokines in fluids from polycystic kidneys. Kidney ml 39:725—731,1991

75. MCDONALD ATJ, CROCKER JFS, DIGouT SC, MCCARTHY SC,BLECHER SR, COLE DEC: Glucocorticoid-induced polycystic kid-ney disease—A threshold trait. Kidney mt 37:901—908, 1990

76. KANWAR YS, CARONE FA: Reversible changes of tubular cell andbasement membrane in drug-induced renal cystic disease. Kidneymt 26:35—43, 1984