Cystic fibrosis and the phenotypic expression of autosomal dominant polycystic kidney disease

Cystic Fibrosis and the Phenotypic Expression of Autosomal Dominant Polycystic Kidney Disease Deirdre A. O’Sullivan, MD, Vincente E. Torres, MD, Patricia A. Gabow, MD, Stephen N. Thibodeau, PhD, Bernard F. King, MD, and Erik J. Bergstralh, MS ● Recent experiments in cultured cyst epithelial cells from kidneys of patients with autosomal dominant polycystic kidney disease (ADPKD) have shown that the cystic fibrosis (CF) transmembrane conductance regulator (CFTR) is present in the apical surface of these cells and mediates chloride (Cl–) and fluid secretion in vitro. To determine whether the presence of CF with the expression of mutated CFTR proteins modifies cyst formation in ADPKD, we studied a large family with both inherited diseases. ADPKD in this family is linked to PKD1. The family is composed of 26 members; 11 members with ADPKD, 4 members with CF, and 2 members with both diseases. Renal volumes measured by computerized tomography (CT), calculated creatinine clearances, and other clinical parameters in the family members with ADPKD and CF were compared with those in the family members with ADPKD alone, as well as to a large population of patients with ADPKD. The patients with CF and ADPKD, but not the CF heterozygote carriers with ADPKD, had less severe polycystic kidney and liver disease, as indicated by normal renal function; smaller renal volume, even when corrected for height and body surface area; and the absence of hypertension and liver cysts. These observations suggest that the coexistence of CF may reduce the severity of ADPKD. r 1998 by the National Kidney Foundation, Inc. INDEX WORDS: Autosomal dominant polycystic kidney disease; cystic fibrosis; cystic fibrosis transmembrane conductance regulator.

Editorial, p. 1081

A

UTOSOMAL DOMINANT polycystic kidney disease (ADPKD) is a common, dominantly inherited, genetically heterogeneous condition characterized by the formation of numerous cysts in the kidney and liver.1 ADPKD caused by mutations at the PKD1 locus is more severe than ADPKD not linked to PKD1.1 The growth of the renal cysts is associated with and believed to contribute to progressive deterioration in renal function.2 Cysts arise from the proliferation of tubular epithelial cells. The majority of cysts become disconnected from the tubular segments from which they originate, and the accumulation

From the Departments of Nephrology, Laboratory Medicine and Pathology, Diagnostic Radiology, and Biostatistics, Mayo Clinic and Mayo Foundation, Rochester, MN; and the Department of Medicine, University of Colorado Health Sciences Center, Denver, CO. Received June 3, 1998; accepted as submitted August 21, 1998. Support in part by grant no. DK44863 from the National Institutes of Health, Bethesda, MD. Address reprint requests to Vicente E. Torres, MD, Department of Nephrology, Mayo Clinic, Plummer 549, 200 First St SW, Rochester, MN 55905. E-mail: torres.vicente@mayo. edu

r 1998 by the National Kidney Foundation, Inc. 0272-6386/98/3206-0009$3.00/0 976

of fluid in these cysts occurs by transepithelial secretion of fluid into the cyst cavity.2 A series of recent experiments has identified that cystic fibrosis (CF) transmembrane conductance regulator (CFTR) is present on the apical surface of cultured ADPKD cells and mediates fluid secretion in vitro.3-5 CFTR functions primarily as a cyclic adenosine monophosphate (cAMP)-dependent chloride (Cl–) channel and is a member of the adenosine triphosphate–binding cassette transporter family.6 Mutations in the CF gene cause structural and functional abnormalities in CFTR, which result in decreased Cl– transport across the apical membrane of the secretory epithelia of the respiratory tract, small intestine, pancreas, and sweat glands. The impermeability of the apical membranes to Cl– in CF disrupts the normal transport of sodium chloride and water across epithelial membranes and leads to the protean manifestations of CF. We have identified a family with both genetic abnormalities who manifest characteristic phenotypic expression of both diseases and suggest that the coexistence of CF limits the severity of ADPKD type 1, as indicated by the absence of hypertension, normal creatinine clearance, and smaller renal volumes relative to age when compared with a large population of ADPKD patients.

American Journal of Kidney Diseases, Vol 32, No 6 (December), 1998: pp 976-983

CYSTIC FIBROSIS AND EXPRESSION OF ADPKD

METHODS

Genetic Analysis Blood for PKD1 linkage analysis and CF mutation analysis was collected in ethylenediaminetetraacetic acid. The DNA was isolated from the leukocytes using the ABI model 340A Nucleic Acid Extractor (Applied Biosystems Inc, Foster City, CA). The family in this study had been previously linked to the PKD1 locus.7 For this linkage analysis, DNA was digested with the appropriate restriction endonucleases and analyzed by Southern blot. Probes used included alpha 38 HVR, DH7, CMM65, and 24.1.8 Allele sizes for each probe/enzyme combination were recorded and linkage analysis was performed for all two-point probe combinations using the programs MLINK and ILINK.9 The agedependent penetrant classes used were 0.32, 0.72, and 0.90 for individuals aged 15 to 20, 20 to 30, and older than 30 years, respectively. The CF mutation analysis was performed at Genzyme Genetics (Framingham, MA). Regions of the CFTR gene were amplified enzymatically and hybridized to specific CF mutation oligonucleotide probes. Positive results were tested for specific mutation identity.

Assessment of Renal Volume Computerized tomographic (CT) scans of the abdomen performed for clinical purposes were identified for all family members with ADPKD, and the age at which these studies were performed was noted. Renal size was determined by manual digitization of CT images of the kidney. Renal volumes were calculated from this data using the Bioquant program (R & M Biometrics, Inc, Austin, TX). Renal volumes were then corrected for height (170 cm) and body surface area (1.73 m2). Liver cysts were noted if there was one or more cysts in the hepatic parenchyma as identified by CT scan.

Collection of Clinical Data Outpatient and inpatient records were reviewed for measurements of blood pressure, serum creatinine levels, glomerular filtration rate (GFR) by iothalamate or creatinine

Fig 1. Pedigree of a family with coexisting ADPKD and CF.

977

clearance, and age at which these measurements were made. Hypertension was defined as a resting systolic blood pressure greater than 145 mm Hg or a resting diastolic blood pressure greater than 90 mm Hg on three separate occasions before the onset of end-stage renal failure and the initiation of renal replacement therapy or the initiation of antihypertensive medication.

Statistical Analysis Renal size predicted values and percentiles for calculated creatinine clearance as a function of age were estimated using linear regression from data collected on a large population of patients with ADPKD seen at the University of Colorado Health Sciences Center.10 Patients with known ADPKD type 2 were excluded from this data set; therefore, this population consisted mainly of ADPKD type 1 patients. Renal volumes in these patients were measured by ultrasonography using the formula for a modified ellipse: 4/3 ␲ (length/2)(anterior-posterior diameter/4 ⫹ width/4). Creatinine clearances were calculated using the Cockcroft-Gault formula.11 Renal volume versus age and creatinine clearance versus age for members of this family with CF were compared graphically with family members without CF and to the Colorado data. Two sample t-tests using log10 of renal volume and calculated creatinine clearance and z statistics (renal volume or calculated creatinine clearance minus agespecific predicted values from the University of Colorado ADPKD 1 data/standard deviation [renal volume or creatinine clearance]) for each were used to compare the family members with ADPKD and CF and the family members with ADPKD alone. All tests were two-sided with alpha level equal to 0.05.

RESULTS

The identified family is composed of 26 members, 11 members with ADPKD, 4 members with CF, and 2 members with both genetic diseases (Fig 1). Of the 11 members with ADPKD, only 10 had medical records available for review. As

978

O’SULLIVAN ET AL

members were hypertensive and 3 members were normotensive. Of the three subjects without hypertension, two had both ADPKD and CF (Table 1). Seven of the 10 affected family members had liver cysts and 3 members did not. Of the three family members without liver cysts, two members had coexistent ADPKD and CF. Six family members had reduced GFR by iothalamate clearance or serum creatinine measurement, all with ADPKD alone. Three of the 10 family members with ADPKD had a single CFTR mutation. All three patients had hypertension, liver cysts, and reduced GFR (Table 1). Renal volumes were smaller for family members with coexistent CF and ADPKD compared with kindred with ADPKD alone, even when corrected for body surface area (469 and 672 mL/1.73 m2 compared with 987 to 4,578 mL/ 1.73 m2; Fig 3 and Table 2). When comparing the log10 of renal volume, family members with CF had a mean value of 2.77 mL (range, 2.68 to 2.87 mL) versus family members without CF who had a mean value of 3.33 mL (range, 2.98 to 3.72 mL). This difference was of borderline statistical significance (P ⫽ 0.040). When compared with a large population of predominantly type 1 ADPKD patients, the family members with CF had smaller renal volumes for age (mean z score, –0.75; range, –1.31 to –0.18) and family members without CF had greater renal volumes for age (mean z score, 0.81; range, –0.40 to 2.09; P ⫽ 0.058). The imaging studies of the two patients with ADPKD and CF are compared with those of

Fig 2. Domain structure of the CFTR indicating the location of the mutations detected in these families. Modified from Sheppard and Ostedgaard.6

part of a previous study, this family had been found to have ADPKD type 1 (lod score 2.38, ␪ ⫽ 0 with 38 HVR). Using polymerase chain reaction–based assays, this family was found to bear three mutations for CF: ⌬F508, a mutation in the nucleotide-binding domain of CFTR that results in abnormal cellular processing; E60X, which encodes for a stop codon at amino acid 60; and 3849 ⫹ 10kb C⬎T, a splice mutation at the 58 end of exon 19 (Fig 2). Clinically, patients III-14 and III-16, with ⌬F508 and E60X mutations, had pancreatic insufficiency, infertility, and moderate pulmonary symptoms requiring daily bronchopulmonary drainage and frequent albuterol nebulization. Patient IV-3, with ⌬F508 and 3849 ⫹10kb C⬎T mutations, had unlimiting pulmonary symptoms and required no pancreatic enzyme replacement. Patient III-11 died at the age of 2 years of complications of CF and the genotype for CF was not determined but presumed to be ⌬F508/E60X. Of the 10 family members with ADPKD, 7

Table 1. Family Members With ADPKD, CF Genotype, and Clinical Manifestations of ADPKD

Genotype

Patient

⌬F508/E60x III-16 ⌬F508/3849 ⫹ 10kb C ⬎ T IV-3 ⌬F508/N III-7 III-10 E60x/N II-3 N/N II-2 III-3 III-9 III-12 IV-1

Sex and Age

HTN S Creat Reduced Renal Renal Renal (age of Liver Elevation GFR (age Age Height Weight Volume Volume Volume detection) Cysts (age) measured) at CT (cm) (kg) (mL) (mL/170 cm) (mL/1.73 m2)

M30

N

N

N

**

27

178

75

738

705

672

F32 F45 M49 M65 F72 F46 F45 F43 M24

N Y (42) Y (38) Y (60) Y (40) Y (40) N Y (38) Y (24)

N Y Y Y Y Y N Y Y

N Y (42) Y (35) Y (48) Y (43) N N N Y (20)

** Y (42) Y (43) Y (58) Y (65) Y (44) ** ** **

31 38 41 61 70 40 41 37 24

170 162 183 177 163 166 165 167 185

67 73 90 82 74 55 66 62 73

480 1542 5282 2907 4737 1049 3142 964 1128

480 1618 4906 2792 4956 1074 3238 981 1036

469 1507 4250 2647 4578 1148 3124 987 1001

Abbreviations: HTN, hypertension; S Creat, serum creatinine level.

CYSTIC FIBROSIS AND EXPRESSION OF ADPKD

979

Fig 3. Plot of the log of renal size v age for the members of this family with ADPKD alone (䉰) or coexisting ADPKD and CF (䉳) compared with a large population of ADPKD type 1 patients seen at the University of Colorado Health Sciences Center (dots). Lines correspond to percentiles (2.5, 5.0, 50.0, 95.0, 97.5) estimated from the Colorado data.

DISCUSSION

family members of the same sex and similar ages with ADPKD alone in Figs 4 and 5. When comparing calculated creatinine clearances, the family members with CF and ADPKD had a mean creatinine clearance of 113 versus 49 mL/min/1.73 m2 in the family members with ADPKD alone (P ⫽ 0.039; Fig 6 and Table 3). When compared with a large population of predominantly type 1 ADPKD patients, family members with combined disease appeared to have slightly higher creatinine clearances than expected for age (mean z score, 0.65; range, 0.04 to 1.26). Family members with ADPKD alone on average had lower than expected creatinine clearance (mean z score, –0.63; range, –1.85 to 0.21; P ⫽ 0.051). Contrary to patients with ADPKD and CF, the renal volumes and creatinine clearances in the ADPKD patients with only one CFTR mutation were not different from those observed in the family members with ADPKD and no CFTR mutations (Table 1).

CFTR protein has been immunolocalized to the proximal tubule and thin limb of Henle’s loop and to the apical membrane region of the distal tubule, principal cells of cortical collecting duct, and inner medullary collecting duct.12-14 Studies using mouse collecting duct cell lines have shown that Cl– secretion in these cells involves a cAMPactivated efflux of Cl– that crosses the apical membrane through CFTR channels.14-16 Despite some published observations of hypercalciuria and microscopic nephrocalcinosis, decreased ability to excrete a sodium chloride load, decreased ability to dilute and concentrate the urine, and altered excretion of certain drugs,14,17 other studies have not been able to show any alterations of tubular electrolyte handling under basal conditions in CF patients.18,19 Because of the lack of gross abnormalities in renal function in patients with CF, it has been suggested that the function of CFTR may be complemented by other renal

Table 2. Comparison of Renal Volumes in the ADPKD Patients of This Family Without or With CF No CF (n ⫽ 8)

CF (n ⫽ 2)

Renal Volume

Mean

SD

Range

Mean

SD

Range

P*

Log10 scale z statistics (deviation from predicted, SD units)

3.33 0.81

0.30 0.90

2.98-3.72 ⫺0.40-2.09

2.77 ⫺0.75

0.13 0.80

2.68-2.87 ⫺1.31-⫺0.18

0.040 0.058

Abbreviation: SD, standard deviation. *A two-sample t-test using the log of renal volume and z statistics (renal volume–predicted renal volume from the University of Colorado PKD data/SD [RV]) was used to compare the 2 ADPKD patients with CF and the 8 ADPKD patients without CF.

980

O’SULLIVAN ET AL

Fig 4. CT scans of the abdomen in a 27-year-old man with ADPKD and CF (III-16, left panels) and in a 24-year-old male relative with ADPKD alone (IV-1, right panels). There are fewer cysts and the sizes of the kidneys and the cysts are smaller in the patient with ADPKD and CF. A few small liver cysts can be seen in the posterior aspect of the right lobe of the liver in the patient with ADPKD alone. Additional liver cysts were detected in other CT sections in this patient only (not shown).

Cl– channels that overcome the functional consequences of mutant CFTR in an otherwise healthy kidney.19,20 In ADPKD, cyst formation is responsible for the renal enlargement and likely contributes to the progressive renal failure encountered in this disease. The cysts arise from focal dilatation of the renal tubules. They then gradually dissociate from the parent tubule and eventually become isolated, fluid-filled sacs. Accumulation of fluid within the cysts occurs by transepithelial secretion.2 Studies using cell monolayers and intact cysts from human ADPKD kidneys have shown that sodium and potassium–adenosinetriphosphatase is present on the basolateral surface of cyst epithelial cells and establishes the electrochemical gradient responsible for secondary transport mechanisms driving fluid secretion.21 It was then established that Cl– secretion is the secondary transport mechanism responsible for fluid secre-

tion into cysts and that it is cAMP-mediated.22 Subsequent studies showed the immunolocalization of CFTR to the apical membrane of the epithelial cells lining the cysts, the translocation of CFTR from an intracellular pool to the apical plasma membrane induced by cAMP, and the inhibition of the cAMP activation of Cl– secretion in the cyst-derived cells by antisense oligonucleotides to CFTR. These observations strongly support the view that CFTR is the Cl– channel responsible for active Cl– and fluid secretion in the cysts.3-5 The observations in a family with ADPKD and CF reported here are consistent with a role for CFTR in the pathogenesis of ADPKD. The present study suggests that the coexistence of ADPKD and CF limits the clinical severity of ADPKD. The two family members with coexistent diseases had smaller renal volumes and higher calculated creatinine clearances relative to fam-

CYSTIC FIBROSIS AND EXPRESSION OF ADPKD

Fig 5. T-1 weighted magnetic resonance imaging of the abdomen with fat saturation in a 31year-old woman with ADPKD and CF (IV-3, left panels) and CT scan of the abdomen in a 37-year-old female relative with ADPKD alone (III-12, right panels). There are fewer renal cysts and the sizes of the kidneys and the cysts are smaller in the patient with ADPKD and CF. Liver cysts were detected only in the patient with ADPKD alone.

Fig 6. Plot of the calculated creatinine clearances v age for the members of this family with ADPKD alone (䉰) or ADPKD coexisting with CF (䉳) compared with a large population of ADPKD type 1 patients seen at the University of Colorado Health Sciences Center (dots). Lines correspond to percentiles (2.5, 5.0, 50.0, 95.0, 97.5) estimated from the Colorado data.

981

982

O’SULLIVAN ET AL

Table 3. Comparison of Calculated Creatinine Clearance in ADPKD Patients of This Family Without or With CF No CF (n ⫽ 8)

CF (n ⫽ 2)

Creatinine Clearance

Mean

SD

Range

Mean

SD

Range

P*

Calculated value (ml/min) z statistics (deviation from predicted, SD units)

49 ⫺0.63

34 0.68

7-97 ⫺1.85-0.21

113 0.65

22 0.86

97-128 0.04-1.26

0.039 0.051

*A two-sample t-test using calculated creatinine clearance and z statistics (calculated creatinine clearance–predicted creatinine clearance from the University of Colorado PKD data/SD [RV]) was used to compare the 2 ADPKD patients with CF and the 8 ADPKD patients without CF.

ily members with ADPKD alone and relative to a large population of predominantly type 1 ADPKD patients, even when corrected for age. The onset of hypertension before the age of 35 years has been established as an independent risk factor for the progression of renal failure in ADPKD.10 Both family members with CF and ADPKD had normal blood pressures without the need for antihypertensive medication at the time of the study compared with only two of the eight family members with ADPKD alone. This, however, could reflect the fact that the two family members with coexistent diseases are younger than the other family members (aged 30 and 32 years). The mean age of diagnosis of hypertension in the family members with ADPKD alone was 42 years (range, 24 to 60 years). Liver cysts are the most common extrarenal manifestation of ADPKD and the severity of liver cyst formation correlates with that of ADPKD. The frequency of hepatic cysts increases with age, from 20% of patients in the third decade to approximately 75% in the seventh decade.23 Women develop more cysts at an earlier age than men, and women who have multiple pregnancies or who have used oral contraceptive agents or estrogen replacement therapy postmenopausally have worse disease.24 Liver cysts were noted to be absent in both family members with CF and ADPKD, even in the woman who had used oral contraceptive agents chronically for more than 14 years and who had one pregnancy. Seven of the eight family members with ADPKD alone had liver cysts. Despite the normal renal function, the absence of hypertension and liver cysts, and the smaller renal volumes, the presence of CF is not entirely protective against the renal manifestations of ADPKD. These patients still form renal cysts, albeit they are smaller and fewer than in family

members with ADPKD alone. This is not unexpected because CFTR mutants can reach the apical membrane to variable extents and form Cl– channels with altered properties and because other Cl– channels in renal tubular epithelial cells20 may compensate for the loss of CFTR. Finally, no protective effect was detected in patients with ADPKD with only one CFTR mutation. Whether CF heterozygote carriers show any phenotypic manifestations of the disease has been a controversial subject. Nevertheless, carefully performed studies have shown that, when controlling for age, CF carriers with the ⌬F508 mutations have significantly increased sweat electrolyte concentrations.25 Whereas an effect of a single CFTR mutation on the expression of ADPKD cannot be excluded because of the small number of patients, the observations in this study indicate that this effect is minor or nonexistent. ACKNOWLEDGMENT Ryan Bolduan provided programming support and Pat Urban provided secretarial support.

REFERENCES 1. Gabow PA: Definition and natural history of autosomal dominant polycystic kidney disease, in Watson ML, Torres VE (eds): Polycystic Kidney Disease. New York, NY, Oxford, 1996, pp 333-355 2. Grantham JJ: Fluid secretion, cellular proliferation, and pathogenesis of renal epithelial cysts. J Am Soc Nephrol 3:1843-1857, 1993 3. Brill SR, Ross KE, Davidow CJ, Ye M, Grantham JJ, Caplan MJ: Immunolocalization of ion transport proteins in human autosomal dominant polycystic kidney epithelial cells. Proc Natl Acad Sci U S A 93:10206-10211, 1996 4. Davidow CJ, Maser RL, Rome LA, Calvet JP, Grantham JJ: The cystic fibrosis transmembrane conductance regulator mediates transepithelial fluid secretion by human autosomal dominant polycystic kidney disease epithelium in vitro. Kidney Int 50:208-218, 1996 5. Hanaoka K, Devuyst O, Schwiebert EM, Wilson PD, Guggino WB: A role for CFTR in human autosomal domi-

CYSTIC FIBROSIS AND EXPRESSION OF ADPKD

nant polycystic kidney disease. Am J Physiol 270:C389-399, 1996 6. Sheppard DN, Ostedgaard LS: Understanding how cystic fibrosis mutations cause a loss of Cl– channel function. Mol Med Today 290-297, 1996 7. Dawson DB, Torres VE, Charboneau JW, Thibodeau SN: Detection of a family with autosomal dominant polycystic kidney disease loosely linked to DNA markers from 16p. The American Society of Nephrology 21st Annual Meeting, San Antonio, TX, 1988 8. Breuning MH, Snijdewint FG, Brunner H, Verwest A, Ijdo JW, Saris JJ, Dauwerse JG, Blonden L, Keith T, Callen DF, Hyland VJ, Xiao GH, Scherer G, Higgs DR, Harris P, Bachner L, Reeders ST, Germino G, Pearson TL, Van Ommen GJB: Map of 16 polymorphic loci on the short arm of chromosome 16 close to the polycystic kidney disease gene (PKD1). J Med Genet 27:603-613, 1990 9. Lathrop GM, Lalouel JM: Efficient computations in multilocus linkage analysis. Am J Hum Genet 42:498-505, 1988 10. Johnson AM, Gabow PA: Identification of patients with autosomal dominant polycystic kidney disease at highest risk for end-stage renal disease. J Am Soc Nephrol 8:1560-1567, 1997 11. Cockroft DW, Gault MH: Prediction of creatinine clearance from serum creatinine. Nephron 16:31, 1976 12. Crawford I, Maloney PC, Zeitlin PL, Guggino WB, Hyde SC, Turley H, Gatter KC, Harris A, Higgins CF: Immunocytochemical localization of the cystic fibrosis gene product CFTR. Proc Natl Acad Sci U S A, 88:9262-9266, 1991 13. Morales MM, Carroll TP, Morita T, Schwiebert EM, Devuyst O, Wilson PD, Lopes AG, Stanton BA, Dietz HC, Cutting GR, Guggino WB: Both the wild type and a functional isoform of CFTR are expressed in kidney. Am J Physiol 270:F1038-F1048, 1996 14. Stanton BA: Cystic fibrosis transmembrane conductance regulator (CFTR) and renal function. Wien Klin Wochenschr 109:457-464, 1997

983

15. Vandorpe D, Kizer N, Ciampollilo F, Moyer B, Karlson K, Guggino WB, Stanton BA: CFTR mediates electrogenic chloride secretion in mouse inner medullary collecting duct (mIMCD-K2) cells. Am J Physiol 269:C683-C689, 1995 16. Letz B, Korbmacher C: cAMP stimulates CFTR-like Cl– channels and inhibits amiloride-sensitive Na⫹ channels in mouse CCD cells. Am J Physiol 272:C657-C666, 1997 17. Katz SM, Krueger LJ, Falkner B: Microscopic nephrocalcinosis in cystic fibrosis. N Engl J Med 319:263-266, 1988 18. Bentur L, Kerem E, Couper R, Baumal R, Canny G, Durie P, Levison H: Renal calcium handling in cystic fibrosis: Lack of evidence for a primary renal defect. J Pediatr 116:556-560, 1990 19. Windstetter D, Schaefer F, Scha¨rer K, Reiter K, Eife R, Harms HK, Bertele-Harms R-M, Fiedler F, Tsui LC, Reitmeir P, Horster M, Hadorn HB: Renal function and renotropic effects of secretin in cystic fibrosis. Eur J Med Res 2:131-436, 1997 20. Jentsch TJ: Chloride channels: A molecular perspective. Curr Opin Neurobiol 6:303-310, 1996 21. Grantham JJ, Ye M, Gattone VH II, Sullivan LP. In vitro fluid secretion by epithelium from polycystic kidneys. J Clin Invest 95:195-202, 1995 22. Wallace DP, Grantham JJ, Sullivan LP: Chloride and fluid secretion by cultured human polycystic kidney cells. Kidney Int 50:1327-1336, 1996 23. Torres VE: Polycystic liver disease, in Watson ML, Torres VE (eds): Polycystic Kidney Disease. New York, NY, Oxford, 1996, pp 500-529 24. Gabow PA, Johnson AM, Kaehny WD, MancoJohnson ML, Duley IT, Everson GT: Risk factors for the development of hepatic cysts in autosomal dominant polycystic kidney disease. Hepatology 11:1033-1037, 1990 25. Farrell PM, Koscik RE: Sweat chloride concentrations in infants homozygous or heterozygous for F508 cystic fibrosis. Pediatrics 97:524-528, 1996