Bicarbonate secretion in vivo by rat distal tubules during alkalosis induced by dietary chloride restriction and alkali loading

Bicarbonate secretion in vivo by rat distal

tubules during alkalosis induced by dietary

chloride restriction and alkali loading.

D Z Levine, … , M Iacovitti, V Harrison

J Clin Invest.

1991;

87(5)

:1513-1518.

https://doi.org/10.1172/JCI115161

.

To examine in vivo the separate effects on distal tubule JtCO2, of dietary chloride restriction,

bicarbonate loading, and changes in luminal chloride concentration, we microperfused

distal tubules at a physiologic flow rate (8 nl/min) with solutions containing either 45 or 0

mM chloride (after gluconate substitution). Rats were fed a diet containing zero, minimal, or

normal amounts of chloride, while drinking either water or a solution of 0.15 M sodium

bicarbonate. Neither extracellular fluid volume contraction nor negative chloride balance

ensued. Analysis of covariance with repeated measures demonstrated that dietary chloride,

drinking sodium bicarbonate, and perfusion with either 45 mM or zero chloride, each have

separate and significant modulating effects on distal tubule bicarbonate secretion. During

mild alkalemia, there is modest bicarbonate secretion which is significantly different from

zero (-9.9 +/- 3.2 pmol.min-1.mm-1, P less than 0.01), and which is suppressed after

perfusion with zero chloride. In contrast, during more pronounced metabolic alkalosis after

supplemental bicarbonate drinking, the bicarbonate secretory flux is brisk (-26 +/- 3

1.mm-1) and significantly different from zero and persists (-11 +/- 3

pmol.min-1.mm-1) even during perfusion with zero luminal chloride. Accordingly, in this two-day

model of alkalosis induced by dietary chloride restriction, there is regulatory secretion of

bicarbonate by distal tubules in vivo which is modulated by luminal chloride concentration.

Research Article

Bicarbonate Secretion

In

Vivo

by Rat Distal Tubules during Alkalosis

Induced

by Dietary

Chloride Restriction and Alkali Loading

DavidZ.Levine, Michellelacovitti, and Virginia Harrison

With statistical analyses by RamaNairand David Vandorpe

DepartmentsofMedicineandPhysiology, UniversityofOttawa andOttawa GeneralHospital, Ottawa, OntarioKJH 8M5 Canada

Abstract

To examine in vivo the separateeffectsondistaltubuleJtCO2,

of dietary chloride restriction, bicarbonate loading, and changes in luminalchloride concentration, we microperfused distaltubules at aphysiologic flowrate(8 nl/min) with solu-tions containing either 45 or0 mM chloride(aftergluconate

substitution).Rats werefedadiet containing zero,minimal,or normal amountsof chloride, whiledrinkingeither water or a

solutionof 0.15 M sodiumbicarbonate. Neitherextracellular

fluidvolumecontraction nornegative chloridebalanceensued. Analysis of covariance with repeated measures demonstrated thatdietarychloride, drinking sodium bicarbonate,and

perfu-sion witheither 45 mM or zerochloride,each have separate andsignificant modulating effectsondistaltubule bicarbonate secretion.Duringmildalkalemia, thereismodest bicarbonate secretionwhich issignificantlydifferentfrom zero(-9.9±3.2

pmol *

min'

*

mm-',

P <0.01),andwhich issuppressed after

perfusion withzerochloride. In contrast, duringmore pro-nounced metabolic alkalosisafter supplemental bicarbonate drinking, the bicarbonate secretory flux is brisk (-26±3

pmol-min'-

mm-')

and significantly different from zeroand

persists (-11±3 pmol min-

-mm-')

even during perfusion

with zero luminalchloride. Accordingly, in this two-day model

of alkalosisinduced by dietarychloriderestriction,thereis reg-ulatory secretion ofbicarbonatebydistaltubules invivowhich is modulated by luminal chlorideconcentration.(J. Clin. In-vest. 1991. 87:1513-1518.)Key words:acid-base-

kidney.

renalmicropuncture* alkalemia

Introduction

We have recentlyreported that net bicarbonate

secretion'

can occur in ratdistal tubules in vivoandthisresponseis

modu-lated by changes in luminal chloride concentration (1). Be-cause we foundB-type intercalated cells in ratsurface distal

Portionsof this work have appeared in abstract form(1990.Clin. In-vest. Med. 13:470).

Receivedforpublication 24 September 1990 and in revised form 10December 1990.

1.Throughout this paper, unless otherwise stated, the term"secretion

of bicarbonate" referstoabsolute netbicarbonate accumulation in the tubular lumen which is determined by the sum of opposing unidirec-tionalsecretoryandreabsorptive fluxes. The unidirectional secretory

flux may involve apicalCl/HCO3exchange,passive paracellular entry ofbicarbonate,or othermechanisms.

tubules(2) and thesearepresumedtohaveanapical

HCO3/C1

exchanger,ourfindingsarebroadly consistent withnumerous

invitro studies which have defined the bicarbonate secretory characteristics of the exchanger in rabbit cortical collecting

tu-bules (3).However, insofarasthe bicarbonate secretory flux wasobserved in association with the alkalosis after administra-tionofpharmacologic doses of desoxycorticosteroneacetate,

and athigh perfusion rates, it is uncertain whether secretion would occur inmetabolic alkalosis associated with dietary chlo-ride restrictionandunderconditionswhen distal tubular flow rateand fluidcompositionreflect the free-flow state. Indeed, twofree-flow studies which assessed distal tubule bicarbonate flux in chloride depleted alkalotic rats indicate that this seg-ment reabsorbs ratherthansecretesbicarbonate (4, 5).

Itis obvious that these considerations have profound impli-cations for acid-base regulation by distal nephron segments

capableofbothreabsorbingandsecretingbicarbonate. Contin-ued bicarbonate reabsorption inthe face of severe alkalemia

indicatesthatthe distaltubuleplaysa"nonhomeostatic" role

in chronic metabolicalkalosis, contributingtothe mainte-nanceofthedisturbance. Incontrast, ifdistal tubulescanbe

shown to secrete bicarbonate underphysiologic conditions,

this response would beregulatoryinsofar as theseverityof the alkalemia should bemitigated.

Anotherareaof controversy concernsthe functionof the

distaltubuleduring the

repair

ofchloride

depletion

metabolic alkalosis by provision of chlorideand to what extent luminal

chloride concentrations modulatenetbicarbonate flux.

Wes-son(6) has reportedthatduring chloride

correction,

distal tu-bulechloridereabsorption is increased presumably witha

con-comitant reduction in bicarbonate

reabsorption.

In contrast,

Galla et al. (7) could detect no change in the rate of either chlorideorbicarbonatereabsorption inratswith maintainedor

correcting alkalosis.

In the studies referred toabove, bicarbonate fluxes were measured under different experimental conditions which mightaccountfordifferingresults.Variables includethe

dura-tion ofthe

alkalosis,

netbalancesof chlorideandpotassium, supplemental alkali

feeding,

distaltubulebicarbonate load,the

degree of extracellular fluid

(ECF)2

volume

depletion,

and whether or notanimalsarefasted beforetheexperiment.The presentstudies wereundertaken todetermine in fed ratsthe

separateeffectson distal tubulebicarbonateflux of shortterm

dietary

chloride

restriction, drinking

0.15 M

NaHCO3,

and

changesinluminal chloride concentrations.To control the

ef-fects of changes in bicarbonate loadandtubularflow rate,we

microperfused distal tubules atphysiologic flow rates

(8

nl/ min) with solutions

containing

either45 or0mM

Cl,

butwitha

composition

similarto

early

distaltubular fluid in other

re-spects. Rats were fed a diet containingzero, minimal, or

nor-2.Abbreviationusedin thispaper: ECF, extracellularfluid.

J.Clin.Invest.

©TheAmericanSocietyfor ClinicalInvestigation,Inc.

0021-9738/91/05/1513/06 $2.00

mal amounts ofchloride for twodays,with or withoutdrinking

abicarbonate solution. This resulted in neithersignificantECF volume depletion nornegativechloride balances.Inthis way

westudiedthemild alkalemiaassociated with selectivedietary

chloride restriction as well as themoreseverealkalemia asso-ciated with both chloride restriction and bicarbonateloading.

Our results indicate that dietary chloride restriction, alkali drinking, and luminal chloride concentration eachhave sepa-rate andsignificant effects on distal tubulebicarbonate secre-tion in vivo. In mildly alkalotic rats, bicarbonate is secreted into the 45 mM C1 perfusate, but is arrested with a zeroCl perfusate. In the morealkaloticrats,thebicarbonate secretory

flux is striking and persists to asignificant degree even with zeroClperfusion.

Methods

Experimentaldesign and statistical methods (Fig. 1). These studies were undertaken to assess theinfluences on distal tubule JtCO2 ofthree factors:selectivedietary chloridedepletion,alkaliloading,andchanges inluminal chlorideconcentration. Thisassessment wasmade under

microperfusionconditions soastoavoid variations in flow or load of bicarbonate during pairedperfusions. Atthesametime,weensured thatperfusionflowrateandcompositionreflected free-flow conditions. Westudied separately distal tubuleJtCO2inmildlyalkalemicratsafter short-termingestion ofazerochloride diet (8) and in more severely alkaloticratsafterdrinkingbicarbonate. The effects of these maneu-vers could be assessed bycomparison with other animals ingesting a normalchloride diet. Accordingly, as shown in Fig. 1, three groups of

rats werefirst studiedwhile drinking water and eating a diet either free of chloride, containing minimal chloride, or containing normal

amountsof chloride.Ineach case,wedocumentedingestionofthe diet by measuring the amount of food eatenandweight gained, andby demonstrating thaturinechloride excretionreflectedintake. Toassess

the separate effects of alkaliloading,comparisonsweremadewith ani-malsdrinking water,aswellaswith animalseatinga zero ornormal chloride diet. In each of these five groups of animals theinfluenceof

Group A Group B

Drink:

_H20

Drink:

_H20

Drink:

_H20

Diet: ZeroCl Diet:Low Cl Diet:NormalCl

Analysis of

Covariance

With

Repeated

Measures

Group A

Group

C

Drink:

_H20

Drink:

_H20

Diet: Zero Cl Diet:Normal Cl

t

I

Groug

D

GroupE

Drink:HCO3 Drink:HCO3

Diet: Zero Cl Diet:Normal Cl

Two-Way

Analysis of

Covariance

With Repeated Measures

Figure1. Experimentaldesign. See also Methods and Results for de-tails of thestatisticalapproach.

changingthe luminal chloride concentration of theperfusatewasalso studied.

Thestatisticalanalysesemployedmustreflect the intent of the

ex-perimentaldesign,namely,to_identifyfactors that

independently

influ-encedistal tubuleJtCO2whilequantifyingthe effect ofconfounding

variables.One-wayand two-wayanalysisofcovariance withrepeated measures wereusedto assessthe separate effects ofdietarychloride,

alkaliloading,andluminal chlorideondistal tubuleJtCO2

(Fig.

1),

whilecontrollingfor the effects of concomitantchangesinflowrate, bicarbonateload, JvandJ0.Thus,comparisonsbetween groupswere

basedonadjustedmeansofJtCO2.Fstatisticswere_computedtotest

forequalityofadjustedmeansinaccordance with standard statistical

practices,andtheresultingprobabilitiesarestatedthroughoutthetext

and inTableI. Therepeatedmeasuresdesignin theANCOVA

analysis

tookintoaccountthatpairedcollectionsweretakenfrom each tubule

perfused(i.e., perfusionwitha45 mMClsolution and thena0 mM Cl

solution,orviceversa).Differences betweenmeansshown in TableI weretestedfor statisticalsignificanceby ttest,adjustedformultiple

comparisonsusingBonferroniprobabilities.

Ratsand diets(Table II).Thesestudieswereperformedonfed male

Sprague-Dawleyratsweighing 300g,bred and raised inaclimate controlledfacilityattheUniversityof Ottawa. Ratsweredivided into five experimentalgroups.Group Arats ate achloride-free diet and drankdistilledwaterforaperiodof2d. Thediet

contained,

in grams per kilogram diet: Na2CO3 1.15,

K2CO3 1.74,

K2SO4

1.81,

MgCO3*Mg(OH)2-5H2O 1.59,BasalTekladDiet Powder 993.7. This dietprovidedtheNational Research Council-recommendedamounts

ofsodium, potassium, magnesium,andsulphatelevels without

pro-ducingnephrocalcinosis(9,10).GroupBratswerefedaminimal chlo-ridediet and drank distilledwaterfor 2 d. This dietwas

_prepared

by

adding0.82 gNaCltoeachkilogramofchloride-free diet.

Group

Crats ate anormal chloride diet(14.0gNaCl/kg)and drank distilledwater

for 2d,resultinginurinarychlorideexcretionrates

comparable

with thoseof normal chow-fedratsfrom thislaboratory (1,

8).

Group

Drats

atethechloride-free diet and dranka0. 15M

NaHCO3

solution for 2

d,

whereas groupErats atethe normal chloride diet and also drank the

NaHCO3solution for 2 d. Animalswerehoused in individual

meta-bolic cages thatpermittedurine

collections,

under oil with

thymol

as a

preservative,for 16hbefore the

microperfusion

experiment. They

had freeaccess tofood and drink until the time of the

experiments.

Microperfusionexperiments.Ratswereanesthetized with 100

_mg/

kg Inactin(BYKGulden, Konstanz,FRG)andwerepreparedfor mi-cropunctureaspreviouslydescribed(1).Allanimalswereinfused with

1.0%bodyweightof donorplasmafollowedby1%b.w./hofasolution

containing92mMCl,30mM

HCO3,

120mMNa,and 2mM Kfor groupsA andD,with few

exceptions,

orwith 1%

b.w./h

of isotonic saline for groups B,C,and E.The data derived from

paired

microperfu-sioncollections,from six additional

unpaired

collections

(group

A)

and fromfree-flow collections. 43pairsof distal tubular collectionswere

obtained from 37rats.The

perfusion

solutionscontained

_(in

millimo-lar):55Na,2K,12HCO3,26 urea, and 45Clor_gluconatefor chloride substitution. The sequence of

perfusion

with

_{chloride-containing}

or

chloride-freeperfusateswasvaried.In

addition,

24free-flow

early

distal

sampleswereobtained

often,

butnot

always,

from therats

_providing

thepaireddata. Thesesamplesconfirmed that total

CO2

and chloride concentrations of 12 and45 mM,

respectively,

andaflowrateof 8

nl/min reasonably reflectedfree-flow conditions. All

_perfusion

solu-tions contained 0.05% FD&C greendyeandwere

_gassed

with humidi-fied 9%C0J91%

02

toattaina

_pCO2

of 65mm_Hg.All in vivo collec-tionswerequantitativeand

timed,

and the concentration of

[3H]inulin

wasdeterminedin theperfusateand in thesamplestodetect

significant

departuresfrom the nominal

_perfusion

rates. Inthe_present

experi-mentstheinvivoperfusion rate was7.8±0.1

nl/min

forthe86

paired

collections.

Analyticalmethods andcalculations.Total carbon dioxide

concen-trationwasmeasuredby

microcalorimetry

as

_previously

described

(1).

TableI.Blood and BalanceData

GroupA Group B GroupC Group D GroupE

Diet ZeroCi LowCl Normal Cl ZeroCI NormalCI

Drinking Solution H20 H20 H20 NaHCO3 NaHCO3

No.of rats 9 9 6 11 9

Bodyweight(g) (overnight) 9±1 11±2 11±1 8±1* 15±1

Food consumed (g) 22±1 23±1 24±1 22±1 25±1

Blood data

pH 7.42±0.01 7.39+0.01 7.41±0.01 7.52±0.03* 7.44±0.01

PC02(mmHg) 51.3±1.7 53.4±1.3 45.7±1.7 47.6±2.7 48.4±2.1 HCO3(mEq/liter) 32.1±0.5* 31.3+0.5$ 27.8±0.2* 37.3±1.2*1 31.6±1.0

Na(mEq/liter) 145±1 142±1 144±2 143±1 145±1

K(mEq/liter) 4.5±0.1 4.6±0.1 4.1±0.2 3.3±0.1$ 3.8±0.1

Cl(mEqiliter) 98±0 100±1 100±1

89±2*f

98±1

Hct(%) 47.1±0.7* 46.9±0.8 44.1±1.2 44.6±0.7 43.4±0.8

Prot

(mg/db

6.4±0.1* 6.5±0.2 5.8±0.2 5.7±0.1 5.8±0.2

Urinedata

Volume(ml/16h) 22±2 15±2* 18±2

53±8"1

37±5

pH

5.99±0.051

5.93±0.07§ 6.09±0.04 7.71±0.11* 7.87±0.5*

HCO3(uEq/16 h) 23±6' 13±3w 30±6' 4350±795* 3241±602*

Na(,uEq/16 h) 331±42'1 649±113iI 5375±597§ 8862±1342 9935±739

K(Eq/16h) 396±42 580±121 591±93 1236±161I 234±42

Cl

(,Eq/J6

h) 30±11* 259±52*1 3615±257

275±641

3654±396

* P<0.05 compared with group

E.

*P<

0.01

compared with group C.I_{P <}

_0.01

_{compared with groupE. P <0.05compared with group C.}

Micropuncture pipettes and sample handling pipettes contained Results

Hepes-bufferedmineral oil and samples were placed on a quartz dish

flooded with Hepes-buffered mineral oil. A separate tray contained

Whole animal

data (Table

I)

fourNaHCO3standards(5, 10, 20, 30 mM NaHCO3) in water-equili- Ratsdrinkingwater(groupsA-C).These rats ate a diet contain-brated mineral oil. Astandard curve wasrun beforesample analysis ing nochloride, minimal chloride, or normal amounts of chlo-andstandardsbracketed the determination ofthe sample and perfusate

ride

whiledrinking water. All animals gained - 6-8 g body tCO2concentrations. weight during the night before the experiment and consumed

Tubularfluid chloride concentrationsweredeterminedbyconstant _{20gof food. Plasma bicarbonate concentration in group C} currentelectrotitration withpotentiometric end-point sensing

accord-ingtoamodifiedmethodofRamseyetal.(I Aseriesof stan- was not

different

thanthat

previously

reported byusin

nor-ingto amodifiedmethod of Ramsey et al. (11). AseriesofNaCIstan- ma.fe ras whl ru asetn azr hoiede a

dards was runbefore andduring sample andperfusateanalysis togener-

mally

fed rats,

while

group Arats

_eahng

azero

chloride diet

had ate aregression line and assess instrument stability. For low chloride a modest elevation of plasmabicarbonate. Plasmapotassium

concentrations, a midcurve chloride standard was added to the sample concentration was similar in all three groups. Neither

signifi-orperfusate before titration. cantECF volume

depletion

or

negative

chloride balanceswere

Other methodsand calculations havebeenrecently detailed(1). present.Therewere nosignificantchangesin body weight nor

Table I. ExperimentalProtocols

Diet Infusion

GroupA 2d,zeroCldiet, drinking water 1%b.w. donorplasmafollowed by1%b.w./hinfusion composed of 92 mM Cl,

(8 rats) 30 mM HCO3, 120 mM Na, and 2 mM K

GroupB 2d, lowCldiet, drinking water 1%b.w. donor plasmafollowed by1%b.w./h of isotonic saline

(9rats)

Group C 2 d, normalCldiet,drinkingwater 1% b.w. donor plasmafollowed by 1% b.w./h of isotonic saline (6 rats)

GroupD 2d,zeroCldiet, drinking 0.15M 1%b.w. donor plasmafollowed by1%b.w./hinfusioncomposed of 92 mMCl,

(7 rats) NaHCO3 30mMHCO3, 120 mM Na, and 2 mM K

GroupE 2d, normal Cldiet, drinking 0.15M 1%b.w. donor plasmafollowed by 1% b.w./h of isotonic saline

15 E E

E

00

cm

0

-15

Zero Chloride LowChloride Normal Chloride

Diet Diet Diet

S

X~~~~~~*

45 0 45 0 45

PERFUSATE [CII (mM)

* p< 0.05 vs. zero

Figure 2. Distal tubule JtCO2 in rats drinking water on different chloride-containing diets. See also Results and TableIII.

urine chloride excretion reflected dietary intake being

ex-tremely lowin group A (some urines contained only zero or 1

mEqCl concentration),and in the normalrange (3.5 mEq/16

hours) in Group C.

Ratsdrinking NaHCO3 (GroupsD, E).Group D rats, which

drank NaHCO3 and ate the zero chloride diet, were frankly

alkalotic and excreted an alkaline urine without significant ECF volumedepletionor negative chloridebalances. Although

urine chloride excretionrates wereminimal,there was a differ-enceof - 7 gin overnight weightgainwhen compared with

group E rats.Becausehematocritandplasma protein

concen-trationswerenotdifferentfromgroup E rats, webelieve signifi-cantECF volume contractioncould nothavebeen present

dur-ing this 2-d period. Plasma potassium concentration was 3.3

mM,a valuesignificantlyless than ingroup A (P < 0.05). In contrast, group Erats,whichatethe normalchloride diet,did notbecome alkalotic.

Microperfusion data

Rats

drinking

H20(groups

A-C).

Fig. 2 and Table III

summa-rize microperfusiondataforratsperfused withzero or 45 mM

Cl. The rateof total

CO2

secretionwith45 mM

Cl

perfusate

was modest and becamelessnegative orsignificantly positive

with zero Cl perfusate. These data were analyzed by the method ofrepeated measures analysis ofcovariance and re-vealed a significant effect of both diet (P < 0.05) andperfusate [Cl], (P<0.05), on modulating JtCO2. As shown in Fig. 2,

group A ratseatingthe zerochloride diet hadsignificant bicar-bonate secretion(P<0.05) usingaone-tailed t test when

com-paredwith zero. In view of the interest of this measurement

and the borderline significance value,weadded six data points whereonly the 45 mMperfusate collectionwassuccessful.For

these 14 samples (-9.9±3.2 pmol

-min-'

-mm-'),

the differ-encefromzeroissignificantatthe P<0.01 level. IngroupC rats, thepositive reabsorptive fluxseenwith zerochloride per-fusatewassignificantly different fromzero(P<0.01).

Table III alsogivesdata on water movementsandchloride

fluxes, which showed marked and consistent changes when per-fusate [Cl]waschanged. Inalltubules, perfusion withzeroCl

wasassociated withalower Jv andbrisk chloride influx when comparedto thehigherrateofwaterreabsorption andnet chlo-ride reabsorption during45 mMClperfusion. Thesemeasures,

Jc0

and

_Jv,

were

_significant

_{covariates among the three groups}

(P < 0.01) butneitherload(JVxperfusate

[tCO2J)

nor

flow

(V), potential determinants of JtCO2,weresignificant covariates.

RatsdrinkingNaHCO3(Table

III,

Fig. 3). Groups D and E

providedatafromratseatinga zeroCldietorthe chloride-con-taining diet and drinking

NaHCO3.

These groups should be compared withgroups AandC,asoutlined in Methods and in Fig. 1. Despite the 8

nl/min perfusion

rate, group D rats showedstrikingtotalCO2 secretion with45 mM Cl. The perfus-ate[total

C02]

rosealmostthreefold, despitethelower plasma

potassium concentration. Surprisingly,evenwithzeroCl

per-fusion, JtCO2 remained negative, albeitnot to thesame degree aswith the45 mMClperfusion. Thesedata areincontrast to thosefromratseating the chloride-containing dietwhere JtCO2 was notdifferent fromzero with both perfusates.

Atwo-wayanalysis of covariance with repeatedmeasures wasperformedongroups A,C,D, and E asdepicted in

Fig.

1.

Theanalysisprovides conclusions identicaltothose described forgroups

A-C, namely

that both

perfusate

Cl and

dietary

Cl

TableIII. SummaryofMicroperfusionData

Mean

No.tubules/ Tubular Perfusion Perfusate luminal Perfusate Collected

No. rats length rate [Cli ICJl [tCO2J [tCO2J Jy ja JtCO2

mm nilmin mMf mML mm mm nl-min-l * mml nmn-min-I.*mm-1 nmnl-min-I.*mm-1

GroupA

Drink:H20

Diet: Zero Cl

GroupB Drink:H20

Diet:LowCl GroupC

Drink:H20

Diet: Normal Cl

GroupD

Drink: 0.1 5MNaHCO3 Diet: Zero Cl

GroupE

Drink: 0.15MNaHCO3

Diet:NormalCl

8/8 1.5±0.0 8.0±0.2 0±0 15.8±2.6 12.4±0.4 19.0±1.2 7.5±0.2 44.3±0.2 52.5±2.0 12.1±0.5 25.6±1.7

9/9 1.5±0.1 7.6±0.2 0±0 20.8±4.1 11.6±0.3 15.0±0.5 7.9±0.2 44.9±0.2 55.2±1.9 12.1±0.4 22.6±1.2

9/6 1.4±0.1 7.7±0.2 0.4±0.3 15.1±2.3 12.9±0.5 16.3±1.0 7.6±0.2 46.2±0.7 57.5±2.1 12.6±0.4 24.8±1.4

9/7 1.4±0.0 8.1±0.3 0.6±0.2 11.8±0.8 11.7±0.2 21.9±1.2 7.9±0.4 45.9±1.0 48.5±2.5 11.9±0.4 34.5±3.9

8/7 1.6±0.1 8.0±0.2 0±0 12.3±1.0 12.4±0.3 18.4±1.2 7.8±0.3 44.8±0.2 48.2±1.5 12.8±0.2 27.1±3.1

1.9±0.2

2.3±0.1

1.7±0.3 2.2±0.3

2.0±0.3 2.6±0.4

2.0±0.2 2.7±0.2

1.9±0.3 2.4±0.4

-111±20 61±6

-132±20 35±9

-107±16 56±26

-82±9 125±13

-84±12 85±22

1+6* -8±4*

8±5* -7±7*

13±4*

-1±7*

-12±3*

-26±3*

4±6* 0±7*

*_{Statistical analyses, describedin Methods andResults,revealsignificanteffectsondistal tubuleJtCO2ofdietarychloriderestriction,bicarbonateloading,and}

per-fusate chloride concentration.

independently influenceJtCO2. Inaddition, there isa

signifi-cant effect of the bicarbonate-drinking solution as well (P

<0.01). Again, neither loadnor

flow

rate

influenced

JtCO2,

whereas both

Jv

andJa weresignificant covariates.In group E

animals,as shown inFig. 3, theJtCO2at45 mM and0mMC1 perfusionwas notsignificantly different fromzero(P<0.01).

Itisnoteworthy, that thereabsorptivefluxin groupCanimals withzeroClperfusatewasalsosignificantly different fromzero

(P

<

0.01),

suggesting

a

suppression

ofthe

reabsorptive

flux

_by

bicarbonate ingestion (see Discussion).

Discussion

The presentstudiesweredesignedtoexamine theseparate ef-fectsinvivoof dietary chloride restriction, ingestion of

supple-mental bicarbonate, and luminal chloride concentration on

distal tubule bicarbonatetransport

(JtCO2)

inthe rat. Weused

amodel ofmetabolic alkalosis induced by

dietary

chloride

re-striction without ECF volume contractionornegative chloride

balance.Our resultsdemonstrate

for

thefirst time that (a)

di-etarychloriderestriction and

drinking

sodiumbicarbonate

in-dependently modulate

JtCO2, (b)

intratubular

perfusion

with either 45mM or zerochloride

significantly

altersJtCO2 inboth control andalkalotic rats, and (c) modest butsignificantnet

bicarbonate secretionoccurs atnormalflow

during

mild

alka-losis, whereaswhenthealkalosis ismoremarkeddistal tubule bicarbonate secretion is

brisk,

thereby

tending

to

mitigate

rather thanmaintainthealkaloticstate.

Comparison

with

free-flow

studies. Wesson (4) has recently

notedthat

conflicting

dataonsegmental bicarbonatehandling

in metabolic alkalosis mayderive from differences in the meansof inducingthedisturbance,the presenceof potassium

depletion, changes

in tubular load of

bicarbonate,

andthe dura-tion of thedisturbance beforethemeasurementof bicarbonate

flux. The studies ofWesson (4) andofDe Mello Aires and

Malnic(5) provide distaltubule dataonbicarbonatetransport

during

chloridedepletion metabolic alkalosis of1 or 4 wk

dura-tion. Theseratsreceivedfurosemideand supplemental bicar-bonateand ate achloride-free diet.At thetime ofstudy, they were potassium depleted and had been fasted overnight. In Wesson's study (4), punctureoflateand early distal segments

ofthe sametubulerevealedincreasedratesofbicarbonate

reab-10

I

E

E

I

E_"I 0

0 0

-10

F

-20

-30

Normal Cl

DietI

.

I

w.

Zero ci

45 0

PERFUSATE [CII (mM)

* p< 0.01 vs. zero

Figure3.Distal tubuleJtCO2inratsdrinking0.15MNaHCO3on

differentchloride-containingdiets. See also Results and TableIII.

sorption when comparedwith control animals. The load of bicarbonate delivered to the punctured tubules was increased

incomparisontocontrols andcontributed,atleast in part,to

the enhanced reabsorptive rate, with potassium depletion

thought to also play a role (4). These valuable and difficult free-flow measurements follow the studies undertaken almost

20years agoby De Mello Aires and Malnic(5).Themeansof

inducing chloridedepletionmetabolic alkalosiswassimilar but atthetime ofthe acuteexperimentamannitol and bicarbonate

diuresiswasinducedtofacilitatesample collection,and

signifi-canthypokalemia(plasma [K] = 2.55 mM) was also present (12). These authors also concluded that the distal tubule

reab-sorbedgreater amounts of bicarbonate than under control

con-ditions, althoughabsolutereabsorptiverates were notdirectly

measured for each tubule punctured.

Factors

influencing distal tubule JtCO2

1.

Feeding

andfasting.

Thebicarbonate

secretory

fluxesin our

studieswere measured inratsfed before the

experiments.

We

havealready demonstrated thatwhetherrats arefastedor not

before theexperiment can determine the directionof JtCO2

(13).Itispossible,therefore,that thereabsorptiveflux in fasted ratsreportedbyWesson(4)may be

suppressed by feeding.

2.

Tubular bicarbonate load.

In our

studies,

bicarbonate loadwasrigidlycontrolledinourpaired

perfusions

and docu-mented not tobea

significant

covariate,

whereasin the free-flow studies loadwasincreased.Itis well establishedthat load

canincreasethebicarbonate

reabsorptive

flux in free-flow dis-taltubule studiesasreported byCapassoetal.(14).

3. Potassium

depletion.

Potassium

depletion

wasabsent or

minimal in our rats

secreting

bicarbonate butin therats

stud-ied byDeMelloAires andMalnicand Wesson(4,

5),

potas-sium

depletion

was present and couldhave stimulated distal

tubulebicarbonate

reabsorption

as

suggested by

Capasso

etal.

(14). Presumably, ifgroup D

potassium

concentrationswere

normal, the

striking secretory

flux weobserved would have been even morepronounced.

4. ECF volume contraction and

negative

chloride balance. Excretionratesfor chlorideandhematocritandplasma

protein

concentrations indicatetherats westudiedwere notECF vol-umecontracted norin

negative

chloridebalance. Inchloride

depletion

associated withdiuretic use or

gastric

alkalosis,

it is possible that the attendant chloride losses and ECF volume

depletion

wouldstimulate bicarbonate

reabsorption.

5. Duration

of

the metabolic alkalosis. Our

experiments

alsodiffered

markedly

fromthesefree-flow studiesin that the

alkalosisweinducedwasonly of2-d duration.Itis

possible

that cell

changes

leading

toenhancedproton

secreting

ATPase activ-ity, or suppression of unidirectional bicarbonate

secretory

mechanisms (see below),

might

evolveoversubsequent

days

or

weeks. The persistence of bicarbonate reabsorption in distal

tubules

during

the phase ofrebound alkalosis that follows NH4Clloading isanexample ofaresponsewhich isfirst

regula-tory

during

the acidotic phase,and then becomes

nonregula-tory in thatit contributestothemaintenance ofthemetabolic alkalosis

(15).

In chloride

depletion

metabolic

alkalosis,

it is

possible

thatdistal tubules

respond

ina

regulatory

mannerin thefirst 48hand then

change

their

JtCO2

fromnetsecretionto

reabsorption.

6.

Histological

variability.

It is

likely

that our

two-looped

microperfusion

technique requires

alength ofdistal tubule

col-lecting tubule where B-typeintercalated cells are found (2). On the other hand, free-flow collectionscan bemadeproximalto theinitialcollectingtubule.Accordingly,thebicarbonate

reab-sorptive fluxes reported byWesson(4) and De MelloAiresand

Malnic(5) mayreflecttransport byportionsofthedistal tubule

devoid of bicarbonate secretingcells.

7.

Supplemental bicarbonate

ingestion. As already noted,

in our laboratory, overnight fasting before the acute micro-punctureexperiment determines whether thenetbicarbonate flux isreabsorptiveorsecretory, but alkaliloading hasan addi-tional effect (13).Inbicarbonate-loadedratsstudiedby Wesson

(16), JtCO2 fellfrom 25to 11 pmol.

min-'

*mm-' despitethe

fastedstateofthe animals.Inourprevious microperfusion

stud-ieswe obtainedsimilarresultsafter bicarbonateloading in

fast-ingrats. Indeed, even in the presentinvestigationsonfed

ani-mals, the reabsorptive flux in group E rats is lower than in group C ratsafter drinking bicarbonate. Similarly, the micro-perfusiondataofChan et al.(17)show asuppressionof JtCO2 asaresultof alkali

loading

infedrats. Just asnormal

feeding

mayallowregulatorybicarbonate secretiontooccur,

bicarbon-ateloadingmay bethoughttoresultinaregulatorysuppression of JtCO2.

8.

Jv

and

Jc1.

Jv

andJ0arealsopotential modulatorsof JtCO2. Inall cases,perfusates containingzerochloride showed highly significant C1 secretion and lower Jv when compared

with chloride-containing perfusates. As already noted in

Re-sults,these weresignificant covariatesand were accountedfor

in thedesignation of significance ofthe three parameters of interest, i.e., dietary chloride, alkali ingestion,andluminal

chlo-ride concentration.

9. Luminal chloride concentration and transtubular bicar-bonateconcentration

gradients.

Because wemeasuredonlynet

bicarbonate fluxes in these studies, changes in net secretion mightbe the resultofalterationsin theunidirectionalsecretory flux (whether duetoapical anion exchange orparacellular

entry,etc.) orchangesin theunidirectional reabsorptive flux. Accordingly,theinterpretation ofourresultsofperfusion with

zerochloride should alsoinclude the

possibility

that the

pre-dominant effect isoneofanincreased

reabsorptive

flux. Whatmightaccountfor the lowernetbicarbonate

secretory

flux inthe

mildly

alkaloticrats(group A),versusthethreefold

greaterflux in the moreseverely alkalotic

animals

(groupD)?

Ineachcase,chloride-free

perfusion

reducedthe netsecretory

flux,

but onlyin groupArats wasit completely suppressed.The meanluminal chloride concentrations werecomparable after

zerochloride

perfusion,

andabove thesuggested Km of4-1 1

mMforthe

HCO3/C1

exchangerintherabbit (3).Inthisinvivo

setting,

it is

possible

that thecharacteristics ofthe exchangerare alteredbythealkalosis per se,

dietary

chloride

restriction,

or

alkaliloading. Moreover,our

experimental

variablesmay have also modulated otherdeterminants ofnetbicarbonate trans-port. Thetransepithelial potential

difference,

the

transepithe-lialdiffusional permeabilitytobicarbonate, and theproton se-cretory rate mayhavevariedin control andalkalotic animalsin response tochanges in lumen chloride concentration,

trans-tubularbicarbonate concentration

gradients, dietary

chloride restriction,andalkali loading.

Implications for

maintenance

of

chloride

depletion

metabolic alkalosis

Ourshort-termstudies separatedtheeffects of dietary chloride restriction fromthoseof alkaliloadingin the absenceof ECF

volumecontraction and

negative

chloride balance.

Neverthe-less,

ourdata may be relevanttoatleastonefacet ofthe clinical

presentation

of metabolic alkalosisassociated with

vomiting.

Inthis

situation,

threeobservationshave been

consistently

made. Withinthefirst 36h there is

(a)

asustained

alkalemia,

(b)

excretioninthe urine of

negligible

amountsof

chloride,

and

(c)

excretioninthe urine of

large

amountsof bicarbonate

(I18).

The

provision

ofalkali inourgroup Dratswould

correspond

to

the

generation

ofnewbicarbonate

by

loss of

gastric juice,

and

the bicarbonate

secretory

fluxweobserved could lead tothe appearance of bicarbonate in theurine within the first 36 h.

Thisbicarbonaturiahas been

variably

describedas a

disequilib-riumstate

(18)

or a stateinwhich the

"reabsorptive

capacity

for bicarbonate" is exceeded

by

distal

nephron

bicarbonate

load.Our

findings

suggestthis could representadistal tubule

andcortical

collecting

ductbicarbonate

secretory

flux.

Acknowledgments

WearegratefultoBeth Caskie and DianeGagnonfor_expert

_help

in

preparingthemanuscript.

This workwas_supportedbyagrant from the Medical Research Council of Canada.

References

1.Levine, D.Z.,D. Vandorpe,andM. lacovitti. 1990.Luminal chloride modulates distal tubulebidirectionalbicarbonatefluxin vivo. J. Clin. Invest. 85:1793-1798.

2.Levine,D. Z.1990.Single-nephronstudies:implicationsfor acid-base regu-lation.KidneyInt. 38:744-761.

3.Schuster,V. L. 1989.Physiologyand cellbiology update:control mecha-nisms forbicarbonatesecretion.Am.J.KidneyDis. 13:348-352.

4. Wesson,D. E. 1989.Augmentedbicarbonatereabsorption byboth the

proximalanddistalnephronmaintainschloride-depletemetabolicalkalosis in

rats.J. Clin.Invest.84:1460-1469.

5.DeMelloAires, M.,and G.Malnic. 1972.Micropuncture studyof

acidifica-tionduringhypochloremicalkalosisintherat.PfluegersArch. Eur. J.Physiol.

331:13-24.

6.Wesson,D. E.1990.Depresseddistal tubule acidificationcorrects

chloride-depletedalkalosis inrats.Am.J.Physiol.259:F636-F644.

7.Galla,J.H.,D. N.Bonduris,and R.G. Luke.1989.Superficialdistaland

deepnephronsincorrection ofmetabolicalkalosis.Am.J.Physio!.

257:F107-F113.

8.Levine,D.Z.,D.Roy,G.Tolnai,L.Nash,and B.G. Shah. 1974. Chloride

depletionand_{nephrocalcinosis.}Am.J._Physiol.227:878-883.

9.National Research Council. 1962. Nutrientrequirementsof domestic ani-mals.Nat!.Acad.Sci.PubL.990:68.

10.Bernhart,F.W.,and R. M.Tomerelli.1966.Asalt mixturesupplyingthe National Research Council estimatesofthemineralrequirementsoftherat.J.

Nutr.89:495-500.

11. Ramsay,J. A.,R. H. J.Brown,and P. C.Croghan. 1955.Electronic

titration of chloride in smallvolumes. J.Exp.Biol. 32:822-829.

12. De MelloAires, M.,andG. Malnic..1972. Renalhandlingof sodiumand

potassiumduringhypochloremicalkalosis in therat._{PfluegersArch.}Eur. J.

Phys-iol. 331:215-225.

13.Levine,D.Z.,M.Iacovitti,L.A.Nash,andD.Vandorpe.1988. Secretion of bicarbonatebyratdistal tubulesin vivo.Modulationby overnight fasting.J. Clin.Invest.81:1873-1878.

14.Capasso, G.,P.Jaeger,G.Giebisch,V._Guckian,andG. Malnic. 1987. Renalbicarbonatereabsorptionin therat.II.Distal tubule loaddependenceand effect ofhypokalemia.J. Clin. Invest. 80:409-414.

15.Vandorpe,D. H.,and D. Z. Levine. 1990. Distal tubule bicarbonate

reabsorption JtCO2in vivopersistsduringrebound alkalosis.KidneyInt. 37:548.

(Abstr.)

16.Wesson,D. E.1990.Dietarybicarbonate reducesratdistalnephron

acidifi-cation evaluated in situ.Am.J.Physiol.258:F870-F876.

17.Chan,Y.L.,G.Malnic,and G. Giebisch. 1989. Renal bicarbonate

reab-sorptionin therat. III. Distal tubuleperfusion studyofloaddependenceand bicarbonatepermeability.J.Clin.Invest.84:931-938.

18.Kassirer,J.P., andW. B. Schwartz. 1966. The response of normalmanto

selectivedepletionofhydrochloricacid. Factors in thegenesisofpersistentgastric