Kidney International, Vol. 15 (1979), PP. 7-19

Nephron obstruction in nordihydroguaiaretic acid-inducedrenal cystic disease

ANDREW P. EVAN and KENNETH D. GARDNER, JR.

Departments ofAnatomy and Medicine, University of New Mexico School of Medicine, Albuquerque, New Mexico

Nephron obstruction in nordihydroguaiaretic acid-induced ren-al cystic disease. Studies were performed to characterize condi-tions in rat kidneys whose nephrons were made cystic by feeding2% nordihydroguaiaretic acid (NDGA) to the animals. Using twomicropipettes, we monitored intratubular hydrostatic pressureswhile perfusing single surface nephrons in NDGA-exposed (5 to7 weeks) and normal rat kidneys. The introduction of 50 nI ofRingers solution labeled with 3H-inulin at a flow rate of 25 nllminwas associated with a significant mean (± SEM) increase (16761%; P < 0.02) in pressure in cystic but not in nondilated (—0.5

27.2%) or normal (31 23%)nephrons, respectively. The rela-tive amount of 3H-inulin excreted in 40 mm from cystic (4.02.0%) was less than that excreted from either nondilated (197%; P < 0.05) or normal (105 26%; P < 0.01) nephrons. Intra-lumenal pressures in nondilated but not other nephron groupscorrelated with urinary flow rates (r = 0.51; P < 0.02). Singlenephron filtration rates and tubular-fluid-to-plasma 3H-inulin ra-tios in additional rats were similar among all groups of tubules.Concluding that these data reflected increased resistance to out-flow from cystic nephrons, we examined these and additionalNDGA-exposed (1 to 24 weeks) kidneys. 3H-thymidine radio-autography demonstrated maximum collecting tubular cell hy-perplasia (13% labeling) at 2 to 3 weeks of NDGA-exposure. Mi-croscopy and microdissection demonstrated tiny mural polypsalong outer medullary segments of collecting tubules. Thirteentubules were traced to their outlets; polyps impinging on outflowlumens were found in all 13 instances. We conclude that partialnephron obstruction exists in NDGA-exposed kidneys and thatobstruction is a likely contributor to cyst formation in this model.

Obstruction nephronique dans Ia maladie kystique du rein dé-terminée par l'acide nordihydroguaiarétique Cette étude a portesur des reins de rats dont los néphrons ont été rendus kystiquespar l'addition a l'alimentation d'acide nordihydroguaiarétique(NDGA) a 2%. Au moyen de deux micropipettes les pressionshydrostatiques intratubulaires ont été enregistrées pendant laperfusion de néphrons superficiels chez des rats exposés aNDGA (5 a 7 semaines) et des rats normaux. L'introduction de50 nl de Ringer contenant de l'inuline 3H au debit de 25 nI/mm estassociée a une augmentation significative (167 61%; P < 0,02)de la pression dans les néphrons kystiques mais pas dans lesndphrons intacts (—0,5 27,2%) ou normaux (31 23%). Laquantité relative d'inuline 3H excrétée en 40 mm par les néph-rons kystiques (4,0 2,0%)est inférieure a celle excrétée par lesnéphrons non dilates (19 7%; P < 0,05) ou normaux (10526%; P < 0,01). Les pressions luminales des néphrons non di-latés, mais non celles des autres groupes, sont correlées aux dé-bits urinaires (r = 0,51; P < 0,02). Les debits de filtration glom-érulaires des néphrons individuels et le rapport de concentrationde l'inuline tube/plasma, étudiés chez d'autres rats, sont scm-blables pour tous les groupes de tubules. Puisque la conclusionétait que ces observations reflètent une augmentation de Ia résis-

7

tance a l'écoulement dans les néphrons kystiques nous avonsétudié d'autres reins exposés a NDGA pendant I a 24 semaines.Des autoradiographies avec de Ia thymidine 3H ont montré unehyperplasie maximale (13% de marquage) des cellules tubulairesdes collecteurs entre 2 et 3 semaines d'exposition a NDGA. Lamicroscopie et Ia microdissection ont montré des petits polypesmuraux Ic long des segments médullaires externes des tubes col-lecteurs. Treize tubules ont été suivis jusqu'à leurs terminaisonset dans les 13 cas des polypes génant l'écoulement ont ététrouvés. Nous concluons qu'il existe une obstruction nephro-nique partielle dans les reins exposés a NDGA et que cette ob-struction contribue vraisemblablement a la formation de kystesdans ce modèle.

A defect in the elastic properties of tubular base-ment membrane has been cited as a major factor inthe pathogenesis of drug-induced renal cystic diseasein rats [1]. Carone et a! reached this conclusion afterdemonstrations that glomerular function, intra-tubular pressures, and solute and water reabsorp-tion were normal in cystic kidneys of rats fed di-phenyl thiazole (DPT). Subsequently, we reportedstudies of similar parameters in rat kidneys madecystic by exposure to diphenylamine (DPA) [21. Inthat study, we recorded increased intratubular pres-sures in, and delayed 3H-inulin excretion from, di-lated and presumably precystic nephrons. We con-sidered partial obstruction to be the better ex-planation for these findings and were reticent toaccept altered compliance of nephron walls as anoperative mechanism in every form of drug-inducedrenal cystic disease.

The study to be described here examined condi-tions in a third variety of drug-induced renal cysticdisease in rats. The provocative agent was nordihy-droguaiaretic acid (NDGA), an antioxidant known

Received for publication January 23, 1978and in revised form April 10, 1978.

0085—2538/79/0015-0007 $03.80© 1979 by the International Society of Nephrology

8 Evan and Gardner

Fig. 1. Structural formulae of diphenylamine, diphenyl thiazole,and nordihydroguaiaretic acid. Cystic disease induced by eachof these antioxidants in rats has been studied structurally andfunctionally [1—3, 12].

to induce kidney cysts when fed to rats [3], and it issimilar in structure to both DPA and DPT (Fig. 1).

Nephrons in NDGA-exposed and normal rat kid-neys were perfused with known microvolumes ofisotopically labeled fluid while pressure changeswere monitored in the same nephrons. Pressure in-creases were significant with microperfusion only incystic nephrons of NDGA-exposed kidneys. Wehave concluded that partial obstruction, in prefer-ence to increased filtration, decreased reabsorption,and decreased mural compliance, is the likely ex-planation for this finding. Its cause appears to befocal cellular hyperplasia (polyp formation), whichdevelops along outer medullary segments of therenal collecting tubules in NDGA-exposed animals.

Methods

Male Sprague-Dawley rats initially weighing 100to 125 g were used in all studies. Animals weremaintained under standard laboratory conditionsand were fed a chow diet (Wayne Lab Blocks; Tek-lad Division of Mogul Corp., Madison, Wisc.) ei-ther alone (control group) or with 2.0% by weightNDGA added (experimental group) for periods of 1through 24 weeks.

Functional studies

Functional studies were performed on controlrats and on rats exposed to NDGA for 5 to 7 weeks.After overnight food deprivation with free access towater, animals were anesthetized with mactin (Pro-monta, Hamburg, West Germany), 100 mg/kg ofbody weight, i.p. Through a midline abdominal in-cision, ureteral catheterizations were performed,and the left kidney was mounted for micropuncturestudies, as described previously [2]. Ringers lactatesolution containing 5% mannitol was infusedthrough intrajugular venous catheter at 1 mI/hr foreach 100 g of body weight or fraction thereof for 60to 90 mm before and during each experiment. Diam-eters of surface nephrons were measured with a fi-lar micrometer eyepiece (Bausch and Lomb, Inc.,Rochester, N.Y.) and were expressed in microns.Cortical nephrons exceeding 50 in diameter wereclassified as cystic; those with diameters under 35 pwere considered nondilated. Intermediate nephronswere not studied.

Intranephron hydrostatic pressures were mon-itored with a servonulling device [21. Pressureswere recorded constantly during initial and experi-mental periods. For the purpose of data analyses,each period was divided into five equal intervals;mean pressures were calculated from readings atthose points that were 0, 20, 40, 60, 80, and 100% ofthe total duration of the initial and experimental pe-riods, respectively.

Animals were selected randomly for inclusion inone of two study groups. In the first, whole kidneyand single nephron glornerular filtration rates(SNGFR) were derived with 3H-inulin (New Eng-land Nuclear Corp., Boston, Mass.) that was addedto the parenteral Ringers infusate in a concentrationof 25 Ci/ml. Timed urine specimens were collectedfrom a left ureteral catheter, their volumes deter-mined by weight, and their isotopic activities deter-mined by scintillation counting as described pre-viously [2].

In the second group of animals, unlabeled Ring-ers solution was infused, and recovery of 3H-inulinafter single nephron ,nicroperfusion was accom-plished as follows: After insertion of the servo-nulling pressure pipette containing 1% lissaminegreen in saline, a small amount of the colored solu-tion was allowed to enter the nephron and to flowdownstream. The second or third convolution distalto the puncture site was identified and the micro-perfusion pipette was inserted at this point. Using aprecalibrated microperfusion pump that was ca-

Diphenylamine

Diphenyl thiazole

N—CII IIC C

S

--/H2N

Nordihydroguaiaretic acid

H3C CH3

HO_._ C — — C - C OH

NDGA-induced renal cystic disease 9

pable of delivering accurate amounts of fluid againsthydrostatic pressure resistances as great as 60 cm ofwater [2] (Wolfgang Hampel, Berlin, West Germa-ny), we perfused into individual nephrons over a 2-mm period at 25 nI/mm known amounts of Ringerssolution made visible with 1% lissamine green andcontaining 3H-inulin in a concentration sufficient toyield 85 counts/nllmin (approximately three timesbackground). Consecutive 10-mm urine collectionswere begun at the time the pressure monitoring pi-pette was inserted and were continued after eachperfusion until radioactivity in urine from both kid-neys returned to preperfusion baselines of radio-activity or until at least four collection periods (40mm) had elapsed.

3H-inulin recovery was expressed as a percent ofcounts recovered in ureteral urine from the left (per-fused) and right (nonperfused) kidneys relative tothe total amount of 3H-inulin injected.

Statistical analyses were performed using Stu-dent's r test to establish the significance of dif-ference between means [4]. Because of the relative-ly wide ranges that sometimes were encountered inabsolute numerical values for some parameters, forexample, intralumenal pressures, data that did notachieve significance with the t test were analyzedby the Mann-Whitney U test to establish whetherdifferences existed in the numerical rank order ofobservations between various nephron groups [4].

Structural studies

In preparation of the renal tissue for morphologicanalysis, kidneys from animals used in the abovestudies and additional animals that received NDGAfor 1 to 24 weeks were studied. After anesthesiawas given, the thoracic cage was opened, and an 18-gauge needle attached to a reservoir was insertedinto the left ventricle. Initially, the animal was per-fused with 50 ml of 0.9% sodium chloride containing0.1% sodium procaine to flush out the blood. Thiswas followed by 100 ml of 2.5% glutaraldehyde in0.15 M cacodylate/hydrochloric-acid buffer (pH,7.3) at room temperature for 10 mm. The right atri-um was opened to allow the perfusate to circulatethroughout the vascular system. The kidneys wereremoved, cut in half, and further fixed in the origi-nal fixative for at least 24 hours.

Light microscopy. Half of each kidney was pro-cessed routinely for light microscopy and embed-ded in paraffin. Sections were cut 5-ji thick andstained either with periodic acid-Schiff(PAS) or he-matoxylin-orange G.

Transmission electron microscopy (TEM). Por-tions of the cortex and medulla were minced into 1-mm2 blocks. The tissue then was washed in 0.075 Mcacodylate/hydrochloric-acid buffer for 1 hour at18° C, postosmicated in 1% osmium oxide at 4° C in0.075 M cacodylate/hydrochioric-acid buffer for 1hour, dehydrated through a series of graded eth-anols, passed through two changes of propyleneoxide, and flat embedded in Epon 812. One-micronsections were cut and stained with toluidine blue,whereas thin sections were stained with uranyl ace-tate and lead citrate. Sections were examined with aPhilips 200 EM.

Scanning electron microscopy (SEM). A portionof each kidney was washed in 0.075 M cacodylate!hydrochloric-acid buffer for 1 hour, postosmicatedin 1% osmium oxide in a 0.075 M cacodylate/hydro-chloric-acid buffer for 1 hour, and dehydratedthrough a series of graded ethanols. To examine tis-sue that was not cut by a razor blade, we fracturedthe tissue according to the following procedure.First, the tissue was placed in a sac containing 100%ethanol. The whole sac then was immersed in liquidnitrogen until the tissue fragments were frozen com-pletely; then it was fractured with a hammer andchisel. The fractured tissue was transferred in fresh100% ethanol to a Samdri critical-point dryer andwas dried with liquid carbon dioxide. The driedspecimens were secured to aluminum slugs bydouble stick tape and placed in a vacuum-evapora-tor with a rotating stage for coating with gold-palla-dium. Samples were examined and photographed inan ETEC Autoscan operating at an acceleratingvoltage of 10 kV.

Microdissection. One-centimeter cubed pieces oftissue, including cortex and medulla, were rinsedseveral times in 0.1 M phosphate buffer to removethe fixative and then placed in 8 N hydrochloric acidsolution for 50 to 70 mm at 60° C. The specimenswere rinsed several times in distilled water, andsingle nephrons were microdissected free withsharpened needles and were photographed with astereomicroscope (Wild M5).

Autoradio graph)'. To determine if hypertrophywas a mechanism of cyst formation, we gave NDGA-treated and control animals an i.p. injection of ra-dioactive thymidine (specific activity, 15.2 Ci/mM)at a dose of 0.2 Ci/g of body weight 45 mm beforethe animal was whole-body perfused. The kidneyswere fixed as previously described and were pre-pared for light microscopy. The kidneys were em-bedded in paraffin and cut at 4-j.. The sections were

10 Evan and Gardner

Table 1. Renal function in nordihydroguaiaretic acid (NDGA)-exposed and normal rat kidneys (mean SEM)a

NDGA-exposed

NormalCystic Nondilated

SNGFRt,nlImin 26.5 6.1 (6) 28.0 4.1 (15) 31.9 4.7 (20)TF/Pinulin 3.34 1.27 (7) 2.51 0.83 (15) 2.17 0.20 (20)Intranephronpressure,cm water 15.5 2.3 (22) 16.9 1.9 (28) 13.8 1.4 (30)

Parentheses contain the numbers of observations. SNGFR is single nephron glomerular filtration rate; TF/P inulin is tubular-fluid-to-plasma-inulin ratio.

b Whole kidney GFR in normal rats was 3.41 0.28 (14), and in NDGA-exposed rats it was 1.93 0.39 (9) (P < 0.02, by t test fordifference from mean value in normal rats).

deparaffinized, dipped in NTB2 nuclear tract emul-sion, and stained with hemotoxylin and eosin.Labeling indices were determined for each regionof the kidney, based on a ratio of the number ofcell nuclei demonstrating thymidine uptake to 500cells counted, multiplied times 100.

Results

Functional studies. Experimental animals weighed240 to 310 g at the time of study, a range not sig-nificantly different from that of controls.

Table 1 presents data on whole kidney andSNGFRs, and intralumenal pressures in nephronsof NDGA-exposed and control rat kidneys. Wholekidney GFR was depressed significantly amongNDGA-exposed rats. No significant differences ex-isted between cystic and nondilated surface neph-rons or between NDGA-exposed and control kid-neys in their mean SNGFRs, tubular-fluid-to-plasma (TF/P) inulin ratios, or intranephron hydro-static pressures.

The excretion of 3H-inulin following its micro-injection into individual nephrons was depressedsignificantly and was delayed (a) from NDGA-ex-posed kidneys as compared to controls, and (b)from cystic as opposed to nondilated nephrons in

the same NDGA-exposed kidneys (Table 2). In nor-mal rats, virtually 100% of perfused label appearedin the urine within 20 mm. In contrast, by 40 mm,only some 20% of injected label appeared in urineafter perfusion of nondilated tubules in the NDGA-exposed kidneys. After the microperfusion of cysticnephrons in these kidneys, the urinary excretion ofisotope was even lower; only 4% was recoveredwithin 40 mm. No significant radioactivity appearedin urine from the contralateral (nonperfused) kidneyin any experiment.

The changes that occurred in intranephron pres-sures during microperfusion are shown in Table 3.Pressures increased to a significantly greater degreein cystic nephrons when compared both with non-dilated nephrons in NDGA-exposed kidneys andwith control nephrons in the kidneys of normal rats.The mean pressure changes of —0.5% and +31% innondilated and control nephrons, respectively,were significantly lower than the mean pressurechange recorded in cystic nephrons, and were notsignificantly different either from zero change orfrom each other.

During measurements of preperfusion pressuresin 28 nondilated NDGA-exposed nephrons, urinaryflow rates from the ipsilateral kidney ranged be-

Table 2. Cumulative excretion of 3H-inulin following its microinjection into individual nephronsa

Time after injection

%3H-inulin excretion

total 3H-inulin injected

NDGA-exposed

NormalCystic Nondilated

10mm20mm30 mm

40mm

1.3 l.0(4)3.3 2.l(4)3.8 2.35.c(4)4.0 2.4c (4)

15.5 5.2c(6)21.3 5.5e(6)21.0 59c (6)19.3 6.6e(4)

72.9 15.5(7)96.6 8.5(7)

100.3 9.1 (6)104.9 25.6(2)

a Parentheses contain the numbers of observations.Significantly less than that of nondilated: P < 0.05 or less.Significantly less than that of normal control: P < 0.05 or less.

NDGA-induced renal cystic disease 11

Table 3. Mean (± sEM) pressures and pressure changes in eight NDGA-exposed and eight normal rats during single nephronmicroperfusiona

NDGA-exposed

Normal(N = 18)

Cystic(N = 11)

Nondil(N =

ated8)

Preperfusionpressure,cmH2O 13.6 3.9 19.2 3.1 13.7 1.7Perfusion pressure, cm H20 27.1 6.6 20.1 5.3 21.7 5.6

Percent change5 167.7 6l.1' —0,5 27.2" 31.0 23.0"

a N is the number of nephrons.

perfusion pressure — preperfusion pressureCalculated as follows:

Significantly different from zero percent change; P < 0.02.d Significantly different from cystic nephrons; P < 0.05.

tween 0.2 and 14.3 j.dlmin. There was a low (r =0.51) but significant (P < 0.02) direct correlation be-tween these parameters (Fig. 2). In contrast, duringpreperfusion pressure measurements in 22 cysticNDGA-exposed nephrons, urinary flow rates rangedbetween 0.1 and 12.9 td/min and in 30 normal neph-rons between 1.1 and 17.8 p11mm and did notcorrelate with pressures in either instance (r = 0.12and 0.05, respectively). No correlation was foundbetween preperfusion pressures and SNGFRs inany of the three groups of nephrons.

In three large cysts, it was technically possible tofollow pressures, whereas cyst volume either wasexpanded by the injection of a mineral oil droplet orwas contracted by the withdrawal of fluid. Pres-sures tended to return toward the premanipulationbaseline (Fig. 3) in all three instances.

Structural studies. The labeling index for collect-ing tubular cells increased to 13% of all cells after 2weeks of drug exposure. It remained at that levelfor several weeks before returning to lower levels(Fig. 4). To relate this change to a cellular hyper-plastic response instead of a degeneration/regenera-tion phenomenon, we closely examined kidneystreated with NDGA for evidence of cell death. Cellnecrosis was not seen until the eighth week, wellafter increased thymidine labeling was documented.Hyperplasia was restricted to the collecting tubulatsegments of nephrons; none was documented inother nephron segments. The number of cells com-prising the periphery of collecting tubules increasedwith time in NDGA-exposed kidneys (Table 4).

Morphologic change appeared rapidly and dra-matically in kidneys of the NDGA-exposed animals(Figs. 5-17). The first apparent abnormality oc-curred along collecting tubules and was visible after

1 week of dietary exposure to the toxin. There wasan increase in the number of nuclei around the cir-cumference (Table 4), definite and slight dilation ofcollecting tubular lumens, and the appearance ofpolyp-like structures along tubular walls (Figs. 5,8).

Dilation of collecting tubules occurred in the ab-sence of cast formation in the early stages of theNDGA-induced lesion, and no evidence of inter-stitial change was seen.

Polyp formation likewise was concentrated alongcollecting tubules, particularly along the outer med-ullary segments (Figs. 8, 11—13). Cells forming thepolyps generally were columnar in type and con-tained few cellular organelles (Figs. 11, 12, 14). Ex-cept for the shape of these cells, their ultrastructure

0

E

a,=a,0.C0Ca,E=t0

C

preperfusion Px 100.

Urinary flow, /.L//mrn

Fig. 2. Correlation-regression analysis of urinary flow and intra-lumenal pressures in 28 nondilated NDGA-exposed nephrons(P <0.02).

12 Evan and Gardner

0

E0

00.0Co

0•0>0>0Co

Fig. 3. Effects of intracystic volume changes on inrracystic hydrostatic pressure.

closely resembled that of normal collecting tubulardark cells.

In 13 instances, it was possible to follow dilatedtubules distally to points at which tubular diametersbecame normal. Consistently, poiyps were found atthese junctures (Figs. 8, 11—13). For example, Figs.11 and 12 clearly show a polyp that occludes 65% ofthe visible lumen; this tubule was dilated proximallybut not distally to this point. Small polyps also werefound randomly scattered along the walls of dilatedtubules (Fig. 13).

Microdissection of nephrons was performed onanimals treated with NDGA for 1 week to 6 monthsand on controls of comparable ages (Figs. 8-10). At1 week, the only change noted was dilation of thestraight portion of collecting tubules (Fig. 8). Thechange in lumenal diameter was abrupt. By 1month, many collecting tubules were dilated, andthe first evidence of gross cyst formation was vis-ible (Fig. 9). By 6 months, cysts were foundthroughout the kidney, perhaps with a greater num-ber located at the polar regions (Figs. 7, 10).

Microscopy studies reflected the progressivechanges that occurred with NDGA exposure. Di-lated tubules later came to contain PAS-positivecast material, and cyst formation was notedthroughout cortex and medulla (Fig. 6). Many cystscontained cast material, debris, and macrophages;some tubules appeared occluded by such material inthe 5- and 6-month kidneys. Cells forming the cystwall generally were flattened with enlarged lateralintercellular spaces (Fig. 15). In older kidneys,changes appeared in other segments of the nephron.Proximal tubules showed large aggregations ofendoplasmic reticulum and increased numbers oflysosomes. Proximal tubule cell necrosis did not oc-cur until after 2 months of NDGA-exposure, a timeat which interstitial changes also were noted, con-sisting of infiltrates of small lymphocytes, poly-morphonuclear leukocytes, and numerous macro-phages. In these areas, tubular atrophy, basementmembrane thickening, and fibrosis were evident.

Time, mir,

12 -

I II I

8x0-wCCCC0

-J6

4

2

I I I I II I

00 2 4 6 8 12 16 20 24Time on NOGA, weeks

Fig. 4. Pattern of NDGA-induced collecting tubular cell hyper-plasia plotted against time on 2% NDGA diet.

NDGA -induced renal cystic disease 13

Table 4. Number of nuclei present on cross section of collectingtubules after NDGA exposure

Length of treatmentweeks Number of nucleia

9.9 0.61 15.1 0•7b2 18.2 1.6

25.5 10b8 48.8 51b

24 94•9 8•3b

a Mean SEM of counts on 25 tubulesb Significantly different from preceding values, P < 0.01.

Discussion

Obstruction is the venerated hypothesis to ac-count for cyst formation in the susceptible kidney.In their classic review of the pathogenesis of poly-cystic kidneys, Osathanondh and Potter cite thefact that in the middle 1880's, Virchow blamed cystformation on the obstruction of nephrons by crys-tals or by atresia of intrapapillary ducts secondaryto connective tissue proliferation [51. Nonunion ofprimitive collecting and secretory tubules (obstruc-tion because of discontinuity of the nephron) was apathogenetic mechanism raised later [51. More re-cently, the experimental work of Fetterman,Ravitch, and Sherman demonstrated that ureteralobstruction in fetal rabbits leads to cyst formationin the ipsilateral kidney [6]. In our study of DPA-induced disease, findings of increased intratubularpressures, occlusion of tubules by casts and debris,and local tubular narrowing, led us to conclude thatobstruction to tubular flow probably was present inthat model [2]. In humans, Baert and Steg [7] foundmacroscopic cysts in three and microscopic cysts inall of five kidneys from patients with partial but nottotal ureteral obstruction of 3 to 12 month's dura-tion.

In contrast to these studies, the observations ofLambert [81 and Bricker and Patton [9] are cited inarguments against obstruction as a cause of cystickidney disease. Lambert found no evidence of lu-menal occlusion in careful anatomy studies of ter-minal adult polycystic kidneys, and he demon-strated inulin in cysts after its systemic injection intwo preterminal subjects. Bricker and Patton con-firmed this finding and went on to conclude that cys-tic nephrons participate in the formation of finalurine.

Carone et al found no evidence of obstruction inthe DPT-induced model of cystic disease in rats [1].They recorded no differences from normal values inintralumenal pressures, GFRs, and tubular-fluid-to-

plasma (TFIP) inulin ratios. They concluded that in-creased compliance of tubular basement membranefacilitates cyst formation in DPT-induced diseaseand, by the title of their report, implied that such isthe case in all forms of drug-induced cystic renaldisease.

The major observation of the present study inNDGA-induced cystic renal disease was the findingthat the perfusion of 50 nI of fluid over a 2-mm inter-val caused intralumenal pressures to rise signifi-cantly only in cystic nephrons. To account for thisdifference in response between cystic nephrons andtheir nondilated and normal counterparts, we con-sidered four possibilities: In cystic nephrons, (a)glomerular filtration had increased, (b) water reab-sorption had decreased, (c) mural compliance haddecreased, or (d) outflow resistance had increased.We favor the last of these alternatives for the fol-lowing reasons.

Superfiltration was not present in cystic neph-rons; SNGFRs were similar among the three groupsof tubules. With less assurance, because distancesfrom glomeruli to puncture site were not measured,we also excluded reduced water reabsorption as themajor contributor to relatively increased pressuresduring perfusion. No differences were found in TF/P inulin ratios among the three nephron groups.Conceivably, were cysts to form and therefore bepunctured only at the distal-most end of the proxi-mal tubule whereas other nephrons were puncturedmore proximally, equal TF/P inulin ratios might befound when, in fact, water reabsorption was re-duced in cystic nephrons. No support for this possi-bility could be mustered from our observations. TF/P inulin ratios were, if anything, higher, not lower,in cystic nephrons (Table 1). Microdissection stud-ies showed a random distribution of cysts and di-lated segments along the entire cortical length ofproximal tubules in the NDGA-exposed kidney.The selection of cysts for puncture was based on arandom search of the cortical surface. It would havebeen highly unlikely that only cysts most distantfrom glomeruli could have been selected for exami-nation.

Had distances from glomeruli to puncture sitesbeen measured, their significance would have beendubious. Every cystic nephron has a virtuallyunique configuration. Cysts sometimes hang fromtubules much like grapes from stems; such cystsmay flop randomly around their connections whileundergoing microdissection. Cysts sometimes arepunctured at or near a blind end, one that is notnecessarily in apposition to the juncture of cyst with

9

8

C

NDGA -induced renal cystic disease 15

tubule. Thus, in cystic nephrons, anatomic incon-sistencies and geometric distortion both make it dif-ficult or impossible to measure distance and to chal-lenge any conventional interpretation of TF/P inulinratios relating relative water reabsorption tostraight-line distance down the tubule.

An effectively lowered mural compliance alsocould have caused pressures to increase during per-fusion in cystic nephrons. This possibility is alsounattractive for several reasons. First, other work-ers have argued that increased, not decreased, com-pliance combines with normal pressure and outflowresistance to cause cysts to form in drug-inducedcystic disease [11. Second, cystic nephrons receivedthe lowest relative stimulus to volume expansionamong the three groups of nephrons in our study—50 nI into tubules whose preperfusion volumescould be judged from microscopic appearances tobe greatest. Third, given the outflow resistance andcompliance characteristics of normal and non-dilated nephrons, 50 ni was shown to be an in-sufficient stimulus to increase pressure in normaland nondilated tubules (Table 3). To suggest thatthe cysts studied had reached maximum size andhad stretched their walls to a "limit" ignores obser-vations that NDGA-induced disease is progressiveand that cysts increase in both number and size with

Fig. 5. A low magnification light micrograph of a kidney from ananimal treated with NDGA for I week. Note the appearance ofdilated collecting tubules in the inner cortex and outer medulla(arrows). No other change could be found in the kidney. A: ar-cuate vessels (x4).

Fig. 6. Kidney from an animal treated with NDGA for 1 month.Both dilated (arrow) cystic (C) nephrons are found. The first evi-dence of cast material (CM) is seen in some diseased nephrons(x4).

Fig. 7. Kidney from an animal treated with NDGA for 6 months.Numerous cysts (C) of varying sizes are found through this end-stage kidney. Cysts with diameters as large as 100 were ob-served. A majority of the cysts contain cast material (x4).

Fig. 8. A portion of a dilated collecting tubule microdissectedfrom an animal treated with NDGA for / week. At the pointwhere the dilated portion of the tubule was continuous with thenormal segment, a polyp-like structure (arrow) is noted (X 200).

Fig. 9. A microdissected collecting tubule from an animal treatedwith NDGA for I month. Dilation of the tubule has progressedsuch that a small cyst (C) is observed. The dilated tubule appearsdark due to the presence of cast material in the tubular lumen(x 180).

Fig. 10. A latexed collecting tubule from an animal treated withNDGA for 6 months. This large cyst (C) was juxtapositioned tothe kidney capsule. The exit (arrow) of the cyst is noted. DT:Distal tubule, CT: Collecting tubule (X 150).

duration of NDGA-feeding (Figs. 5-7) [3]. Presum-ably, the cysts we studied would have continued togrow had they been left undisturbed in vivo. Theycould not have both increased (cf., Carone et al [1])and decreased (cf., above) compliance at the sametime. Faced with these considerations, we consid-ered the probability to be poor that low mural com-pliance was the prime defect responsible for in-creased perfusion pressures in cystic nephrons.

Increased resistance to outflow is the most favor-able alternative to explain increased pressures withperfusion. First, micropolyps, evidently occludinglumens to varying degrees, were present along wallsof outer medullary segments of collecting tubules inNDGA-exposed kidneys. Their location at this seg-ment of the nephron is significant. Gottschalk andMylle [101 cite the collecting system as a "limitingfactor in the outflow of urine from the kidney and animportant determinant of the intratubular pressure,particularly at high rates of urine flow." Secondly,micropolyp or papilloma formation occurs in ac-quired human cystic disease [11], in infantile andadult polycystic kidney disease (personal observa-tion), and in not only the NDGA but also the DPA[12] models of cystic renal disease in rats. Thus,these tiny overgrowths of cells represent an ana-tomic abnormality that is more ubiquitous in renalcystic disease than heretofore appreciated. Con-sequently, they are an attractive target on which tofocus pathogenetic speculation.

Obstruction, partial or complete, but not in-creased compliance can explain protracted 3H-in-ulin excretion although if considered out of contextwith the perfusion pressure changes, delayed andreduced 3H-inulin excretion could reflect eddyingand dilution of isotope in "lakes" of cyst fluid or itsescape into the general circulation [2]. The latterpossibility was ruled out here by the absence of sig-nificant radioactivity in contralateral urine.

The ultimate appearance of 3H-inulin in ipsilateralurine indicates that at least some cystic nephronswere not occluded totally.

The fact the 3H-inulin excretion was delayed fromnondilated nephrons but not as drastically as it wasfrom cystic tubules (Table 2) raises the possibilitythat these nephrons were also affected by at leastpartial outflow obstruction, but to a lesser degree.An additional observation supports this possibility(Fig. 2). Urinary flow rates correlated with intra-lumenal pressures in nondilated but not cysticNDGA-exposed nephrons. The correlation, albeit apoor one (r = 0.51; P < 0.02), existed across thespan of recorded urinary flows (2.0 to 14.3 LiJmin).

Fig. 11. A light micrograph showing several cross-sections of adilated collecting tubule. At the exit of the dilated collecting tu-bule, a polyp-like structure (arrow) is noted partially obstructingthe segment of the nephron. The collecting tubule was dilatedproximal to the polyp (x 170).Fig. 12. A higher magnification of the polyp seen in Fig. Ii. Thecells of the polyp do not resemble normal collecting tubular cells(x 1320).Fig. 13. Scanning electron microscopy also showed polyps (ar-row) located at one end of a dilated tubule or cyst (x88).

In normal rats, such correlation between flow andpressure has been reported to exist only above uri-nary flow rates of 10 p.11mm [10]. This difference be-tween normal and NDGA-exposed kidneys couldreflect a more generalized and subtile obstructivephenomenon in the collecting system of the NDGA-exposed kidney, one which, as our anatomy studiesindicate, evolves at different rates and to differentdegrees among nephrons, but one which has pro-gressed to a point in cystic nephrons where outflowresistance is affected significantly by factors (for ex-ample, micropolyps) other than the rate of urinaryflow. Less severe but significant degrees of outflowresistance along nondilated nephrons, by loweringthe threshold at which intralumenal pressures in-crease in relation to urinary flow, could establish abasis for the correlation just cited.

Several findings in our study stand in apparentcontrast with observations of others. Wilson dem-onstrated in chronic (2 to 4 weeks) ureteral obstruc-tion that SNGFRs fall to 76% of normal, from

means (±sEM) of 30.2 1.5 to 23.0 1.2 nllmin1113]. We have argued that cystic nephrons are ob-structed nephrons, yet their SNGFRs were not sig-nificantly lower than normal (Table 1). We record-ed, however, an average decrease to 83% of normalin cystic tubules, and the mean SNGFR of 26.56.1 mllmin among these nephrons does not differsignificantly from Wilson's value obtained duringtotal ureteral occlusion. The decrease to 83% ofnormal also is intermediate to the decreases to 86%of control in plasma-expanded and to 79% in hydro-penic rats observed by Blantz, Konnen, and Tucker[141 during acutely elevated ureteral pressure. In-tranephron versus ureteral, varying degrees of par-tial versus complete, and acute versus chronic ob-struction to urinary flow are differences among thestudies that might explain our less dramatic find-ings.

Gottschalk, Morel, and Mylle [151 recovered vir-tually 100% of tracer inulin within 2 to 4 mm of itsinjection into proximal tubules, whereas we report

16 Evan and Gardner

L

Fig. 14. Transmission electron micrograph of a polyp in an outer medullary collecting tubule. The cells are generally columnar in shapewith widened lateral intracellular spaces (double arrows). Note the numerous cell junctions (single arrows) indicating that the picturedcells are contiguous. BM: basement membrane, L: Lumen (x4700).Fig. 15. Cells of dilated or cystic collecting tubules showed with time an increasing number of lysosomes (Ly) and widening of the lateralintercellular spaces (arrow) (x5300).

1.•.-•"

"° 7

4 .c

L

)• ;'.

a 14

a &

15

18 Evan and Gardner

an apparent delay in the excretion of 3H-inulin fromnormal rats (Table 2). We attribute our apparentlylow recovery during the initial 10-mm period to twodifferences between the Gottschalk study andours—a 2-mm versus a 15-sec perfusion periodand the fact that perfusion and urine collection werenot initiated simultaneously in our study. In ourprotocol, to provide preperfusion baseline data forurinary radioactivity, it was necessary to completeat least one 10-mm urinary collection period beforeinsertion of the perfusion pipette. Actual insertionconsumed a variable amount of time. And it wasdeemed not advisable to delay the start of perfusionuntil the next collection period started in fear that aspontaneous leakage of radioisotope into the neph-ron from the pipette tip might introduce an uncon-trolled variable by randomly lengthening the per-fusion period. Our conclusions are in no way al-tered by this approach, but our data are not directlycomparable to that of other workers because of thisdifference in technique.

The observation that baseline pressures are notincreased in cystic nephrons of NDGA-exposedkidneys differs from our earlier reported finding thatpressures are elevated in dilated nephrons of DPA-induced renal disease [21. This difference is evenmore striking in view of the fact that 5% mannitolwas added to the maintenance Ringers infusion inthe NDGA study to insure adequate urinary vol-umes. No direct explanation for the difference isavailable. Morphologic differences, however, areconsiderable between the two models, especially inthe degree of interstitial change that was present atthe times of study. Whereas NDGA-exposed ratswere studied functionally when cyst formation wassignificant but necrosis and interstitial fibrosis wereminimal (5 to 7 weeks), DPA-exposed rats werestudied much later and at a time when thesechanges were clearly present [12], sometimes to ex-tremes (5 to 20 months). Because of differences induration of drug exposure and in renal morphology(other than cyst formation), we consider it ill-ad-vised to attempt a comparison of functional datafrom the two studies.

Our observations on the nature of NDGA-in-duced damage differ from those of Goodman et al[3J. Those workers cite the proximal, not the col-lecting, tubule as the location of initial damage, anddescribe extensive necrosis of tubular cells and amarked interstitial reaction. Goodman et al in-tended to expedite development of the renal lesionand used, initially at least, 3% dietary NDGA. Weused 2% throughout our study. They recorded

weight differences between control and treated rats;we did not. They began with older and heavier rats(250 g vs. 100-125 g). They describe proximal tubu-lar changes at 6 weeks (their text is not entirelyclear, but it seems to indicate that this was the timethat initial observations were made in treated rats).We did not see such changes until 8 weeks. It maybe that drug-dose and duration of exposure, as theyare in the DPA model [2, 121, and ages of the ani-mals at time of initial drug exposure are importantdeterminants in establishing the location and severi-ty of the NDGA-induced lesion.

We found that whole kidney GFR was reduced toa greater extent than were cortical SNGFRs inNDGA-exposed kidneys: whole kidney GFR wasreduced 46% versus 16% and 11% in cystic and non-dilated surface nephrons, respectively. Qualita-tively, this difference is similar to that observed anddiscussed by Stein et al, in their study of uranyl ni-trite-induced acute tubular necrosis [161. Quan-titatively, it is less. Several ready explanations forthe difference are available. A greater reduction injuxtamedullary nephron GFRs, back diffusion, anda biased selection of nephrons for study are possi-bilities [16]. We saw no evidence of cortical neph-ron collapse in vivo and, as discussed, a significantdegree of back diffusion was not documented. Cystformation, however, tended to be pronounced atcorticomedullary junctions (Figs. 5—7). Conceiv-ably, by their presence in this region, cysts couldmechanically interfere with glomerular hemo-dynamics to a degree sufficient to lower juxtamedul-lary nephron and overall kidney GFRs.

The exact mechanism by which cysts form in theNDGA-exposed kidney, and in any kidney, remainsspeculative. The observed significant increase in in-tralumenal pressures in cystic nephrons indicatesthat they are less tolerant than normal to expansionof their volumes by 50 nl in 2 mm. The progressiveincrease in number of cysts with time suggests thatsome nephrons which are nondilated early growcystic later. The correlation between intralumenalpressure and urinary flow in nondilated tubules sug-gests that their pressures are responsive to changesin the rate of urine formation and excretion. Thedemonstration that this correlation exists over arange of urinary flows that appears to have little orno relationship with intranephron pressures in nor-mal rats [101 suggests that, over time, these tubulesmay be subjected to relatively higher and more fre-quently changing intralumenal pressures in vivo. Itcould be that such rises and fluctuations in pres-sure, occurring at relatively lower rates of urinary

NDGA-induced renal cystic disease 19

flow, ultimately induce cyst formation by causingtubular walls to "blow out."

Reasonably well-sustained filtration into andreabsorption from developing cystic nephrons areprerequisites to this postulate, for either decreasedfiltration or increased water reabsorption couldserve as safety valves and delay or abort the evolu-tion of cysts. Unlike Allison, Wilson, and Gotts-chalk [17] who could show that SNGFR falls as in-tratubular pressure rises in rats with experimentalglomerulonephritis, we found no such correlation inour data from NDGA-exposed kidneys. Thus, a rel-atively unchanged rate of glomerular filtration, noincrease in tubular reabsorption of water, and in-creased resistance to tubular outflow together maydistinguish the potentially cystic kidney from itsmore common nephritic and nephrotic counter-parts.

The extent to which chemically induced modelsmimic human renal cystic disease has not been es-tablished. Several similarities are intriguing. Inulingains entrance to cysts in both model and man [2, 8,91. Pressures in cysts may be elevated [2, 18, 19].And tubular papillomas may occur in both [11, 12].

Drug-induced models of renal cystic disease dis-play differing intratubular conditions. NDGA- andDPA-induced diseases appear to be obstructive inorigin, but DPT-induced disease does not and wellmay result from increased compliance of tubularwalls. Perhaps significant is that in both of theformer conditions the lesion is heterogeneous—on-ly some nephrons are cystic [21—but in the latterthe lesion is homogeneous—virtually all nephronsare dilated [1]. Variations such as these support theconcept that there is more than one cause for renalcystic disease in man.

Acknowledgments

This work was supported by a Basil O'ConnorStarter Research Grant from the National Founda-tion-March of Dimes (Dr. Evan) and U.S. PublicHealth Service Grant AM 17641-01 and -02. PatCooper, Jeff Huser, and Barbara Meyer gave tech-nical assistance.

Reprint requests to Dr. Kenneth D. Gardner, Jr., Departmentof Medicine, Division of Renal Diseases, University of NewMexico School of Medicine, Albuquerque, New Mexico 87131,USA

References

1. CARONE FA, ROWLAND RG, PERLMAN SG, GANOTE CE:The pathogenesis of drug-induced renal cystic disease. Kid-ney mt 5:411—421, 1974

2. GARDNER KD JR. SOLOMON S, FITZGERREL WW, EVAN AP:Function and structure in the diphenylamine-exposed kid-ney. J C/in Invest 57:796-806, 1976

3. GOODMAN T, GRICE HC, BECKING GC, SALEM FA: A cysticnephropathy induced by nordihydroguaiaretic acid in therat: Light and electron microscopic investigations. Lab In-vest 23:93—107, 1970

4. GOLDSTEIN A: Biostatistics: An Introductory Text. NewYork, MacMillan Co., 1964, pp. 51, 56—57

5. OSATHANONDH V, POTTER EL: Pathogenesis of polycystickidneys: Historical survey. Arch Pathol 77:459—465, 1964

6. FETTERMAN GH, RAVITCH MM, SHERMAN FE: Cysticchanges in fetal kidneys following ureteral ligation: Studiesby microdissection. Kidney mt. 5:111—121, 1974

7. BAERTL, STEG A: Cystic changes in adult human kidneysafter ureteral obstruction. Urology 7:526—528, 1976

8. LAMBERT PP: Polycystic disease of the kidney: A review.Arch Pathol 44:34-58, 1947

9. BRICKER NS, PATTON JF: Cystic disease of the kidneys:Study of dynamics and chemical composition of cyst fluid.Am J Med 18:207—219, 1955

10. GOTTSCHALK CW, MYLLE M: Micropuncture study of pres-sures in proximal and distal tubules and peritubular capillar-ies of the rat kidney during osmotic diuresis. Am J Physiol189:323—328, 1957

11. DUNNILL MS, MILLARD PR, OLIVER D: Acquired cystic dis-ease of the kidneys: A hazard of long-term intermittentmaintenance haemodialysis. J Gun Pathol 30:868—877, 1977

12. EVAN AP, HONG SK, GARDNER KD JR, PARK YS, ITAGAKIR: Evolution of the collecting tubular lesion in diphenyl-amine-induced renal disease. Lab Invest 38:242—252, 1978

13. WILsON DR: Micropuncture study of chronic obstructive ne-phropathy before and after release of obstruction. Kidneymt 2:119—130, 1972

14. BLANTZ RC, KONNEN KS, TUCKER BJ: Glomerular filtra-tion response to elevated ureteral pressure in both the hy-dropenic and the plasma-expanded rat. Circ Res 37:819-829,1975

15. GOTTSCHALK CW, MOREL F, MYLLE M: Tracer micro-injection studies of renal tubular permeability. Am J Physiol209: 173—178, 1965

16. STEIN JH, GOTTSCHALL J, OSGOOD RW, FERRIS TF: Patho-physiology of a nephrotoxic model of acute renal failure.Kidney Int 8:27—41, 1975

17. ALLISON MEM, WILSON CB, GOTTSCHALK CW: Patho-physiology of experimental glomerulonephritis in rats. J C/inInvest 53:1402—1423, 1974

18. BJERLE P, LINDQVIST B, MICI-IAELSON G: Pressure mea-surements in renal cysts. Scand J C/in Lab Invest 27:135—138, 1971

19. JACOBSSON L, LINDQVIST B, MICHAELSON G, BJERLE P:Fluid turnover in renal cysts. Acta Med Scand 202:327—329,

1977