Indomethacin secretion in the isolated perfused
proximal straight rabbit tubule. Evidence for
two parallel transport mechanisms.
D de Zeeuw, … , H R Jacobson, D C Brater
J Clin Invest.
1988;
81(5)
:1585-1592.
https://doi.org/10.1172/JCI113492
.
We studied indomethacin as a probe of anion transport across the isolated perfused
proximal straight tubule of the rabbit and discovered that a substantial component of
transport may occur by anion exchange at the basolateral membrane. Various perturbations
involving direct or indirect dissipation of the cellular sodium gradient (ouabain, sodium- or
potassium-free solutions, cooling to 18 degrees C) resulted in only a 50% inhibition of
indomethacin transport, which raised the question of a co-existent alternative pathway for
secretion. Similarly, the anion exchange inhibitor, 4,4'-diisothiocyanostilbene (DIDS),
diminished indomethacin secretion by only 50%. Cooling followed by DIDS or the reverse
sequence resulted in additive inhibition such that the combination abolished active
secretion of indomethacin. We conclude that active secretion of indomethacin by the
proximal straight tubule appears to be in part sodium gradient dependent; the remainder
may be driven by an anion exchanger on the basolateral membrane.
Research Article
Find the latest version:
Indomethacin
Secretion in the Isolated Perfused Proximal Straight Rabbit Tubule
Evidence for Two Parallel Transport Mechanisms
Dickde Zeeuw,* Harry R. Jacobson, and D.CraigBratert
Departments ofInternalMedicine andPharmacology, University of TexasHealthScience Center atDallas,Dallas, Texas75235; *University ofGroningen, Groningen, TheNetherlands;and
tIndiana
UniversitySchool ofMedicine,Indianapolis, Indiana46202Abstract
Westudied indomethacin as a probe of anion transport across theisolated perfusedproximal straight tubule of the rabbit and
discovered that a substantial component of transport may occurby anion exchangeatthe basolateral membrane.Various
perturbations involving director indirect dissipation of the cel-lularsodiumgradient (ouabain, sodium- or potassium-free
so-lutions, cooling to 18°C) resulted in only a 50% inhibition of
indomethacintransport, which raised the question of a co-ex-istent alternative pathway for secretion. Similarly, the anion exchangeinhibitor, 4,4'-diisothiocyanostilbene(DIDS),
dimin-ished indomethacin secretionbyonly50%. Cooling followedby
DIDS or the reverse sequence resulted in additive inhibition such that thecombination abolishedactive secretionof indo-methacin.
We conclude thatactive secretionof indomethacin by the
proximal straighttubule appears to be in partsodiumgradient dependent; the remaindermay bedriven byananion exchanger
on the basolateral membrane.
Introduction
Thesecretion ofawide
variety
of both endogenousand exoge-nousorganic anions is one of theexcretory functions ofthe kidney (1, 2). The latter include numerous drugs. There isgeneralagreement that thesite of action oforganic anion se-cretion ispredominantlytheproximal straightrenaltubule,in
particular the S2segment
(3).
In contrast, the mechanism ofsecretion is much debated. Different carrier-mediated and
passive
transport processeshavebeensuggested
to be present across the basolateral aswell as the luminal epithelial cellmembrane.Anactive
sodium-coupled
aniontransport process appearstobeaprimarycandidate for the transfer of the anionfrom bloodtocell (4,
5),
whereaspassive
diffusionand/or
an anionexchangemechanismhave beensuggestedtoplayarole in transport acrossthe luminal membrane (6). However, re-cent reviewson renal aniontransport byM0llerand Sheikh(1)This workwaspresented in partattheannualmeetingof the American Society ofNephrologyin 1985.
AddresscorrespondencetoDr.Brater,ClinicalPharmacologySec tion, WishardMemorial Hospital,WestOutpatient Bldg,Room316, 1001 W. 10thSt.,Indianapolis,IN46202. Addressreprintrequeststo
Dr. deZeeuw, Dept. ofMedicine, Div. ofNephrology, University
Hospital, 59Oostersingel,TheNetherlands.
Receivedforpublication10July1987 andinrevisedform20
No-vember 1987.
andGrantham(2)illustratethat many other transport systems mayalso be involved in the secretion of organic anions.
Theabovementioned characteristics of the transport sys-tem have derived, for the most part, from studies using
p-aminohippurate
(PAH)'
as a probe. Although the choice of PAHas a prototype seems logical (low protein binding, highsecretionrate, nointracellular metabolism,easy chemical
de-termination), extrapolation ofPAH studies to a general con-ceptof anion secretory mechanisms may be incorrect for sev-eral reasons. First,to allow quantitation, PAH transport mea-surements are oftenperformed at high concentrations of the probe. Thus, the transport mechanism(s) operative at such
concentrationsmay not be applicable to drugs that are present
in blood at much lower
(pharmacological)
levels. Second, thereisevidence that theaniontransport system may bedi-vided intosubsystems, each with specificity for certain anions
(1).
Moreover, the relative contribution of these multiplemechanismsto drugsecretionmay be dose dependent. Hence,
data withPAH may notextrapolatewell toother compounds. For these reasons we chose the isolatedperfusedrabbit proxi-malstraighttubule(PST)to study the processes involved in the renal secretion of another organic anion, the NSAID
indo-methacin, which is highly bound to plasma proteins and is
actively secreted. Moreover, weperformedthesestudies using pharmacological concentrations ofthe drug. In doing so, we
found evidence fortwoparalleltransporters.
Methods
Isolatedperfusedtubule. Thetechnique ofperfusingthe isolated PST was used as previously described by Burg (7, 8). Briefly, slices were cut from kidneys of 1.5-2.0-kgNewZealandWhite rabbits.A PSTwas
dissected free hand in cooled(40C)ultrafiltrate-like solution(in milli-molar: 105, NaCl; 5, KCI;25, NaHCO3;2.3, Na2HPO4; 1, MgSO4;
10,
sodium acetate; 5, alanine; 8.3, glucose; and 1.8,CaCl2;
anditwasbubbled with 95%
02/5%
CO2toreachapH of 7.40). Onlycortical nephron segments of the PSTwereused. Tubuleswerebathed in the ultrafiltrate-like solution with5vol%of filtered fetal calfserum(Gibco, GrandIsland, NY). Thebath wasexchanged continuouslyat aflow rate of - 0.5ml/min. The tubulewassealedintoacollectionpipette using SylgardNo. 184(DowComing Corp., Midland, MI).Inone setofexperimentsasodium-freesolutionwasused. For this purpose the sodium ionwasreplaced byN-methyl-D-glucamine (SigmaChemical Co.,St.Louis, MO)(inmillimolar: 140,N-methyl-D-glucamine Cl; 5,
potassiumacetate;30, N-methyl-D-glucamine HCO3 (made from N-methyl-D-glucamineClbybubblingwith 20%CO2);2.3,KH2P04; 1, magnesium acetate; 8.3, glucose; 5,alanine; 1.8,calciumacetate.Fetal calfserumwasexhaustivelydialyzed againstdouble-distilledwater to
remove thesodium,andwasthen addedtothebathingsolution.Less
1.Abbreviationsused in thispaper:DIDS,4,4'-diisothiocyanostilbene;
Jindo, indomethacin flux; J, volume flux; PAH,p-aminohippurate;
PST,proximal straighttubule.
J.Clin.Invest.
©TheAmerican
Society
for ClinicalInvestigation,
Inc.than 10-7 M of sodium, measured by atomic absorption, could be detected inthis sodium-free bathing solution. The potassium-free
me-dium used in anothersetof experiments was prepared as ultrafiltrate-likesolutionsubstituting5mMKCIby 5 mM NaCl.Inthis protocol, dialyzedfetal calf serum was usedtoprepare thebathing solution. In all experiments,indomethacin(Merck, Sharp & Dohme Research Labo-ratories,WestPoint, PA)aswellasthecompetitive inhibitors sodium PAH (Eastman-Kodak Co., Rochester, NY), probenecid (Sigma Chemical Co.), ouabain(ICN, Irvine, CA), and 4,4'-diisothiocyanos-tilbene(DIDS) (lot04065,Polysciences, Inc., Warrington,PA)were
preparedin the bathingsolution to the desired concentration.All pi-pettesused in theperfusion set-upweresiliconized.
Transepithelial voltage (PD, inmillivolts)wasmeasuredbymeans
of anagarose Ringer's solution-calomel electrode series. Exhaustively dialyzed, tritiated inulinwasaddedtotheperfusateas a marker for volumeflux
(J4).
Indomethacin transport measurement. A 45-minequilibration pe-riodpreceded the actualperfusion experiments. Each indomethacin flux
(Jind.)
measurement(uptothree pertubule)wasbracketed byameasurementofJu,and a 15-minequilibration separated each pertur-bation. Jo, measurements were made with a50-nl constantvolume pipette,and a 110 nl constant volume pipette was used for indometh-acin (seebelow). Only one collection for indomethacinmeasurement
wasmade per perturbation to allow for multiple perturbations within the timespanof one experiment inatubule. The accuracy of sucha
singleJindo measurementwastested infour tubules byperformingthree consecutive periods of indomethacin collections (1 5-min equilibra-tion,J,, Jindo, andJ.).Thus, theinterval between these Jindo
measure-mentswas comparablewith thatusedin theactualexperiments.The
difference among these repetitive sampleswas4.2±3.5% (mean±SD). Allindomethacin sampleswereimmediately transferred to a glass tube (seebelow) and storedat4VC in the dark.
Transfer and measurement ofthe indomethacin sample. The tech-nique used to transfer thecollected indomethacin sample has been described elsewhere(8).Briefly,during collection and transfer, evapo-rationof the sample was prevented byuseof gamma terpinene, an oil thatdidnotinterfere with the indomethacinmeasurement.The 110-nl indomethacin sample was immediately transferred to a small glass tubecontaining 640 nlwater(pH=7.4)containing internal standard (phenylbutazone).
Sampleswereimmediately assayed forindomethacin by HPLC as previously described (9).Asemi-microbore column(10Mm)(Bondpak C18; Waters Assoc., Div. ofMillipore Corp., Milford, MA) allowed small sampleinjections and increased the sensitivity of the measure-mentsby increasing the concentration of drug in the detector. The mobile phaseconsisted of water and methanol (45:55). The flow rate was 130
gl/min,
pumped with a modified Waters solvent delivery system(model 6000 A; Waters Assoc.).Amicroflow cell (1.9Mll)was used in theultraviolet (UV) detector (model 440; Waters Assoc.). Di-rectlyafter the UV detector, 4NNaOH was added to the mobile phase throughaT-connector at aflow rate of 1.33,l/min,
thus maintaining an optimal pH of 12.75 in a reaction loop (130Al,
640C).Thus, indo-methacin was hydrolyzed to a reaction product that could be measuredatlow concentrations by fluorescence detection using an excitation wavelength of 295 nm and emission wavelength of 376 nm (model 650-10 S; Perkin Elmer Corp., Instrument Div., Norwalk, CT). The intraassay variability of 10-7 and10-6Mindomethacin standards was 5 and3%, respectively. Day-to-day coefficients of variation were 6 and 2%, respectively. Recovery ofindomethacin for collected and trans-ferred samples was 99±12%.
Calculations. Jo, was determined by the following formula:
J,
=(Vi - V)/L,whereJ,
is rate ofnet volume flux in nl *mm-'*min-', Viis perfusionratein nanoliters, Vt,is collection rate in nanoliters, and L is tubularlength in millimeters. The twoJ.sperperiod that bracketed theJAndocollection wereaveraged. Jindowasdeterminedbythefollowing
formula:Jindo =[ Vi (perfusate indo)-
Vt,
(collectateindo,)]/L,
where Jindo is rate of indomethacin secretion in fmol *mm- min-', andindo,
is measured concentration ofindomethacin in micromolar.Sinceperfusate
indo,
iszero,thisequation simplifiesto:Jindo = [-Vo (collectateindoj)]/L.
Unless notedotherwisemean±SD are given. Statistical analysis wascarriedoutusinga(paired) ranktest(Wilcoxon).
Results
Indomethacinsecretion rate. The effect of various concentra-tions of indomethacin onJvand PD is shown in Table I. The average tubularlengthwas2.5±0.6 mm. PD and
J,
were mea-suredduring a control period and compared with a period with different bath concentrations of indomethacin ranging from 10-6 to 3 X l0-s M. In 12 of thesetubules, multiple (up to threepertubule) indomethacin concentrations were tested in random order. Table I shows that indomethacin did not signif-icantly affect PD orJo atthe drug concentrationstested, al-though Jvtended to decrease at the highest bath concentra-tions.Jindo rates were measured in 22 PST segments (length,
2.5±0.6 mm) at a bath concentrationof 10-6 M. The
indo-methacin concentrationin thecollectedperfusatewas2.9±1.3 X 10-6 M, indicating bath-to-lumen transport against a chemical gradient; calculated Jindo averaged 16.8±8.8
fmol*mm-'*min-'. Mean PD and Jv were -1.4±0.4mVand
0.39±0.15 nl *mm-'*min-', respectively. Values forJindodid notcorrelatewithPD orJ,
In 9 of these22 tubules and in an additional 8 PST seg-ments, Jindo was measured at higher bath concentrations of indomethacin (5 X 10-6,
10-',
1.6 X10-5,
and 3 X 10-5 M).AJndo
increased withincreasingbath concentrations of thedrug(Fig. 1). If one assumes only one indomethacin transport site atthebasolateral membrane, nointracellular metabolism of indomethacin, andan unrestricted one-waytransfer of
indo-methacinacrosstheluminalmembrane,Michaelis-Menten
ki-neticscan beapplied tothecurvilinear relationship between
Jindo and bathdrug concentration. The inset ofFig. 1 shows such a curve transformation. Although this derivation gives theimpression that indomethacin secretion at leastpartially
follows Michaelis-Menten kinetics, the deviation from a
straight line issuchthatparallelorsequentialtransport
mecha-nismsshould not beexcluded.
Effect
of other organic anions on Jindo. To test whether theobservedindomethacin secretion followedasimilar
transcellu-lar pathwayasdescribed forother organicanions, the effect of twoprototypic substrates fortheorganic acidtransportsystem
(PAH and probenecid) on Jindo was evaluated (Fig. 2).
TableL. Effect of Varying Indomethacin Bath Concentrations on PDandWaterReabsorption(Jo)in PSTSegments
PD J.
[Indomethacin]b,h Control Indo Control Indo mV ni.mm-'.min-'
106 M (n=7) -1.4±0.3 .-1.6±0.2 0.48±0.16 0.44±0.15
5X10-6M (n=7) -1.2±0.4 -1.2±0.3 0.40±0.13 0.35±0.18
10-'M (n= 11) -1.3±0.4 -1.4±0.5 0.39±0.12 0.34±0.15
1.6X 10-5M
(n=6) -1.5±0.4 -1.6±0.3 0.40±0.10 0.33±0.11 3X 10-5M (n=6) -1.5±0.4 -1.6±0.4 0.40±0.10 0.35±0.17
-1ZLI
,-E (7)
E~~
100 ,,'
E
-6 80 C* 2
tCC
c 60 | 15
401 M9)8 X
9-,t
A'~~~~~~~~~~~mQlx
~ ~ ~~~~~~~~r=-0Q921=22S fmol.mm-1. mind1
20 (22)/ 3 Km=18x1046M
;
Ka~~~~1-6
k-5 2;- 15M0
to-6
10-5
2.i0-5
3.;0-5
M[ Indomethacin]
Figure 1. Effect (mean±SEM) of varying indomethacinbath
concen-trationsonbath-to-lumentransport
(J1d.)
inthePST.Theinset showsacurvetransformation (Hanes plot). The values inparenthe-sesrepresentthenumber of tubules studiedat thegiven
indometha-cinconcentration.
First, the effect ofbath concentrations of PAH ranging
from 6 X
1-'
to 4X 10-3 M onindomethacin (10-6M bathconcentration) secretionwasmeasured in sevenPST. Aftera
controlmeasurement,Jindowas measuredafteraddingPAHto
the bathatthedesired
concentration,
starting with either the highestor the lowest PAHconcentration, followedbyanother two PAHconcentrations
in random order, and a recoveryperiod.
AttheseconcentrationsPAHhad nosignificant
effecton PD
(-1.7±0.3
mV atthe lowest to -1.8±0.1 mV atthehighest bath concentrations). In contrast, J, decreased by 33±16% of controlatthehighestPAHconcentrations.Atthe
end ofthe
experiment
Jo recovered to 82±10% of controlvalues.With increasing bath concentrations ofPAH,Jindo de-creased reciprocally from a control value of 19.2±9.4
fmol
-mm-' min'
to < 17% of controlJindo
atthehighest
PAHconcentration used(4X 10-3 M)
(Fig. 2).
Atthe end oftheexperiment,
Jind0
recoveredto 103±30% of control values.n'
a \II
dda*\~~~~\
U'~~~\
I Np
I 0~~~'
I-~~~~~~
10-4
1Inhibitor]
10-3 10-2M
Figure 2. Effect ofvaryingbathconcentrations ofprobenecidand
PAHonindomethacin(bath concentration, 10-6 M)transportrate
(Jind.)in thePST,expressedaspercentchange comparedwith control.
In an additional five experiments a similar dose-dependent
inhibition of indomethacin (10-6 Mbath) secretion was
ob-served by adding varying concentrations ofprobenecid to the bath (10-6to
10-4
M). Thehighestprobenecid bathconcen-trations hadnoeffecton PD (-1.6±0.1 vs. -1.8±0.2mV),nor on
Jh,
(0.46±0.14 vs. 0.47±0.14 nlmm-'
-min-').
Fig. 2shows that
10-6
M probenecid had no significant effect oncontrol
Jind0
(18.4±10.5 fmol-mm-'
*min'),
whereas 10-4Mnearlycompletely inhibited indomethacin secretion.Inall
ex-periments, completerecovery(92±5%)tocontrolJindovalues was observed 30 minafter withdrawing probenecid from the
bathing solution. Apparently, both organic anionscompetefor indomethacin bath-to-lumen transport, and both PAH and
probenecidare able to nearlycompletelyabolish indometha-cin secretion. Although the
inhibition profile
showsasimilarpattern, probenecid is a much more potent inhibitorthan
PAH: -100 times less probenecid than PAH is needed for half-maximal inhibition of
Jindow
Temperature dependency of
Jj,,Mis
The effect of cooling to
180C
on Jindo is depicted in Fig. 3. In thefirst setofexperi-ments (n = 12), bath temperaturewas lowered from
380
to180C.
Asalsoshown byothers(10, 11), coolingto180C
wassufficient to totally inhibit net volume reabsorption in these
PSTsegments,which isreflected by the decrease ofcontrol
J"
from 0.39±0.17 to -0.02±0.04 nl
mm-' min-'.
Similarly, cooling inhibited indomethacin secretion from 24.0±12.0 to10.5±2.9 fmol-
mm-'
*min'.
However, as can be seeninFig.3, in none ofthe tested tubules wasindomethacintransport
completely abolished by
cooling,
the mean inhibitionbeing
48±15%.Toascertain whether the inhibition of
AJnd0
wasstablewith time, the cooling periodwasextendedto 60min infour of
the above
experiments.
No furtherdecrease in Jindo wasob-served
comparing
15 mincooling
(51±6%) with 40-min (51±7%) and with1-h
cooling (54±6%).Infour ofthecoolingexperiments,the
bathing
solutionwasreheatedto370C,
lead-ingto arecoveryof
Jindo
to82±10% of control.Inthe remain-ing four experiments, probenecid(10-4
M)was added to thebath
during
continuedcooling.
Thisresulted in afurther de-crease ofJAndo
to < 15% ofcontrol,
such thatnegligible
amounts of
indomethacin
could bedetectedin the collectedperfusate
of three ofthefour PSTsegments.To further explore the temperaturedependency of indo-methacin transport asrelated to active fluid
reabsorption
inthePST,anadditional setof
perfusion
experimentswasper-formed (n = 5). After control
Jindo
andJo measurements at380C,
the bath temperaturewaslowered in four consecutivesteps
(330, 250, 180,
and120C, respectively).
Ineach of theseC00
so- Figure3.Effect ofcool-100 t t 50 \ ing(180C)andrecovery
80 tocontrol bath
temper-o
E \ ature
(380C)
onbath-ak
60
30to-lumen indomethacin
50 V E
trantransport
(Jindo)
in the*~~~~50
~~~~~PST,
expressedasper-40 centchange compared
with control(left)and
0i .-. . asabsolute fluxrates
38C 1"C 3eC 3 18°C 380C (right). The shaded area (right)indicates the mean±SD ofthecalculated fluxrate,assumingtheindomethacin concentration in the lumenequalsthat in thebathingsolution
(10-6
M).
Indomethacin Secretion
by
ProximalStraight
Tubule 1587
L-
-periods
Jindo
andJv
weremeasured.Jo,decreased from 0.45±15 to 0.30±0.08 (330C),to0.11±0.04(250C), andto0.00±0.03nl mm-'
min'
at 18'Cbath temperature. No further de-crease ofJv
was observed at12'C
(0.00±0.01). Whereas amaximumdecrease of
Jv
wasreachedatthe bath temperatureof 18'C, Jindo inhibition atthis temperature was 53±14% of
control,consistent withprior experiments.Byfurther
decreas-ing bath temperature to
12'C,
Jindo decreased to 31±6% of control values.Fig.4 shows Arrheniusplotsof bothJvand .Jndocalculatedfrom all tested bath temperatures.AssumingthatJv solely reflects active sodium transport across the basolateral
membraneby
Na',K+-ATPase,
one can calculate theactiva-tionenergy of this enzymetobe between 15.5 and 22.6 kcal/ mol(12).Theprofileof the ArrheniusplotforJindois difficult to interpret without
knowledge
of thecarrier(s)
that is (are) involved. However,the dissimilaritybetween theJv
andJindoplots suggeststhatindomethacin secretionisnot
solely
linked to anenergy-dependenttransport process.Sodium
dependency of
Judo
To moreclearly distinguish
between changes in indomethacin transport and changes in
active sodiumtransport in the PST,aseries ofperturbations
were used.First, the effect of adding 10-5 M ouabain tothe
bathingsolutionwastested in five tubules. After 10min
oua-bain had completely inhibitedwaterreabsorptioninthese
tu-bules(control
Ju,
0.40±0.20to-0.04±0.03 nl mm-'*min').
Jindo (control, 36.7±12.4 fmol*
mm'
*min-),
however, wasinhibited by only 47±18%
(Fig.
5,left).
Toinhibit activeso-diumtransportbyadifferent
mechanism,
wealso substitutedN-methyl-D-glucamine for sodiumin thebathandperfusate.
In three experiments with this sodium-free medium, J,, de-creased from 0.33±0.10
(baseline)
to -0.01±0.05 n1 mm-' min-'.Again, Jindo
(control, 47.1±13.4 fmol mm-'*min-')wasinhibited submaximallyby 36±3% (Fig. 5,right).
In the lastseries of experiments (n = 3), bathing andperfusion
solutionwere substituted by potassium-free media.Jo,
wastotally
inhibited(from
0.27±0.03 to 0.02±0.01 nl *mm-'*min-) by thismaneuver,whereasJindoremained at alevel40±4%ofcontrol(34.7±13.1 fmol-ml-'
min-').
Like109I indo 1.4
1.3
1.2
11
1.0'
0.9s
0.8
0.7
380 330 25' 18' 12'C
togJv
-0.3
-0.5
-0.7'
-0.9.
-1.1
-1.3.
-1.5
-1.7.
380 33* 25' 18' 12'C
32 33 34 35 36 32 33
341
35 36 (temperature)-lx104('K1)Figure 4. Arrheniusplotsofbath-to-lumenindomethacin (bath
con-centration, 10-6M)transport(left)andlumen-to-bath net water
reabsorption (Jv,right)inthe PST measured at different bath tem-peratures.
100
80
60
40
20
C
b
Io ouaboi
r-Figure 5. Effect of ouabain (bathconcentration, 10- M) (left)andreplacementof the bathand lumenNa' by N-methyl-D-glucamine(right)on indomethacin (bath
concentra-tion, 10-6 M) transport in the PST.The shadedareas repre-sentthe mean±SD changeof indomethacintransportduring zeroNa' cooling (1 8C).
the resultsduring18'Ccooling, the above perturbations
com-pletely inhibited sodium transport (estimated from
J),
whereas adistinct component of indomethacin transport
re-mained uninhibited. These data raised thequestionof an in-domethacin transport componentindependent of
Na'
trans-port.Effect of
DIDSonJindo.
Since it has been suggested thatanion exchange may play a role in transport of anions in the PST (13-16), weevaluated the effects ofan anion-exchange
inhibitor (DIDS) on
JiAnd.
To find anoptimal dose of DIDS,dose-responseexperimentswerefirst conducted. Fig.6 shows the effect onJindo of adding DIDS to the bathing solution of
fivePST segments.Consecutive increases inthe doseof DIDS
incrementally decreased Jindo to a maximum inhibition of 56±6%of control (23.4±2.5 fmol-
mm-'
*min-').
Theplateau wasreachedwithI0-5
MDIDS. Atlower bath concentrations,DIDS had littleeffectonJ.; onlyat
10-4
MdidDIDS tendtoinhibit!,,,
from 0.37±0.14to0.25±0.13nl-mm-'
*min-'.Infour PSTsegments theeffect of
10-4
MDIDSfollowed bysuperimposedcooling
of the bathing solutionto 18'C wasevaluated. DIDS marginally reduced
Jv
from 0.38±0.05 to0.30±0.07 nl mm-I
min-';
cooling further decreasedJv
to-0.03±0.01
nlo
mm-'* min-'.
DIDS inhibited indomethacintransport to 54±6% of control values (30.4±4.5
fmol-mm-'*
min',
Fig. 7, top). Superimposedcooling fur-therdecreased Jindoto 14±3% of control. The effect ofaddingDIDS to acooled tubulewasevaluated inanother fourPST segments. Cooling decreased
J,
from 0.40±0.20 to-0.03±0.04 nl mm-'1
min-',
whereas superimposed DIDS did not further alter J,, significantly (-0.01±0.05nl *
mm'
*min').
Cooling inhibitedJindoto45±4.5%ofcon-100- 25
.~80- .20
60-
'15
~0 10
20 5
0., 0.
-10-6 10-5 10-4M 10-6 10-5 10-4M
[DIOS]
Figure6.Inhibitory effectof increasingbath concentrations of DIDS
onindomethacin(bath concentration, 10-6M) transport in thePST, expressedasthepercent change compared with control(left)and as
absoluteflux rates (right).
c
20
E 151
E
E
10-R
51
E
0
E
so
c_
_%n
38C 38C 180C
DIDS DIDS
36-
30-24
18-12
6
380C 380C 18C
0ols DIOS
Figure7. Effectof cooling (18C) and superimposedDIDS (bath
concentration, I0- M)onbath-to-lumenindomethacin(bath
con-centration, 10-6M)transportinthePST(top),and the effectof
DIDS andsuperimposed coolingonindomethacintransport (bot-tom).Theplotsonthe leftrepresentthepercentchange compared
with control; the plotsonthe rightrepresentthe absoluteJind.rates.
trol(23.9±3.1 fmol*mm-'
min-'),
whereasadding DIDSto the bath further inhibitedJindoto 14±3% ofcontrol, avaluestatisticallynotdifferentfromzero(Fig. 7, bottom).
Discussion
The present study documents the active secretion of the
or-ganic anion indomethacin across the rabbit renal PST, and addressessomeof theunderlying mechanism(s)forthis
trans-port. Ingeneral, indomethacin behaves like otherorganic
anionstested forsecretionbythePST. At the low (pharmaco-logic) bath concentrations used in thisstudy, indomethacin
wassecreted into the lumen of the PST againstaconcentration
gradient. The process appears tobe saturable at higher bath concentrations ofthe drug; further, secretion is completely inhibited by probenecid and PAH, suggesting atranscellular routefor indomethacintransport.Therelatively high
concen-trationsofcompetitorsneeded for inhibition indicate that in-domethacin has one of the highest affinities for the carrier
quantifiedtodate(2).Variousexperimental approachesinthe
presentstudy revealed dependence ofindomethacinsecretion
on the sodium gradient. Interestingly, however, only about
half of the indomethacin transport appeared susceptible to perturbationsofthisgradient, leavingaconsiderableamount ofdrugsecretion viaindependent mechanism(s). The remain-ing portion of Jindowaseliminatedbytheanion-exchange in-hibitor DIDS, raising the important question of whether a
basolateral anionexchangeprocessaccountsfor theother half of indomethacin secretion.
Beforeinterpretingourobservationsin thecontext of the currentconcepts of anion transport bythe PST, oneshould
address the possible biasintroduced by the study design and
thechoice of indomethacin as thestudied compound. Three
issues may have confounded our experiments: first, protein
binding of indomethacin; second, cellular metabolism of
in-domethacin; and lastly, the absenceofinformation on
indo-methacintransportprocesses across the luminal membrane. As far as protein binding is concerned, indomethacin is
highlybound toalbumin, with various estimates ranging from
90to >99% (17, 18). For reasonsofstandardization and
via-bility ofthestudiedtubule, we added 5%fetalcalfserum to the
bath.Inthis setting, protein binding of indomethacin(1o-6M) was 65±3% (ultrafiltration technique using 14-mm MPS YMT
filters [Amicon Corp., Danvers, MA]). The free amount of drug increased with higher bath concentrations of
indometha-cin (percent binding decreased to 50±4% at 1O-4 M indo-methacin). Ifoneassumes that only the unbound fraction of
the drug in the bathing solution is available for secretion,the
actual
Vm.
and Km arelower than thevalues we calculated.However,correctingthe dose-response curvefor free
concen-trations of indomethacin did not affect theinterpretation of Fig. 1.Theassumption that freeasopposedto totaldrug
con-centration ismost relevantremains
speculative,
sincenocon-clusivedata areavailableas towhether thetransporteris able
to
"strip"
theanion fromalbumin,
aprocesswhich is theoreti-cally dependentupontheaffinity of the anion for thetransport system(19).Futurestudies should address this issue.NeitherPAH, probenecid,cooling (11 C), or ouabain
af-fected the protein binding of indomethacin (61±3, 65±4,
62±3, and 61±5%, respectively). Hence, interpretation of these data wouldnotbeaffected byprotein binding.However,
DIDS
(10-4
M) increased the amount of free indomethacin(48±4% bound), possibly by competition forthealbumin binding site. Not
accounting
for this effectwould have servedonlyto underestimate the
inhibitory
effectsofDIDS on indo-methacin transport, since (if only free drug is secreted) the higher free concentration of indomethacin inthe bathduring DIDScompared with control wouldhave resulted in ahigherJindo.
Interestingly,
review of the literatureconcerning
trans-port ofPAH and other anions by the PST reveals no dataconcerning protein binding
norchangesinbindingcausedby inhibitorsin thebathing solution.Althoughprotein binding ofPAH is claimedtobeverylow(- 25%)(20), largevariations
have been reported (21). As faras the current data are
con-cerned,our
analysis
is in themostconservative fashion. Any unaccounted for effects ofbinding
wouldonlyservetoamplify
our
findings.
Very little information is available concerning renal
me-tabolismofindomethacin.Inhuman
pharmacokinetic
studies, half of the clearance of indomethacin is accounted for by0-demethylation,
and a smallerportion
isN-deacylated
(22).
Mostof theremainder of the active drugaswellastheinactive metabolites are
glucuronidated
and excreted viakidney
andbile. Whether
glucuronidation
occurs in thekidney
is un-known. However, since only very lowconcentrations or noglucuronide conjugates
are measured in serum, andlarge
amounts have beendetectedin the
urine,
glucuronidation
may well occur in the kidney(23). Indeed, we observed an extrapeak
(nexttothe solventfront)
inchromatograms
ofthe col-lectedperfusate
that wasnotpresent in thebathing
solution.Wehave not yet been ableto
conclusively
establishthispeak
to be indomethacinglucuronide (or
anothermetabolite).
Thismay be due to the extreme
instability
oftheglucuronide,
100-- 80
I.-8 60-- 40
20
O- 100-- 80100-- 80-3 60
4-.'.
which can readilydissociate to indomethacin and glucuronic
acid (23). Relative to the indomethacin peak, the extrapeak was small (- 15%), and it did not appear to change signifi-cantly during the various perturbations used in our studies.
Therefore, althoughwe were not able toquantifytubular
me-tabolism(if it exists),we feltit reasonable to assume that
ob-served changesinindomethacin secretion were nottheresult
of changesinPSTintracellular metabolism.
Alloftheconclusions concerningindomethacin secretion
inthe presentstudyare restrictedtothe basolateralmembrane, basedontheassumptionthat indomethacinmovement across
the luminal membrane is driven by passive diffusion (6) or
alternatively by an anion exchange mechanism (16). Subse-quent studies should characterize the transport step at this
barriermorecarefully.
Howdo our data onindomethacintransport contribute to currentconcepts ofanion transport mechanisms in the PST? Todate twomodelshavebeenproposed. The first andoldest theory is thatananion, andinparticularthe prototypePAH,is
secretedacrossthebasolateralmembraneviasecondary active
transport dependent upon
Na'-K'
ATPase; anion thendif-fusesacross the luminal membrane drivenbythehigh
intra-cellular concentration (6). Many studies indifferent species, using different experimental
techniques
anddesigns,havead-dressedthesodium dependency of this active basolateral anion
transport. Somefound PAH transport todecreasealmost
to-tally
using
either ouabain or sodium-freebathing
solutions(3-5, 24-28). In the present study we consistently found a
submaximal inhibition of Jindoby directlyorindirectly
remov-ing the sodium
gradient.
Thereasonforpersistence
ofJindo
inthese
settings
comparedto observationswith otheranionsis unclear.Inaddition, it is uncertainas towhether our data areunique
toindomethacinorrepresentageneralaspectof aniontransport. Because ofoursensitive analytical techniques, we wereableto use a
relatively
low concentration of anion (10-6 M) compared with other studies (210'
M). So doingmay have revealedtheexistence ofa nonsodium gradient-related carrierthat is driven by anion orintracellular counter-anion concentrationsrather thanby theNa'
gradientacross theba-solateral membrane. Because of the higher anion
concentra-tions used by other
investigators,
such pathways may havebeen"swamped" and thereby undetected. In pastreviewson
the anion transport system, authors have suggested that the sodium gradient itselfwas not sufficient to accountfor the
entire PAH
gradient
acrossthe basolateralmembrane (1, 2).This
suggestion
is supportedbythe resultsofthe present study,inthat we were not able to completelyinhibit
JAndo
bydissipa-tion ofthesodiumgradient.
In addition to sodium-dependent anion transport, Dantzler(16) has indirectly offered evidence that a
4-aceta-mido-4'-isothyocyanostilbene-2,2'-disulfonate (SITS)-inhibit-ablecarrier mechanismmay be present on the basolateral (as well as on the luminal) side of the tubule. Although experi-mentshavesuggested,to date none have clearly demonstrated
theexistence ofsuch acarrierindependent from
sodium-de-pendent mechanisms. Thepresent study offersdata that are
consistent
withasecondcarrier(anion exchange?)at thebaso-lateral membrane. The discontinuous shape of the dose-re-sponse curve,and the inabilitytocompletely inhibitJindoby
interference with the sodium gradient (whereas water
reab-sorption
is abolished), together with theclear dissociation ofthe Arrhenius
plots
ofJ,
andJindo, strongly
suggestthepres-enceofasecond carrier(29). Moreover, probenecidand PAH caninhibitthe transport processcompletely,thusmakingthe
possibility of paracellularindomethacinmovement as an
ex-planation for theuninhibitedportionof Jindohighlyunlikely.
The inhibitory capacity ofDIDS itselfonJindo, and the fact thatthesame inhibition can be achieved whenasodium
gra-dient is absent
(cooling),
points againtothe existence oftwo separate transport mechanismsonthe basolateral membrane. In thiscontext itis important to discuss our reasons forusing coolingtoshow sodium dependency during the DIDS
experiments rather than ouabain or
Na'-free
solutions. Though the latter two maneuvers are morecommonlyusedtodissipatethesodium gradient, theyaffected the tubule
prepara-tion inourexperimentsso as topreclude performanceofmore than oneperturbation. Perhapsindomethacin hada noxious
effectsuchthat
superimposed
ouabainor aNa'-free
environ-mentadverselyaffected
viability.
Cooling,
ontheotherhand, didnotaffect thestability
of thepreparation,
as shownby
the similarity of the Jindomeasurementsduring coolingaswellas the recovery dataafterreheating
(Fig. 3).
Thesepractical
justi-fications foruseof
cooling notwithstanding,
thisperturbation
seems a valid probe of energy-dependent sodium transport.
Shaferet al. haveshownthe
similarity
ofcooling
andouabainon thePST (11).Inaddition,our data showequal abolishment of Jv by
cooling
aswithouabain andNa'-free
solutions. More-over,inhibition of Jindowassimilar with all threemaneuvers. Thus, though additivityof
effects of DIDS plus ouabain or DIDSplusNa'-free
solutions wouldrepresentideal confirma-tionof the DIDS plus cooling experiments,wearguethat the interpretation of the additive studies is notobviatedbytheir absence.Intheory, DIDS could simply beacompetitive inhibitor of
theaniontransport system as wasalsosuggested by Koshieret al.(30). However, the fact that DIDS only minimally affected J,and the fact that DIDS could inhibitJindo
by
at most50%,
makesthis unlikely. Itmight be arguedthat albumin in the
bathing solutionmay have bluntedtheeffect ofDIDS(15, 16,
30). Thisseemsunlikely, since the final concentration of
albu-min in the bath (0.1%) was below that which reversed the
inhibition ofPAH transportcaused by DIDS and SITS (15,
30). Furthermore, concentrations of DIDSwere used an order
ofmagnitude higher than those which caused maximal effects
onindomethacintransport(Fig. 6). Thus,evenifsomeof the
DIDS were bound to albumin, it is likely that enough free
concentrationwasavailabletocause amaximal effect. Although the specific effect ofDIDS on the PST is not
known, its documented effectonanionexchange in the eryth-rocyte may in theory be extrapolated to the PST. If so, the second carrier forindomethacin transport on the basolateral membrane could be an anion exchanger. In this respect the second conceptof aniontransport across theproximaltubule proposed byUllrich (31) is of interest.Several groups, using a
varietyof experimentalpreparations(kidney slices, membrane
vesicles,in vivomicroperfusion), foundPAH transport across thebasolateralmembrane to be clearly independent of a
so-dium gradient
(13,
31-34).Therefore,Ullrich et al.(14,31, 36) andalsoDantzler et al.(35) explaintheeffects of perturbationsaimed at dissipatingthe sodium gradient to be the result of
secondary changesin thetubularcellsrather than their
exert-inga specific effect on the PAH carrier. Ullrich argues for a countertransport of anion (PAH) vs. OH- at the basolateral membrane
(14).
Kasheretal. supported the model of an anionexchange
process for PAHtransportacrossbasolateral mem-branevescicles(37). However, they
concluded thistransport wasNa' dependent.
Asfaras theDIDS-inhibitable
compo-nentof Jindointhe presentstudy,
ourdataareconsistent withthe
theory
of basolateral anionexchange.
Thequestion
stillremains, however,
astowhy
we werenotabletocompletelyinhibitindomethacin secretion with either DIDSorby totally
inhibiting
sodium transport(as
measuredby
waterreabsorp-tion).
Assuch,
ourdata remainmostconsistent withtwotrans-port
pathways
forindomethacin. Wecannotexclude thepossi-bility
that thesephenomena
arepeculiar
to use of thetech-nique
of the isolatedperfused
tubule.However,
one canarguesimilarly
that data obtainedfromisolated membranes is lessphysiological. Moreover,
in support of the concept oftwo aniontransportpathways,
Lee etal.(38)
observedin theiso-lated
perfused
ratkidney
thatfurosemide secretionwas best characterizedby
the presence ofbothahigh affinity,
lowca-pacity
system, aswellasalowaffinity, high capacity
system.Further studiesare
clearly
neededtoresolvesuchissues.In
conclusion,
the presentstudy
shows that the NSAID indomethacin istranscellularly
secretedinto the lumen of thePST
by
ananiontransport mechanism likethatdescribed forPAH. The transportis characterized
by
a low transportrate andhigh affinity
for the carrier. Half of the indomethacin secretion is blockedby
maximal inhibition of sodiumtrans-portinthe
PST,
whereas the remaindercanbe abolishedby
the anion
exchange
inhibitor DIDS. The datasuggestthatatpharmacological
concentrations,
indomethacin movementacross the basolateral membrane of the PST is the result of
sodium indomethacin
co-transport
aswell as ofindometha-cin-anion
exchange.
Ratherthanexplaining
themechanism of"active"
basolateral aniontransport
aseithertotally
sodiumgradient
dependent
or astotally
driven via anioncounter-transport,
we feel that ourdata aremost consistent with the simultaneous existence of bothtransport
mechanismson the basolateral membrane of the rabbitPST.Acknowledaments
Theauthors wishtothank Ms. Rebecca Aricheta for her skillful tech-nical assistance.In addition, wegratefully acknowledgethesupport
andadvice ofDr.MichelBaum,Dr. MatthewD.Breyer,andDr.Juha
P.Kokko.
This workwassupported byNationalInstitutes of Healthgrants
AM-27069 (to Dr.Brater)and AM-23091(toDr.Jacobson).Thestudy wascompletedduringasabbaticalyearofDr.deZeeuw,whichwas
supportedinpartbytheUniversityofGroningen.
References
1. Moller, J. V., and M. I. Sheikh. 1982. Renal organicanion
transport system: pharmacological, physiological,and biochemical
aspects.Pharmacol. Rev. 34:315-358.
2.Grantham,J.J.1982. Studies oforganicandcationtransportin isolatedsegmentsofproximaltubules.KidneyInt.22:519-525.
3.Woodhall,P.B., C. C.Tisher, C.A.Simonton,and R. R.
Robin-son. 1978. Relationshipbetweenpara-aminohippuratesecretionand
cellular morphology in rabbit proximal tubules. J. Clin. Invest.
51:1320-1329.
4.Burg, M. B., andJ. Orloff. 1962. Effect ofstrophanthidinon
electrolytecontentandPAHaccumulation ofrabbitkidneyslices.Am. J.Physiol.202:565-571.
5. Dantzler,W. H. 1969. Effects ofK, Na, and ouabain onurate
and PAH uptake by snake and chicken kidney slices. Am. J. Physiol. 217:1510-1519.
6.Tune, B. M., M. B. Burg, and C. S. Patlak. 1969. Characteristics ofp-aminohippuratetransport inproximal renal tubules.Am.J. Phys-iol.217:1057-1063.
7.Burg, M. B.,J. J.Grantham,M.Abramow, and J. Orloff. 1966. Preparationandstudy offragments of single rabbit nephrons. Am. J. Physiol. 210:1293-1298.
8. de Zeeuw, D., D. C. Brater, and H. R. Jacobson. 1987. Micro-HPLC and transfertechniques for directly measuring drug (indometh-acin) transportacross the isolated perfused renal proximal straight tubule. J.Pharmacol. Methods. In press.
9. deZeeuw, D., J. L. Leinfelder, and D. C. Brater. 1986. Highly sensitivemeasurementof indomethacin usinga'micro' HPLC tech-niquecombined with post-column in-line hydrolysis. J. Chromatogr. 380:157-162.
10.Biagi,B. A., andG. Giebisch. 1979. Temperature dependence oftransepithelial potentialin isolated perfused rabbit proximal
tu-bules.Am.J.Physiol.236:F302-F3 10.
11. Schafer, J. A., S. L. Troutman, and T. E. Andreoli. 1974. Volume reabsorption, transepithelial potential differences and ionic permeability propertiesin mammalian superficial proximal straight tubules. J. Gen.Physiol. 64:582-607.
12.Soltoff,S. P., and L. J. Mandel. 1984. Active ion transport in the renalproximal tubule. J. Gen. Physiol. 84:601-622.
13. Berner, W., and R. Kinne. 1976. Transport ofp-aminohippuric acid by plasma membrane vesicles isolated from rat kidney cortex. PfluegersArch.Eur.J.Physiol.361:269-277.
14. Ullrich,K. J. 1976. Renal tubular mechanisms of organic
sol-utetransport.KidneyInt.9:134-148.
15.Hong, S. K., J. M.Goldinger,Y.K.Song, F. J. Koshier, and S. Lee. 1978. Effectof SITS on organicanion transport in the rabbit kidneycorticalslice.Am.J.Physiol.234:F302-F307.
16. Dantzler, W. H., and S. K. Bentley. 1980. Bath and lumen effects of SITSonPAH transportby isolatedperfusedrenaltubules.
Am.J.Physiol.238:F16-F25.
17. Hvidberg, E., H. H. Lausen, and J. A. Jansen. 1972. Indo-methacin: plasma concentrations and protein binding in man. Eur. J. Clin.Pharmacol. 4:119-124.
18.Rane, A., 0. Oelz, J. C. Frolich, H.W.Seyberth, B. J. Sweet-man, J. T.Watson, G. R.Wilkinson, and J.A.Oates.1978.Relation between plasma concentration ofindomethacin and its effectson pros-taglandin synthesisandplateletaggregationinman.Clin. Pharmacol. Ther.23:658-668.
19. Weiner, I. M. 1973. Transport of weak acids and bases. In Handbook ofPhysiology.J.Orloff and R. W.Berliner,editors.
Ameri-canPhysiological Society, Washington,D.C. 521-554.
20. Donoso,V. S., and J. J. Grantham. 1986. Characteristicsof renalp-aminohippurateandurateexcretion in rabbits. J.Lab. Clin. Med. 107:315-321.
21.Richardson,D.R., S.E.Lee, C. W. Bryan,K. D.Nolph, and
M.L. Zatzman.1972.Variable protein bindingof
para-aminohippur-ateandortho-iodohippurate.Clin. Res. 20:606.(Abstr.)
22. Duggan,D.E.,A.F. Hogans, K. C. Kwan, and F. G. McMa-hon. 1972. Themetabolism of indomethacin inman.J. Pharmacol. Exp. Ther. 181:563-575.
23.Faed,E. M.1984.Propertiesofacylglucuronides: implications for studies of thepharmacokineticsandmetabolismof acidicdrugs. Drug Metab. Rev. 15:1213-1249.
24. Chung, S. T., Y.S. Park, andS. K. Hong. 1970. Effects of cationsontransport of weakorganicacids in rabbitkidneyslices.Am.
J.Physiol.219:30-33.
25.Grantham, J. J., P. B.Qualizza,andR. L.Irwin. 1974.Netfluid secretion inproximal straightrenaltubulesin vitro:roleofPAH. Am.
J.Physiol.226:191-197.
26.Gerencser,G. A., and S. K.Hong.1975. Roles of sodiumand potassiumonPAHtransport in rabbitkidneyslices. Biochim.Biophys.
27.Spencer, A.M., J.Sack, and S.K.Hong. 1979. Relationship betweenPAHtransport and Na-K-ATPase activity intherabbit kid-ney. Am. J.Physiol. 236:F126-F130.
28.Maxild, J.,J.M0ller, andM.I.Skeikh. 1981. Involvement of Na-K-ATPaseinPAH transport by rabbit kidney tissue. J. Physio. (Lond.). 315:189-201.
29. Brot-Laroche, E., M.A.Serrano, B. Delhomme, and F. Alvar-ado. 1985. Different temperaturesensitivity and cation specificity of twodistinct d-glucose/Na-cotransport systems in the intestinal brush-border membrane.Ann.NYAcad. Sci. 456:47-50.
30.Koschier, F. J., M. F. Stokols,J.M.Goldinger,M.Acara,and
S.K.Hong. 1980. Effect of DIDSonrenal tubular transport.Am.J. Physiol. 238:F99-F106.
31. Ullrich, K.J., G. Rumrich, G. Fritzsch, and S. Kloss. 1987. Contraluminalpara-aminohippurate (PAH) transport in the proximal tubuleof theratkidney.PfluegersArch.Eur. J.Physic!.409:229-235. 32. Hayashi, H., andT. Hoshi. 1979. Sodium-dependence of p-aminohippurate transport byratkidney cortex slices. Arch. Int. Phar-macodyn. Ther. 240:103-115.
33.Maxild, J., J. V. Moller, and M. I. Sheikh. 1981. An energy-de-pendent,sodium-independent component of activep-aminohippurate transport in rabbit renalcortex.J.Physiol. (Lond.). 310:273-283.
34.Sheikh,M.I., and J.V.Moller. 1983. Nature of
Na'-indepen-dent stimulation of renal transport ofp-aminohippurate by exogenous metabolites. Biochem. Pharmacol. 32:2745-2749.
35. Dantzler,W. H.,and S. K. Bentley. 1976. Low Na effects on PAH transport andpermeabilities in isolated snake tubules. Am. J. Physiol.230:256-262.
36.Ullrich, K. J., G. Capasso, G. Rumrich, F. Papavassiliou, and S. K1oss. 1977. Coupling between proximal tubular transport processes. PfluegersArch. Eur. J.Physiol. 368:245-252.
37.Kasher, Y. S., P. D.Holohan, and C. R. Ross. 1983. Na gra-dient-dependent p-aminohippurate transport inratbasolateral
mem-branevesicles. J.Pharmacol. Exp. Ther. 227:122-129.
38. Lee, L.-J., J. A. Cook, and D. E. Smith. 1986. Renal transport kinetics of furosemide in the isolated perfused rat kidney. J. Pharma-cokinet.Biopharm. 14:157-174.