Diffusion Potentials and Potassium Distribution the ...dm5migu4zj3pb.cloudfront.net/manuscripts/105000/105032/JCI6410… · john m. dietschy and edwardw. moore mvolts m/s +12* + 8

Journal of Clinical InvestigationVol. 43, No. 8, 1964

Diffusion Potentials and Potassium Distribution across theGallbladder Wall *

JOHN M. DIETSCHY t AND EDWARDNV. MOORE(From the Evans Memorial Department of Clinical Research, Massachusetts Memorial Hos-

pitals, the Department of Medicine, Boston University School of Medicine, the MetabolicLaboratories of the Medical Services, Lemutel Shattuck Hospital, Commonwealth of

Massachusetts, and the Department of Medicine, Tufts University School ofMedicine, Boston, Mass.)

During the process of concentration of bile bythe gallbladder, water and electrolytes are ab-sorbed, leaving behind bile acids, bilirubin pig-ment, and other substances to which the gall-bladder epithelium is not permeable. It has re-cently been postulated (1) that this resorptiveprocess depends upon a "neutral" ion pump in thegallbladder epithelium that transports sodium,anions, and water out of the bile into the blood.Hepatic bile is thus progressively converted froma solution in which the major ions are sodium,chloride, bicarbonate, and bile acid to concen-trated gallbladder bile, which consists predomi-nantly of sodium and bile acid ions. Bile re-mains isosmotic with the serum during this trans-formation in composition. Since the osmotic co-efficient of sodium bile salt is, on the average, onlyabout 0.45 (2), the total concentration of ions inconcentrated bile must be about twice that of se-rum. Typically, whole bile sodium and bile acidconcentrations are of the order of 300 mEq perL each.

During this concentrating process, changes alsooccur in the levels of potassium, which increasefrom 4 to 5 mEq per L in newly formed hepaticbile to 10 to 14 mEq per L in concentrated gall-bladder bile. This increase in potassium concen-tration has been assumed to be due to "secretion"

* Submitted for publication June 14, 1963; acceptedApril 2, 1964.

Supported in part by U. S. Public Health Servicetraining grant 2A-5025 (C6) and by research grantsA-3560 (C3) and AM-07417 from the National Insti-tute of Arthritis and Metabolic Diseases.

This work appeared in abstract form in Clin. Res.1963, 11, 182.

t Present address: Department of Medicine, Uni-versity of Texas Southwestern Medical School, Dallas35, Texas.

of potassium ions by the gallbladder epithelium orto a Donnan effect produced by the nondiffusiblebile acid ions (3, 4). However, if a potential dif-ference exists between the mucosal and serosalsurface of the gallbladder and if potassium isfreely diffusible across the epithelial membrane,then this ion would become distributed between thebile and the blood as predicted by the Nernstequation. The present studies were done to meas-ure these potential differences and to demonstrateif they do indeed affect potassium distributionacross the gallbladder epithelium.

Methods

Gallbladder preparations. Three types of rabbit gall-bladder preparations were used: the isolated everted gall-bladder, the isolated noneverted gallbladder, and the invivo cannulated gallbladder. Adult male rabbits weigh-ing 3.0 to 4.5 kg were anesthetized with 100 to 150 mg ofsodium pentobarbital administered slowly into an earvein. The abdomen was opened, the cystic duct wasclamped and cut, and the gallbladder was stripped up-ward f rom its bed on the liver with the aid of smallscissors. The neck of the gallbladder was cut acrossjust distal to where it joined the cystic duct, and theorgan was immediately everted on a fine glass rod. Itwas then placed in a tray of Tyrode's solution at 370 Cand secured on a tapered glass cannula with silk liga-tures. The mucosal (outside) surface of the gallbladdercould now be exposed to different experimental solutionsby simply transferring the cannulated gallbladder fromcontainer to container. Solutions on the inner, or serosalsurface, were changed by introducing a fine catheterdown the glass cannula into the serosal compartment andaspirating and refilling with a small syringe. The mu-cosal and serosal fluid levels were maintained at equallevels. Care was taken not to overdistend the gall-bladder, since this appeared to change the permeabilitycharacteristics of the epithelial membrane.

Studies were also done on gallbladders that werehandled in an identical manner except that they werenot everted. This made it possible to expose the inner

1551

JOHN M. DIETSCHY ANDEDWARDW. MOORE

MVOLTSM/S

+12*

+ 8

+ 4

0'

- 4,

- 8'

IN VITRO GBI

.

i

I

S.

0 4b 80 120 160 260 mEq/L[SODIUM]

MUCOSAL

FIG. 1. POTENTIAL DIFFERENCES ACROSS THE WALL OF SEVENEVERTED GALLBLADDERSWHENTHE SEROSA WAS BATHED IN TYRODESSOLUTION AND THE MUCOSAWASEXPOSEDTO SULFATE TEST SOLUTIONSWITH SODIUM CONCENTRATIONSVARYING FROM20 TO 220 MEQPER L.The solid line represents the mean values.

or mucosal surface to small volumes (0.75 to 1.50 ml)of such test solutions as whole bile. Systematic deter-mination of permeability constants for several ions inboth the everted and noneverted gallbladders gave identi-cal results, indicating that eversion did not alter thepermeability characteristics of the epithelial surface.All in vitro studies were done in a constant temperaturewater bath at 370 C.

In vivo preparations were made in rabbits maintainedon intravenous sodium pentobarbital anesthesia. From10 to 50 mg of the barbiturate was administered intoan ear vein each hour depending upon the apparent levelof anesthesia in the individual animal as judged by itsrespirations and corneal reflex. Animals could be main-tained for up to 4 hours in this manner. After openingthe abdomen the cystic duct was isolated from the cysticartery and tied. A small incision was made distal to thetie, and a polyethylene catheter was inserted into theapex of the gallbladder until it protruded 1 to 2 mmintQ the lumen. Silk ligatures were used to secure thecatheter in place. The mucosal surface could then bewashed and exposed to different test solutions, and thecatheter could be clamped to leave the test solution inthe gallbladder for varying periods of time.

Solutions. The in vitro gallbladders were prepared inmodified Tyrode's solution gassed with 95% 02 and 5%C02. The composition of this solution in millimoles perliter is NaCl, 137; NaHCO3, 12; KC1, 2.9; CaCI2, 1.8;MgC12, 0.5; NaH2PO4, 0.4; and glucose, 11.1. It had an

osmolality of 292 + 5 mOsmper L and a pH of 7.4 +

0.2 after gassing. Tyrode's solution was also used asthe serosal bathing fluid in both the everted and non-everted in vitro gallbladder preparations.

Two types of mucosal solutions were used, a sulfate-containing test solution and rabbit whole bile. Sinceconcentrated bile is composed mainly of the sodium saltof a nondiffusible anion (bile acid), this chemical situa-tion was approximated with solutions containing the so-dium salt of another essentially nondiffusible anion, sul-fate. The sulfate test solutions contained Na2SO4 inconcentrations varying from 20 to 220 mEq per L andK2SO4 at a constant concentration of 3.5 mEq per L.Mannitol was added to give each test solution a final os-molality of 295 ± 5 mOsmper L, which equals the meanosmolality of rabbit serum. Hereafter, the expression"sulfate test solution" always refers to this isosmoticmixture of sodium and potassium sulfate. Bile samplesof varying ionic composition were obtained by aspirationfrom the gallbladders ofi rabbits fasted for 6 to 30 hours.These samples were frozen at - 150 C until used.

Potential measurements. Polyethylene catheters (i.d.,0.034 inches) were filled with a 3%o agar, saturated KC1solution. In the in vitro experiments one such salt bridgewas placed in the outer solution, and another was inserteddown the glass cannula into the inner solution. In thein vivo preparations one bridge was inserted into the gall-bladder cannula to make contact with the mucosal solu-tion, and the other was applied directly to the serosa of

1552

POTASSIUMDTSTRtBUTION ACROSSTHE GALLBLADDERWALL

MVOLTS0

M/S

* IN VIVO

-4 IN VITRO

-8

-12

*~ ~-16

-20 0

-24140 180 220 260 300 340 mEq/L

[SODIUM]BILE

FIG. 2. POTENTIAL DIFFERENCES ACROSSTHE GALLBLADDER WALL PLOTTEDAS A FUNCTION OF THE BILE SAMPLE SODIUM CONCENTRATIONON THE MU-

COSAL SURFACE. The open circles represent data obtained in noneverted, invitro gallbladders where the serosa was bathed in Tyrode's solution. Thesolid circles are for bile samples placed in the in vivo gallbladder. Thesolid line represents the approximate mean of these points.

the gallbladder. To avoid damage to the gallbladderpreparations from potassium diffusion out of the bridges,the KCl-filled catheters were left in place only longenough to obtain reliable voltage measurements; usuallyless than 1 minute was required to reach a stable reading.The salt bridges led to two balanced calomel cells, whichwere connected to a high impedance volt meter.' Ap-propriate corrections were made for day to day electricalimbalance in the system (±+0.5 mv). The sign of thepotentials is always expressed in terms of the mucosalvoltage with respect to the serosa. Variation from themean values is given as 2 SD (95% confidence limit).

Activity determinations. The activities of sodium andpotassium in various mixed solutions were measured bymeans of sodium and potassium specific glass electrodes,respectively.2 The electrodes used are described in de-tail elsewhere (2, 5). Activity coefficients (,y) were cal-culated by dividing the activity of a particular ion, as

determined by the appropriate electrode, by the concentra-tion of that ion, as determined by flame photometry.

ResultsPotentials in the in vitro gallbladder. With the

everted gallbladder the serosal surface was bathedin Tyrode's solution while the mucosal surface

1 Radiometer, Copenhagen, Denmark.2 Obtained from Orion Research, Inc., Cambridge, Mass.

was exposed to sulfate test solutions with sodiumconcentrations varying from 20 to 220 mEq perL. The potentials that developed in seven suchgallbladder preparations are plotted in Figure 1as a function of the sodium concentration in themucosal test solution. At a mucosal sodium con-centration of 20 mEq per L the mucosa was14.3 + 3.6 mv positive to the serosa. The voltagedropped to - 7.9 2.4 mv as the mucosal sodiumlevel was raised to 220 mEq per L. Similar po-tentials were obtained in three gallbladders thatwere immersed in Tyrode's solution containingiodoacetate, 5 x 105 M, for 10 minutes before ex-

posure to the test solutions.Potential differences were also measured when

the mucosa of the noneverted, in vitro gallbladderwas exposed to rabbit whole bile. In these stud-ies Tyrode's solution was again the serosal fluid.In Figure 2 the potential differences that devel-oped across the wall of seven such preparationsare shown as open circles and -plotted as a func-tion of the bile sodium concentration on the mu-

cosal surface. These values varied from - 14.2to - 19.0 mv for bile samples with sodium con-

1553

JOHN M. DIETSCHY AND EDWARDW. MOORE

MVOLTSM/S

+12 . IN VIVO GB\

+ 8

+4

0~~~~~~~

4 0

-8

-120 40 80 120 160 200 mEq/L

[SODIUM]MUCOSAL

FIG. 3. POTENTIAL DIFFERENCES ACROSSTHE WALL OF FIVE ift ViVOGALLBLADDERSTHAT WERESEQUENTIALLY FILLED WITH SULFATE TEST

SOLUTIONS WITH SODIUM CONCENTRATIONSVARYING FROM 20 To 220MEQPER L. The solid line represents the mean values.

centrations that varied f rom 250 to 335 mEqper L.

Potentials in the in vivo gallbladder. To dem-onstrate that similar potentials exist in the in vivogallbladder, sulfate test solutions and whole bilesamples were introduced into gallbladders left insitu, and potential measurements taken. In thisexperimental preparation the animal's blood per-fused the serosal surface of the epithelial mem-brane. Figure 3 illustrates the magnitude of thepotentials that developed in five in vivo gallblad-ders which were sequentially filled with a series ofsulfate test solutions of varying sodium concen-trations. These potentials fell from + 14.6 + 3.2to - 7.4 ± 1.4 mv as the mucosal sodium concen-tration was progressively increased from 20 to220 mEq per L. When the in vivo gallbladderwas filled with whole bile, potentials developedbetween the mucosa and serosa that varied from- 6.5 to - 18.0 mv as the sodium concentrationranged from 180 to 330 mEqper L in the variousbile samples. The data for 20 such experimentsare plotted as the solid circles in Figure 2.

Thus, when either sulfate test solution or wholebile was placed in the functioning gallbladder in

vivo, potential differences developed between themucosal and serosal surfaces. When the intralu-minal solution had a very low sodium concentra-tion, the mucosa became positive with respect tothe serosa; however, over the entire range of so-dium concentrations encountered physiologically inbile, the mucosal surface was negative with respectto the serosal surface.

Potassium concentration in the mucosal solu-tion at equilibrium. The experiments cited indi-cated that the magnitude and sign of the potentialdifference across the gallbladder epithelium variedwith the sodium concentration in the mucosal so-lution. If the distribution of potassium across thegallbladder epithelium were influenced by thesepotential differences, then the potassium concen-tration in the mucosal solution at equilibriumshould also vary with the mucosal sodium con-centration.

To test this hypothesis a series of cannulatedin vivo gallbladders was filled with measured vol-umes of isosmotic sulfate test solutions of varyingsodium concentrations but a constant potassiumlevel of 3.5 mEq per L. After 1 hour each test

sample was removed, net volume changes were

1554

POTASSIUMDISTRIBUTION ACROSSTHE GALLBLADDERWALL

CONCENTRATIONK 16

mEq/L14

12

10

8

6

4

2

* SULFATEo BILE

0

O~~~~ 0° 0

0 0 0 0

0 0 0

00 0 0

000

0

S *

;

.

0 50 100 150 200 250 300 350

[NA] mEq/LMUCOSAL

FIG. 4. POTASSIUM CONCENTRATIONIN THE GALLBLADDER CONTENTS AT

EQUILIBRIUM. -The solid circles represent the potassium concentrationsreached in sulfate test solutions placed in the in vivo gallbladder for 1 hourplotted against the sodium concentration in each test solution. The opencircles are the equilibrated potassium concentrations in bile samples plottedagainst the bile sodium concentration.

noted, and potassium concentrations were deter-mined. Because the test solutions contained sul-fate and mannitol, neither of which readily pene-trates the epithelial membrane, no net electrolyteor water absorption occurred from the gallbladderduring the test period. Hence, after 1 hour thevolume of the test solution either remained un-changed or increased by a slight amount (up to5%o of the original volume). The potassium con-centration shifted up or down from the initialvalue of 3.5 mEq per L depending upon the so-dium concentration (and hence the potential ofthe mucosa) in the test solution. Such changesusually occurred within 30 minutes, but the testperiod was prolonged to 1 hour to make certainthat potassium equilibration across the epithelialsurface was complete. The solid circles in Fig-ure 4 show the concentration of potassium achievedat equilibrium in each mucosal solution plottedagainst the test solution sodium concentration.In samples containing only 20 mEq per L of so-dium, potassium concentrations dropped to 0.9 to1.7 mEqper L. The potassium level rose as thesodium concentration in the test solution was in-

creased and reached 5.2 to 6.0 mEq per L whenthe mucosal sodium concentration was 220 mEqper L.

Similar experiments were done to demonstratethat the same phenomenon occurred when the invivo gallbladder was filled with whole bile. Theopen circles in Figure 4 represent the equilibratedpotassium concentration plotted against the sodiumconcentration in 37 bile samples. The same trendis apparent; the higher the sodium concentration(and the more negative the mucosa), the higher isthe potassium concentration at equilibrium. Di-rect potential difference measurements duringthese equilibration experiments were not possible,since the salt bridges leaked potassium into thetest solutions.

Calculated potassium activities at equilibrium.If potassium distribution across the in vivo gall-bladder wall is dictated solely by passive trans-port phenomena, then at 370 C the following formof the Nernst equation should apply (6): E =61.5 -log (K activity in serum)/(K activity ingallbladder). Since the volume of the blood isvery large compared to the volume of the solution

t j55v

S :

a

JOHN M. DTETS-CHY AND EDWARDW. MOORE

ACTIVITYK 16

mEq/L14

12

10.

8

6

4,

2-

0 *

* SULFATEo BILE

THEORETICAL

THEORETICALK ACTIVITY

IN S04 SOL.\

_

0

0 50 100 150 200 250 30(

ENA]MUCOSAL

FIG. 5. OBSERVEDPOTASSIUM ACTIVITY AT EQUILIBRIUM PLOTTED AGAINST

THE SODIUM CONCENTRATIONIN THE SULFATE TEST SOLUTION OR BILE SAMPLE.

The theoretical activity curves for the sulfate test solutions and for bilesamples were derived from the Nernst equation as described in the text.

contained within the gallbladder, any changes inthe ratio (K activity in serum)/ (K activity in gall-bladder), induced by a potential difference (E),will be reflected by changes in the potassium ac-

tivity in the gallbladder contents while the serum

potassium activity remains constant. In bloodsamples taken from ten rabbits during the experi-mental periods the mean serum potassium concen-

tration was 3.9 + 0.4 mEq per L. Since the ac-

tivity coefficient of potassium at this concentrationin the presence of the other electrolytes in serum

is about 0.75 (2), this corresponds to a mean se-

rum potassium activity of 2.9 + 0.3 mEqper L.For any particular sodium concentration in a

bile sample or sulfate test solution placed in thein vivo gallbladder the expected potential differ-ence across the gallbladder wall can be obtainedfrom the mean experimental curves in Figure 2and Figure 3, respectively. With the values ofthe potential difference and the mean serum potas-sium activity known, the Nernst equation can beused to calculate the expected potassium activityin the gallbladder contents at equilibrium. Forexample, when a solution containing 20 mEq perL of Na2SO4 is placed in the gallbladder, the mu-

cosa is about 14.3 mv positive to the serosa (Fig-ure 3). The calculated potassium activity in thismucosal test solution at equilibrium is 1.7 mEqper L. For a bile sample with a sodium concen-

tration of 300 mEq per L the expected potentialis about - 16.0 mv (Figure 2), and the calculatedpotassium activity in the bile at equilibrium is 5.3mEq per L. One can calculate the expected po-

tassium activities over the entire range of sodiumconcentrations in the sulfate test solutions or bilesamples; such calculations give the two theoreticalcurves shown in Figure 5.

Experimental potassium activities at equilib-rium. The data in Figure 4 show the experi-mentally determined potassium concentrationsfound in the gallbladder contents at equilibrium.To compare these individual points to the theo-retical curves in Figure 5 each potassium concen-

tration value must be converted to potassium ac-

tivity. This requires knowledge of the activitycoefficients of potassium in the presence of varyingconcentrations of sodium, sulfate, and bile acidions. Such data have been determined by means

of selective sodium and potassium glass electrodesfor sodium and potassium sulfate solutions, as

0 350

mEq/L

1556

POTASSIUMDISTRIBUTION ACROSSTHE GALLBLADDERWALL

well as for solutions of the sodium and potassiumsalts of a variety of free and conjugated bileacids (2).

From the data thus obtained it is possible to ap-proximate activity coefficients in mixed solutionsof sodium and potassium sulfate and of sodiumand potassium deoxycholate (Figure 6). For ex-ample, in a solution containing 5 mEq per L ofK2SO4 and 150 mEq per L of Na2SO4, the ac-tivity coefficient for the potassium ion is about0.647. In a similar solution of 5 mEq per L ofpotassium deoxycholate and 150 mEqper L of so-dium deoxycholate, the activity coefficient of thepotassium ion is about 0.539. These, of course,are activity coefficients measured in pure solutions.In the sulfate test solutions left in the in vivo gall-bladder for 1 hour other ions such as chloride andbicarbonate enter the solutions, causing some de-viation from these predicted activity coefficients.However, the deviation should be small, sincethese ions are present in relatively low concen-trations. Concentrated rabbit bile also containssmall amounts of calcium, chloride, bicarbonate,and other ions; however, the concentration ofthese substances is also low compared to sodiumand bile acid ions.

The predominant bile acid present in rabbit bileis the glycine conjugate of deoxycholic acid (7).Since activity coefficient values do not appear tobe affected by conjugation (2), potassium activitycoefficients in rabbit whole bile are approximatedby values obtained in pure deoxycholate solutions.Thus, in the conversion of the observed potassiumconcentration values (Figure 4) to appropriateactivity values each bile sample was considered apure mixture of sodium and potassium deoxycho-late, and activities were calculated by using theactivity coefficients presented in Figure 6. Simi-larly, each sulfate test solution was considered apure mixture of sodium and potassium sulfate,and the conversion of potassium concentration atequilibrium to potassium activity was effected bythe use of the appropriate activity coefficient fromFigure 6.

The resulting "observed" potassium activitiesin sulfate test solutions and in 37 native bile sam-ples are plotted in Figure 5. There is close cor-relation between these observed potassium activi-ties and the calculated values (solid lines) predicted

[K]mEqAL S

10.

5'

01

20[K]

mEq/L

10.

5.

50 100 i1o 200 250 *300 350[N A mEq/L

0 50 100 150 200 250 300 350[NA] rnEq/

FIG. 6. ACTIVITY COEFFICIENTS OF POTASSIUM IN

MIXED SOLUTIONS OF SODIUM AND POTASSIUM SULFATE

AND IN SODIUM AND POTASSIUM DEOXYCHOLATE (BILESALT). For a particular ionic mixture, project linesperpendicular to the ordinate and abscissa from the ap-propriate concentrations; the point of interception of thetwo lines gives the activity coefficient for potassium inthe mixed solution (2).

from the Nernst equation. The data indicatethat over a wide range of potential differencesacross the gallbladder wall potassium becomes dis-tributed between bile and serum according to ex-isting electrochemical gradients.

DiscussionThese experiments demonstrate the relationship

between the transmucosal potential difference inthe gallbladder and the bile potassium concen-tration. Although not specifically investigated inthe present study, enough information is nowavailable concerning ion movement in the gall-bladder to allow speculation as to the source ofthis potential difference. At least three phenom-ena may result in electrical potentials across bio-logic membranes: 1) active ion transport acrossthe membrane (transport potentials), 2) trans-membrane osmotic gradients (streaming poten-tials), and 3) transmembrane ionic activity gradi-ents (diffusion potentials). It is unlikely thateither of the first two possibilities, i.e., transportor streaming potentials, accounts for the potentialdifferences measured in these experiments. Al-though the gallbladder epithelium actively trans-ports sodium, unlike many other epithelial sur-

SU'LFATE

0

04

BILE SALT \ \

.',I

1557

JOHN M. DIETSCHY AND EDWARI) W. MOORE

faces it develops essentially no transport potential.For example, the fish (1) and everted rabbit gall-bladder (8) transport electrolytes, but when bathedon both surfaces with identical physiologic solu-tions only a few millivolts potential difference de-velops across the epithelial surface. This ob-servation apparently is explained by interde-pendent transport of sodium and anions resultingin a "neutral" pump mechanism (1). Further-more, iodoacetate, which is known to stop activesolute transport in the gallbladder ( 1, 8), does notalter the magnitude of the potential differencesmeasured in in vitro gallbladders in this study.Streaming potentials have been measured acrossthe isolated gallbladder when water movement isinduced by large osmotic gradients across the epi-thelial surface (8, 9). In the present study, how-

ever, no such gradients existed, since all test solu-tions were kept isosmotic with blood.

In the various experimental situations largedifferences in ionic composition usually existedbetween the sulfate test solutions or whole bilebathing the mucosal surface of the gallbladder andthe Tyrode's solution or animal's blood perfusingthe serosal surface. These transmembrane ionicactivity gradients undoubtedly result in the pro-duction of diffusion potentials that probably ac-count for the potential differences which weremeasured in the present study. If this is true, theHodgkin-Katz equation (10) should predict themagnitude of the potential difference (6). Withthe major univalent ions in serum, bile, and thesulfate test solutions considered, the equationtakes the following form at 370 C (1):

E = 61.5 log (Na)8 (yNa) (PNa) + (K). (YK) (PK) + (Cl)m (Yci) (Pc1) + (BS)m (YBS) (PBS)(Na)mQ('YNa)(PNa) + (K)m,(YK)(PK) + (C1) (-Yci)(Pci) + (BS)8Q(YBS)(PBS)'

where E is the potential difference; (ion)8. and(ion)m refer to ionic concentration of the serosaland mucosal fluid, respectively; (BS) representsbile salt; (y) is the activity coefficient of the ionin its respective solution; and (P) is the perme-ability constant for each ion relative to sodium.

anU)

IE.0

E

to

(0

0

E

8 .6 .4 .2

(Na]Slog

[NO]MIN vIWRO

20 'I,

5 \\ Predicted

\w Curve

10-

5-

0~~~~~0 x

Experimental-5- Curve

-10~~~~~~~- lo

1.0 .8 .6 .4 .2i -.?-

[Nao]

IN V VO

FIG. 7. POTENTIAL DIFFERENCES ACROSS THE WALL OF

THE GALLBLADDERFILLED WITH SULFATE TEST SOLUTIONS.

The potential difference is plotted against the log of theratio of the serosal to mucosal sodium concentration.See legend to Table I for detailss of the calculations of

the predicted curves.

The relative permeability constants in the rabbitgallbladder have been determined (8), and forNa+: K+: Cl-: BS- they equal 1.00:1.50:0.33:0.00.

A comparison of the potential differences foundexperimentally and those calculated from theHodgkin-Katz equation is shown in Figure 7 andTable I. The data in Figure 7 demonstrate theclose agreement between the measured and theo-retical potential differences in in vitro and invivo gallbladders filled with sulfate test solutions.Actually the correlation between these two sets ofcurves is probably even closer than shown in Fig-ure 7 because, for the purpose of calculation, thegallbladder mucosa was considered to be com-pletely impermeable to sulfate ion, i.e., in theHodgkin-Katz equation Pso4 was considered equalto zero. This assumption is not strictly true forin the rabbit sulfate ion does penetrate the gall-b)ladder epithelium to a limited degree (8).3 Be-cause of this sulfate "leak," the magnitude of the

3 When the rabbit gallbladder is perfused on both sur-faces by sodium sulfate solution, a slope of only 50.0mv rather than 61.5 mv (at 370 C) is obtained for atenfold change in sodium activity; thus, in contrast to thefish, the rabbit gallbladder does not l)ellave in this situa-tion as a sodium electrode, hut, rather, sulfate ion mustpenetrate the lmucosatl surface to sonie degree.

1558

POTASSIUMDISTRIBUTION ACROSSTHE GALLBLADDERWALL

calculated potentials would be lower so that inFigure 7 the predicted curves would even more

closely approximate the experimental curves. InTable I the comparison is made for in vivo andin vitro gallbladders filled with whole bile. Hereagain there is close agreement between the poten-tial difference measured experimentally across thegallbladder wall and the magnitude of the diffu-sion potentials calculated from the Hodgkin-Katzequation. These data strongly suggest that ionicactivity gradients with consequent diffusion po-

tentials are the principal source of the potentialdifferences measured in the present study.

As bile becomes progressively concentrated, itssodium concentration rises to around 300 mEqper L, and its chloride concentration drops to levelsas low as 5 to 10 mEq per L. Thus, in the insitu gallbladder the largest activity gradients be-tween serum and bile are for sodium (about 100mEqper L) and for chloride (about 70 mEqper

L). Since the epithelial surface is relatively per-

meable to both of these ions, both significantly in-fluence the magnitude of the potential difference.Other ions in serum and bile are either presentin such relatively low concentrations on both sidesof the membrane or else have such low perme-

ability constants, e.g., the divalent ions and bilesalt, that they contribute only slightly or not atall to the voltage difference between the mucosaland serosal surfaces.

In these experiments, over a wide range of po-

tential differences, potassium distribution betweenserum and bile was always predictable from theNernst equation. This suggests passive transportof the ion secondary to electrochemical forces.In no instance did potassium move against an elec-trochemical gradient so that there was no evidenceof an active secretaryy" mechanism for this ion.This observation is supported by the recent reportof Wheeler (11) that the flux ratio for potassiumion across the rabbit gallbladder wall is consistentwith passive diffusion of the ion.

The sequence of events that occurs in the gall-bladder during the process of concentration ofbile may be outlined as follows. Newly formedhepatic bile is delivered into the gallbladder as an

isosmotic solution with sodium the principal cationand chloride, bicarbonate, and bile acids the prin-cipal anions. Active solute absorption by the gall-

TABLE I

Potential differences across the wall of thebile-filled gallbladder*

Bile composition Mucosal potential

Na K Cl Measured Calculated

mEqILIn vitro 189 8.5 37 - 8.7 - 7.4

216 11.0 21 -10.0 -11.5280 12.1 9 -16.2 -16.5321 13.9 5 -18.9 -19.7

In vivo 260 8.0 40 -14.1 -12.3325 12.1 10 -18.2 -19.1300 9.2 12 -16.1 -16.6340 13.4 5 -19.8 -20.4

* These data compare the potential differences measured across thewall of the in vitro and in vivo gallbladders filled with whole bile andthose calculated from the Hodgkin-Katz equation. For these calcula-tions the relative permeability constants used were 1.00: 1.50: 0.33:0.00for Na+:K+: Cl: BS-. Cationic activity coefficients (-Y+) were deter-mined with glass electrodes (2); the values used for 'YNa+ in Tyrode'ssolution and rabbit blood were 0.755 and 0.750, respectively; 'yNA+ inwhole bile varied from 0.60 to 0.67. -ycr- and -YK+ were considered equalto the value of YN,+ in Tyrode's solution and blood, but in whole bile,the potassium activity coefficient varied from 0.45 to 0.55.

bladder epithelium decreases the chloride and bi-carbonate concentration and raises the level of so-dium bile salt to about 300 mEq per L. Ionicactivity gradients between serum and bile, prin-cipally for sodium and chloride, cause the mucosato become 16 to 19 mv negative to the serosa.Because of this electrical gradient, the activity ofpotassium in bile must be about twice as high asin serum, if this passively transported ion is toremain in electrochemical equilibrium across thegallbladder epithelium. Because the activity coeffi-cient of potassium in whole bile is lower than theactivity coefficient of potassium in serum the bilepotassium concentration is about three times thatof the serum.

Summary

During the process of bile concentration by therabbit gallbladder, the potassium concentration in-creases from 4 to 5 mEq per L in newly formedhepatic bile to 10 to 14 mEqper L in concentratedgallbladder bile. This increase in potassium con-centration has been assumed to represent "secre-tion" of potassium ions by gallbladder epitheliumor a Donnan effect produced by nondiffusible bileacid ions. The purpose of this study was to de-termine whether this bile-serum concentrationgradient for potassium could be explained byelectrical potentials across the membrane. Po-tential differences were measured across in vivo

1559

JOHN M. DIETSCHY AND EDWARDW. MOORE

and in vitro rabbit gallbladders and were foundto be directly related to sodium concentration inthe mucosal bathing solution. Theoretical bile-serum potassium distribution was then calculatedfrom the Nernst equation. Potassium activities insulphate test solutions and bile were estimatedf rom previously derived activity coefficients de-termined by sodium- and potassiumn-selective glasselectrodes. A close correlation was found betweenobserved potassium activities and those predictedby the Nernst equation according to existing elec-trochemical gradients. There was no evidence ofan active secretaryy" process. The data suggestthat the high potassium levels in concentrated gall-bladder bile result from passive transport phe-nomena.

Acknowledgment

Wethank Dr. Franz J. Ingelfinger for his helpful sug-

gestions and comments during the course of these ex-

periments and the preparation of the manuscript.

References1. Diamond, J. M. The mechanism of solute transport

by the gall-bladder. J. Physiol. (Lond.) 1962,161, 474.

2. Moore, E. W., and J. M. Dietschy. Na and K ac-tivity coefficients in bile and bile salts determinedwith glass electrodes. Amer. J. Physiol., in press.

3. Herman, R. H., T. H. Wilson, and L. Kazyak.Electrolyte migrations across the wall of theguinea pig gall bladder. J. cell. comp. Physiol.1958, 51, 133.

4. Sobotka, H. Physiological Chemistry of the Bile.Baltimore, WVilliams & Wilkins, 1937.

5. Moore, E. WV., and D. W. Wilson. The determina-tion of sodium in body fluids by the glass elec-trode. J. clin. Invest. 1963, 42, 293.

6. Ussing, H. H. The Alkali Metal Ions in Biology.Berlin, Springer-Verlag, 1960.

7. Haslewood, G. A. D., and U. Wootton. Comparativestudies of "bile salts," I. Preliminary survey.Biochem. J. 1950, 47, 584.

8. Dietschy, J. M. Water and solute movement acrossthe wall of the everted rabbit gallbladder. Gas-troenterology, in press.

9. Diamond, J. M. The mechanism of water transportby the gall-bladder. J. Physiol. (Lond.) 1962, 161,503.

10. Hodgkin, A. L., and B. Katz. The effect of sodiumions on the electrical activity of the giant axon ofthe squid. J. Physiol. (Lond.) 1949, 108, 37.

11. Wheeler, H. 0. Transport of electrolytes and wa-ter across wall of the rabbit gall bladder. Amer.J. Physiol. 1963, 205, 427.

1560