Exp. Eye Res. (19921 54. 921-931

Cl-. Transport

in Frog Retinal MORTEN

Department

of General

(Received

Physiology Blegdamsvej

Houston

Pigment

LA COUR

and Biophysics, University 3, DK-2200 Copenhagen

7 1 February

Epithelium

7991 and accepted

of Copenhagen, N, Denmark

in revised

Par-urn Institute,

form 29 July

1991)

Cl transport across the retinal membrane of the frog retinal pigment epithelium was studied by means of double-barrelled Cll selective microelectrodes. Three types of experiments were performed. In the first group of experiments, the ionic dependence of Cl- influx across the retinal membrane was studied. The intracellular CIl activity was first decreased by perfusing the retinal side of the epithelium with low Cl solutions ( 3.6 rnM Cl-): then the perfusate was changed to high Cl- solutions (90.1 mM1. and the resulting Cll influx was studied. In these experiments. the combined presence of extracellular Na+ and K’ was a necessary condition for CIl influx across the retinal membrane. This supports the hypothesis of Na+.K*.Cl co-transport across this membrane. In a second group of experiments, the effect of furosemide was studied. Furosemide (100 pi) inhibited Cl influx when the retinal extracellular CIl concentration was increased from 3.6 to 90.1 mM. When administered to cells in steady state. furosemide in concentrations between 5 and 1000~~ decreased the intracellular Cll activity. Michaelis-Menten analysis yielded a Ki for furosemide of 7 + 2 ,ILM.The effect of furosemide on the intracellular Cl- activity required the combined presence of extracellular Na+ and K+. When the retinal extracellular K+ concentration was increased to between 0 and 10 mM, the furosemidesensitive Cl- influx across the retinal membrane increased. Michaelis-Menten analysis yielded a half maximal stimulation at an extracellular K’ concentration of 0.5 mM. Stimulation of the epithelium with 1 mM CAMP and 0.5 mM IBMX reduced the effect of furosemide on the intracellular Cll activity by 2h0/,. In a third group of experiments, the effect of transepithelial currents on the intracellular Cll activity was investigated. Currents that depolarized the choroidal membrane potential increased the intracellular Cl activity: currents that hyperpolarized this membrane potential decreased the intracellular Cl activity. These findings are compatible with conductive Cl transport across the choroidal membrane. The apparent Cl- conductance of this membrane was estimated to be 0.59 mS cm-?, This represents 27”/, of the total conductance in the choroidal membrane. Administration of 1 rnM CAMP and 0.5 rnM IBMX caused a 2 1 (jj increase in the apparent Cl- conductance of the choroidal membrane. Key words: Na+,K+,Cl- co-transport : conductive Cl transport: furosemide; cyclic AMP ; retinal pigment epithelium : frog.

1. Introduction The retinal pigment epithelium

(RPE) is a single cell

layer in the back of the vertebrate

eye. It separates the

neurosensory retina from the choriocapillaries and is an important part of the blood retinal barrier. Transport of fluid. ions, and metabolites across the RPE is crucial for retinal homeostasis. The isolated frog RPE-choroid preparation has been shown to transport Cl- in the retina-to-choroid direction (Miller and Steinberg, 1977a; DiMattio, Degnan and Zadunaisky. 1983: Miller and Farber, 1984): administration of furosemide to the retinal bath blocked this Cl- absorption (DiMattio et al., 1983) and CAMP stimulation of the epithelium reduced it by 50% (Miller and Farber, 1984; Hughes, Miller and Farber, 1987). Furosemide (or bumetanide) sensitive Cll absorption has also been described in isolated RPEchoroid preparations from a variety of other species including dogs (Tsuboi, Manabe and Iizuka, 1986). chick embryos (Frambach and Misfeldt, 1983) and cows (Frambach, Valentine and Weiter, 1989; Miller and Edelman, 1990). Intracellular measurements in frog RPE cells showed that the intracellular Cl- activity 0014--1835/92/0(,0921+

11 $03.00/O

was above electrochemical equilibrium and retinal furosemide administration caused a decrease in this activity (Wiederholt and Zadunaisky. 198 5). Since furosemide is an inhibitor of Na+:Cll co-transport, these results indicate that some form of Nat-coupled Cl- transport is responsible for the influx of Cl- across the retinal membrane. Furosemide-sensitive Na’ : Cll co-transport is known both in a K’-independent form as Na’,Cll co-transport (Musch and Field, 1989) and as Na+,K+,Cl- cotransport (O’Grady, Palfrey and Field. 1987). Although some indirect evidence has recently been presented (Adorante and Miller, 1990 : Kennedy, 1990; Miller and Edelman, 1990) it is not known whether Cl- influx across the retinal membrane in the frog RPE is K’ dependent. This is of particular interest since Na’,K+,Cl- co-transport could influence the retinal extracellular K+ activity, and thereby be important for K’ homeostasis in the extracellular space between the photoreceptors and the RPE. Little is known about the transport mechanisms for Cll in the choroidal membrane. It has been suggested that conductive Cll transport across the choroidal membrane is responsible for Cll exit across this 0 1992 Academic Press Limited

922

M. LA COUR

Soltrtiarts (concentratiom in mu) -__

Solution Solution Solution

1 2 3

Solution 4 Solution Solution Solution Solution Solution Solution Solution Solution Solution Abbreviations:

5 6 7 8 9 10 11 12 13

Na-

K-

110 110 0 1 10 0 110 110 110 110 110 110 0 0

2 2 2 0 2 0 0.5 1 5 10 2 2 2

Mg”-

pj2-

1 1 1 1 1 1 1 1 1 1 1 1 1

1.X 1.X 1.X 1.X 1.8 1.x 1.8 1.8 1.8 1.8 1.8 1.X 1.8

Glu. glucose: Glnt. gluconate: NMDGH’.

NMIXX 0 0 110 0 110 0 0 0 0 0 0 110 110

N-methyl-o-glucamin

membrane (DiMattio et al., 1983). In isolated chick retinas, electrophysiological evidence for the existence of a Cl- conductance in the choroidal membrane of the RPE has been reported (Gallemore and Steinberg, 1989). In the isolated bovine RPE preparation Clabsorption is blocked by millimolar concentrations of DEDS in the choroidal bath (Miller and Edelman, 1990). It was the purpose of the present study to apply double-barrelled ion-selective microelectrodes to investigate the hypothesis of Na+,K+,Cl- co-transport across the retinal membrane of the frog RPE. In addition, an attempt was made to quantify the rate of this co-transport and its dependence on the retinal extracellular K’ concentration and the effects of CAMP stimulation. Further, the hypothesis of conductive Cltransport across the choroidal membrane was investigated: the apparent Cl- conductance of this membrane was estimated and the influence of CAMP stimulation determined.

2. Materials and Methods Microelectrodes Double-barrelled Cl- selective microelectrodes were made as described by Zeuthen (1980). The ionselective barrel of the microelectrode was filled with Corning cocktail 477913 (Corning Medical, Medifield, MA) and the reference barrel was filled with 0.5 M potassium gluconate or 0.5 M potassium acetate. A W + W multi-channel chart recorder (Kontron Elektronik GmbH. Miinchen, Germany) was used to record the signals from the reference barrel and the Cl-selective barrel of the microelectrode. The signal from the Cl--selective barrel was electronically subtracted from the reference barrel signal before being recorded. The Cl- electrodes were calibrated in solutions that approximated intracellular fluid: 105 mM K’, 15 mM Na’, 15 mM HCO,-, 5, 10, 20 and 40 mM Cl-, and

C’l

HCO,,

Glnt

Hepes

90.1 3.6 90 1 90~ 1 3.6 3.6 901 90.1 90.1 90.1 113.2 I 13.2 3.6

27.5 27.5 27.5 27.5 2 i.5 27.5 27.5 2i.5 27.5 27.5 0 0 0

0 X6.5 0 0 86.5 86.5 0 0 0 0 I) 0 111

0 0 0 0 0 0 0 0 0 0 10 10 10

C&l

5 5 5 i 5 5 5 i ; 5 5 5

acid.

100, 105, 95 and 75 mM gluconate. In experiments where the tissue was perfused with Hepes buffers (solutions 11, 12 and 13. Table I) the Cl- electrodes were calibrated in HCO,- free calibrating solutions (NaHCO, was replaced with NaHepes). The difference AE in the electrical potential between the control solution (solution 1, Table I) and the calibrating solutions was fitted by linear regression to the log Clconcentration : AE = E, - S log ([Cl-]. 90.1-l 1, S is the sensitivity (Nernstian slope) of the Cl- electrodes, and averaged 47.0 f 0.7 mV (mean 2 s.E.M.. II = 2 5 1. E,, is an empirical constant that contains the HCO:, selectivity factor and the concentration difference for this ion between the control solution and the calibrating solution : E, averaged - 0.41& 0.73 mV (mean+s.E.~., n = 22). The intracellular Cl- activity (measured as apparent concentration, LI, Cl-) was determined by the equation: a,& = 90.1 exp,, [(E,, - AE’) . s-l]. where AE’ is the difference in electrical potential between the control solution and the intracellular fluid. In RPE cells, the average intracellular HCO,- concentration is close to that of the calibrating solutions (la Cour, 1989). Thus the influence of the HCO,- selectivity factor for the steady state measurements of a$- was accounted for by the constant E,,. The HCO,- selectivity factor was not routinely determined. In six electrodes this coefficient was estimated to be 0.12 f 0.01 (average_+ s.E.M.1 by non-linear regression to the Nicholsky-Eisenman equation. In the same way the gluconate selectivity factor was estimated to be 0.00 5 i 0.009 (average + II = 4). The influence of the gluconate S&M., selectivity factor was ignored in the electrode calibration. Preparation Large bull-frogs (Ram cutesbeiunu) were obtained from Carolina Biological Supply Co., Burlington. NC. Frogs were dark-adapted and killed by decapitation.

Cl

TRANSPORT

IN FROG

923

RPE

The eyes were enucleated, the RPE and choroid were excised and mounted in a small Ussing chamber as previously described (la Cour. Lund-Anderson and Zeuthen, 198(i). and a surface area of 0.07 cm? of epithelium was exposed. The transepithelial potential was recorded by Ag/AgCl electrodes connected to the retinal and the choroidal bath via agar bridges. The retinal agar bridge was filled with 2 M KCl. the choroidal agar bridge was filled with the control solution (solution 1, Table I). Circular Ag/AgCl electrodes in both the retinal and choroidal bath allowed passage of transepithelial currents. Currents passed in the retina to choroid direction were denoted as positive. The retinal side of the epithelium was superperfused and the perfusate could be changed within 10 set between different solutions. The perfusion system was designed to minimize loss of CO, (la Cour, 199 1). The choroidal compartment contained about 30 itI of the control solution (solution 1, Table 1). and was not perfused. In the experiments performed under nominally HCO:,- free conditions the choroidal compartment contained the Hepes-buffered HCO,,- free solution (solution 1 1, Table I ).

Table I shows the composition of the solutions used to superfuse the epithelium. Solutions l-10 were HCO,,--buffered, and gassed with 95% 02/5”/o CO,. Solutions 11, 12 and 13 were HCO,--free. Hepesbuffered, and gassed with 100% 0,. After being equilibrated with the gas mixture. solutions were titrated to pH 74. No compensation was made for chelation of Ca”+ by gluconate (Christoffersen and Skibsted, 1975). However, in eight control experiments no changes in membrane potential, or in intracellular Cll activity were observed when the retinal extracellular Ca’)+ concentration was reduced from 1.8 mM to 0.4 mM. Furosemide, ouabain, CAMP and 3-isobutyl-lmethylxanthine (IBMX) was obtained from Sigma (Sigma Chemie GmbH, Griinwalder weg 30. D-8024 Deisenhofen, Germany). Unless otherwise stated, results are presented as average +_S.E.M. (n, number of experiments). The Michaelis-Menten functions were fitted by non-linear regression using the Marquardt algorithm.

Stedy State Measurements In the control HCO,- buffer (solution 11Table I) the cells had a membrane potential across the retinal membrane of - 83.9 + 0.4 mV (n = 63). the intracellular Cll activity was 14.5 F 0.5 (II = 6 3 I and the voltage divider ratio was 0.54 f0.03 (n = 63). The transepithelial potential was 9.9 -t 0.3 mV (n = 26). and the transepithelial resistance was 134 & 3 Qcrn? (n = 26). In the control Hepes buffer (solution 11. Table I) the retinal membrane potential was - 84 + 1 mv (n = 7). the intracellular Cl. activity was 16 k 1 mM, the voltage divider ratio was 0.64 * 0.04. the transepithelial potential was loll mV (n = 3), and the transepithelial resistance was 15 5 -t 1 3 Qcmz.

When the perfusate was changed from the control solution (90.1 rnlv Cl-, solution 1, Table I) to one of the low Cl solutions (3.6 mM Cl-, solutions 2, 5 and 6. Table I) the intracellular Cl- activity decreased. When the perfusate was returned from the low Cll solution and back to the control solution, the intracellular Cl- activity increased rapidly towards control levels (Pigs 1. 2 and 3). It was investigated whether this increase in intracellular Cl- activity required the presence of extracellular Na- and K+. N~I” tiependence. In nine experiments

the change

I mln

LOW cl-

3. Results The cells were punctured with the microelectrode while the tissue was perfused with the control solution (solutions 1 or 11. Table I). During the first 2-6 min the cells were studied in steady state. Then three groups of experiments investigated: the K+ and Na+ dependence of Cll influx across the retinal membrane ; the effect of furosemide on the intracellular Clactivity: and the effect of transepithelial currents on the extracellular Cl activity.

-

0mt.1

Not

Fro 1. The changes in intracellular Cl- activity (a$-) and in retinal membrane potential (E,) in responseto changes in the retinal extracellular concentrations of Cl- and Na+. The Cli concentration was changed from 90.1 to 3.6 mM (low Cl ). The Na+ concentration was changed from 110 to 0 mM. In the first experiment (left). Cl.- and Na’ were simultaneously changed back from 3.6 to 90.1 mM and from 0 to 110 mM. In the second experiment (right), Cl- was changed back from 3.6 to 90.1 mM, 1.5 min before Na+ was changed back from 0 to 110 mM.

924

M. LA

-

m

I

OmM

COUR

K+

FIG. 2. The changes in intracellular Cl- activity (u,CIF) and in retinal membrane potential (E,) in response to changes in the retinal extracellular concentrations of Cl- and K’. The Cl- concentration was changed from 90.1 to 3.6 mM (low Cl-). The Kconcentration was changed from 2 to 0 mM. In the first experiment (left), Cl- was changed back from 3.6 to 90.1 mM. 1.5 min before K+ was changed back from 0 to 2 mM. In the second experiment (right), Cl- and K- were simultaneously changed back from 3.6 to 90.1 mM and from 0 to 2 mM.

-

100 pM

furosemhde

PIG. 3. The changes in intracellular Cl- activity (a&Y) and in retinal membrane potential (E,) in response to transient changes in the retinal extracellular Cl- concengration from 90.1 to 3.6 mM (low Cl-). In the second experiment (right), Cl- was changed back from 3.6 to 90.1 mM in the presence of 100 ,UM furosemide. from low to high Cl- was made while the epithelium was perfused with Na+-free solutions (solutions 3 and

5, Table I). In the absence of Na+ the change from low to high Cl- caused no increase in the extracellular Clactivity. Instead, this activity continued to decrease while the epithelium was perfused with Na+ free solutions (Fig. 1, right). When the perfusate was returned from the Na+-free solution to the control solution, the intracellular Cl- activity increased rapidly towards control levels (Fig. 1, right). Similar results were obtained in five experiments performed under nominally HCO,--free conditions (solutions 11, 12. and 13, Table I).

K’ dependence. In seven experiments the change from low to high Cl- was made while the epithelium was perfused with K+-free solutions (solutions 4 and 6, Table I). In the absence of K+, the intracellular Clactivity remained low after change from low to high Cl- (Fig. 2. left). Only after the perfusate was returned. from the K+-free solution to the control solution, the intracellular Cl- activity increased towards control levels (Fig. 2. left). Similar results were observed l-12 min after the Na+/K+ pump had been inhibited by 100 ,UM ouabain (n = 4), and l-l 8 min after 50 [tM Bad+ had been administered into the perfusate (n = 4). Baz+ (50 ,UM) completely inhibited the hyperpolar-

Cl-

TRANSPORT

IN

FROG

925

RPE

5

1c

20

100

50

1000

@M furosemidc

4. The changes in intracellular Cl- activity (n,Cl ) and in retinal membrane potential (E,.)in responseto administration of 5, IO. 20. 50. 100 and lOOO/(M furosemide. FIG.

70 _ T

i 50

60.

4 1

Furosemide

concentrahon

(,u~)

FIG. 5. The initial rates of decrease in concentration Cl- activity (-da$m .dt-l) observed after retinal administration of furosemide plotted against the furosemide concentration. Values are mean k S.E.M.

ization of the retinal membrane potential caused by K’ removal. Effects offurosemide. In five experiments, the change from low to high Cl- was performed in the presence of 100 11~ furosemide (solution 2 and solution 1 + 100 /.M furosemide, Table I). In these experiments the intracellular Cl- activity remained low after the change from low to high Cl-. Only after furosemide was removed. the intracellular Cl- activity returned toward control levels (Fig. 3. right). In the absence of furosemide the intracellular Cl- activity returned towards control levels immediately after the change from low to high Cl- (Fig. 3. left). The change from low to high Cl- caused depolarizations of the retinal membrane potential (Fig. 3). In the absence of furosemide these depolarizations averaged 7.5 10.7 mV (n = 26); in the presence of 100 PM furosemide they averaged 7.8 f 1,2 mV (n = 5). The difference was not significant (unpaired t-test, P > 0.1). When administered to cells in steady state, during perfusion with the control solution, 100 //M furo,111

semide caused a decrease in the intracellular Clactivity, the initial rate of which was 61 F 2 /IM see-’ (Fig. 4). A range of furosemide concentrations between 5 PM and 1 mM was tested. Figure 5 shows the initial rates of decrease in the intracelluiar Clactivity ( - daiClk. dt-‘, jrmol . 1-l . see-’ ) plotted against the concentrations of furosemide. Between 5 and 100 pi furosemide the initial rates of decrease in the intracellular Cl- activity increased when the furosemide concentration was increased. The effect saturated between 50 and 100 ,u~ (Figs 4 and 5). The initial rates of decrease in intracellular Cl- activity could be fitted to a single Michaelis-Menten function of the furosemide concentration (Fig. 5). The affinity constant for furosemide was Ki = 7f 2 1tM (estimate f SD.). Administration of furosemide in concentrations above 20 /IM caused small depolarizations of the retinal membrane potential. Administration of 100 ,u~ furosemide caused a depolarization of the retinal membrane potential of 1.2 F 0.2 mV. a depolarization of the transepithelial potential of 0.6 i: 0.1 mV, and an increase in the voltage divider ratio of 0.1 1 F0.02 (n = 50). KEKii

926

M LA COUR

Extracellular

Kt concentration

(mM)

FIG. 6. The initial rates of decrease in the intracellular Cl- activity ( -dn,Clk .dt-I) observed after retinal administration 100 pM furosemide plotted against the retinal extracellular K‘ concentration. Values are mean+s.~.~.

2

-80

of

-

4; -90 L-

I mM CAMP

+ O-5

mM

IBMX

FIG. 7. The changes in intracellular Cl- activity (a$-) and in retinal membrane potential (E,) in response to retinal administration of 100,~~~ furosemide before, during and after the epithelium was perfused with solutions containing 1 rn,vi cyclic AMP and 0.5 mM IBMX (1 mM CAMP + 0.5 InM IBMX).

Dependenceof Nu’ and K+ If furosemide affected the cells only by inhibition of inward Na+,K+,CIl cotransport, the drug should not decrease the intracellular Cll activity in the absence of extracellular Nat and K+. In seven experiments no changes in the intracellular CI- activity were observed when 100 ,U~M furosemide was administered while the epithelium was perfused with the Na+-free solution (solution 3, Table I). While the epithelium was perfused with the K+-free solution (solution 4, Table I) 100 ,LLM furosemide caused a slight increase in the intracellular Cll activity (Fig. 6). Furosemide (100 yM) decreased the intracellular Cll activity at an increasingly faster rate

when

administered

while

the epithelium

was

perfused with solutions containing increasing K’ concentrations (0.5, 1. 2, 5 and 10 mM K’, Fig. 6). The effect saturated between 1 and 5 mM K-. Michaelis-Menten analysis of the initial rates (-da,Cl-.dt-‘) with respect to the extracellular K’ concentration yielded an apparent affinity constant (K,,,) for K.+ of (> 5 + 02 mM (estimate + S.D.).

Effect ofcAMP on thefurosemide response.Administration of 1 mM CAMP and O-5 mM IBMX to the perfusate caused a depolarization of the membrane potential across the retinal membrane of 2.7 + 0.2 mV (II = L 3 ), a hyperpolarization of the transepithelial potential of l.Of0.2 mV, an increase in the voltage divider ratio of 0.3 8 10.04 (measured 1 min after administration of CAMP), and a decrease in the transepithelial resistance of 7 + 1 Qcm’ (measured 1 min after CAMP administration). In three out of 13 experiments, the intracellular Cl- activity did not change after administration of CAMP and IBMX. In five experiments a slight increase in this activity was observed (one of these experiments is shown in Fig. 7) and in four experiments a small decrease was observed. On average the intracellular Cl- activity was 0.4 rnM* 0.4 mM higher 1 min after administration of CAMP and IBMX. than before administration of the drugs (n = 13). The effect of furosemide on the intracellular Clactivity was investigated in CAMP stimulated tissues.

CI~ TRANSPORT

IN FROG

927

RPE

-m -10

-20

m

m

-mm

-30

-10

IO

30 pA

20

I mln FIG. 8. The changes in intracellular Cl- activity (a,CI-) in responseto the passageof transepithelial currents of - 30. - 20. - 10. 10, 20 and 30 //A. Positive currents were passedin the retina-to-choroid direction

150 100 50 0 -50 -100 -150 -200 -250 -300

-60

-40

-20

0

20

Afc,,,,hV, FIG. 9. The effectsof transepithelial currents of - 30. - 20, - 10, 10.20 and 30 /tA. The initial rates of change in intracellular Cl- activity (da,Clk.dt-‘) plotted against the change in the membrane potential across the choroidal membrane (AE,,,,,). Also shown are lines obtained by linear regression through the origin of da,Cl- .dtt’ on AE,,,,.. Open symbols and the straight line represent the effectsof transepithelial currents under control conditions. Closedsymbols and the dotted line represent the effects of transepithelial currents in tissues stimulated with 1 mM CAMP and 0.5 mM IBMX.

In nine experiments 100 ,uM furosemide was administered 2-9 min after administration of CAMP and IBMX (Fig. 7). The initial rate of decrease in the intracellular Cl- activity observed in these experiments was 45 + 9 irmol.1 ‘.sec-I. This is 26% smaller than what was observed when 100 /LM furosemide was added without simultaneous stimulation with CAMP and IBMX; the difference was significant (unpaired t-test, P < 0.02 5). The Effects of Transepithelial Currents To investigate the putative Cl- conductance in the choroidal membrane of the frog RPE, the effects of transepithelial currents on the intracellular Cl- activity were studied. Currents of - 30, - 20, - 10, 10.20 and 30 ,uA were passed. Positive currents (retina-tochoroid) hyperpolarized the retinal membrane potential, depolarized the choroidal membrane potential, and caused an increase in intracellular Cl- activity. Negative currents (choroid-to-retina) had the opposite effects on membrane potentials and intracellular Cl-

activity (Fig. 8). Figure 9 shows a plot of the current induced changes in choroidal membrane potential (AF,.,,,,.) versus the initial rates of change in the intracellular Cll activity ida$?~dt~‘). Linear regression through the origin yielded a slope (da+? dtt’ AE,.,,,,,.-’) of 4.1 t 02 jtmol . 1-l . see-’ mV-’ (estimate &LO.). It is not likely that the current induced changes in intracellular Cll activity were caused by altered Na*,K+,Cl- co-transport through the retinal membrane. In three experiments, 100 /LM furosemide was present in the retinal bath while - 10, -20 and - 30 ld”A were passed: current induced decreases in the intracellular Cl- activity were still observed and, da,Cl- AE,.,,,,.?. was 4.5 * 0.8 /[mol. ll’ . see-’ rnv’ (estimate f S.D.). In two control experiments the retinal side of the epithelium was perfused with the low Cll solution (solution 2, Table I) while 30 /rA were passed. This resulted in depolarizations of the choroidal membrane potential of -25 and - 26 mV and increases in the intracellular Cl- activity, the initial rates of which were 200 and 241 lcmol.I-’ .sec-I.

928

The effect of transepithelial currents on the intracellular Cl- activity was also investigated l-l 2 min after 1 mlM CAMP and 0.5 mM IBMX had been administered into the retinal perfusate. Figure 9 displays the results of 39 such experiments (filled symbols). Linear regression through the origin, of the initial rates of change in the intracellular Cl- activity on the changes in the choroidal membrane potential, yielded the dotted line shown in Fig. 9. The slope, daiC1- dt-’ A&,,,,. -‘, was 5.0 + 0.3 HIMsee-’ mV-‘. which is significantly larger than the slope obtained in the absence of CAMP and IBMX (F-test. P < 045). For each experiment da,Cl- dt-’ AE,.,,Ol.-l was calculated and plotted against the time from CAMP administration : linear regression showed no significant development in da,Cl- dt-’ A&,,,,.-’ over time after CAMP administration (F-test, P > 0.1). Similarly, no significant development over time from CAMP administration was found for the voltage divider ratio (VDK) (F-test, P > 0.05).

4. Discussion Na+,K+,Cl- Co-transport Intracellular Cl- was accumulated above electrochemical equilibrium : with an average intracellular Cl- activity of 14.5 mM, the Cl- equilibrium potential across the retinal membrane was 3 7 mV more positive than the membrane potential across this membrane. Thus, the Cl- influx, observed when the perfusate was changed from Na+-free or K’-free solutions to the control solution, proceeded against an electrochemical gradient (Figs 1 and 2). The K+ dependence of this Clinflux was not secondary to effects on Na+/K+ pumping, since similar results occurred when this pump was blocked with ouabain. Neither was the K’ dependence secondary to K+-induced changes in membrane potential since similar results were observed after K+ channels were blocked with BazC. The Na+ dependence of Cl- influx was not secondary to altered Na+:HCO,- co-transport (Hughes et al., 1989: la Cour, 19 8 9 ), since similar results were observed under nominally HCO,--free conditions. It is concluded that Cl- influx across the retinal membrane of the frog RPE was dependent on the simultaneous presence of extracellular Na+ and K’. Thus the present work supports the hypothesis of Na+.K+,Cl- co-transport across this membrane. It remains, however, to be shown that Na+ and K+ can be driven against their electrochemical gradients by an imposed Cl- gradient and that the fluxes of Cl-, Na’, and K’ are coupled in a constant stoichiometric relation (Geck and Heinz, 1986). Furosemide is an inhibitor of Na+,K+,Cl- co-transport (O’Grady et al., 1987). Furosemide (100 ,f/M) inhibited Cl- entry into the RPE cells after a change in the extracellular Cl- concentration from 3.6 to 90.1 mM (Fig. 3). When administered to cells in steady

M LA COUR

state. furosemide in c.oncentrations between 5 and 1000 mM decreased the intracellular Cl- activity (Figs -L and 5). Thus cells in steady state exhibited furosemide-sensitive Cl~- influx across the retinal membrane. This confirms a previous study (Wiederholt and Zadunaisky. 1985). The effect of furosemide on intracellular Cl- activity saturated between 50 and 100 ,L~M(Fig. 5). Michaelis-Menten analysis showed half maximal effect when the furosemide concentration was 7 l(M (Fig. 5). This agrees well with previous studies in the flounder intestine where half maximal inhibition of Na’ .Kf.Cl- co-transport was obtained at furosemide concentrations between 5 and lO/r~ (O’Grady et al., 1987). Effects of 100 /TV furosemide on Na” and K’ independent Cl- transport systems are unlikely, since furosemide did not decrease the intracellular Cl- activity in the absence of these ions. It is assumed that the effect of 100 ,L~Mfurosemide on the intracellular Cl- activity was caused solely by inhibition of Na+,K+,Cl- co-transport across the retinal membrane. Thus the initial rate of decrease in the intracellular Cl- activity (da,Cl-.dt-I). observed after administration of 100 i/M furosemide, is interpreted as being proportional to the rate of Na+,K-.Cl cotransport (I,., ) just prior to furosemide administration. The proportionality factor is the estimated intracellular volume to surface ratio of the epithelium (h = 0~00 1 5 cm) : J,., = h da$- dt-’ . The relevant intracellular volume is the volume of the cytoplasm. The cytoplasm is the intracellular distribution space that is immediately available for Cl- ions after they enter the cells through the plasma membrane, and it is the space where intracellular microelectrodes measure the Cl- activity. The volume of the cytoplasm generally constitutes 50-60% of the total cell volume (Alberts et al., 198 3 ). The total cell volume is the volume of the cell soma plus the volume of the 30-60 /lrn long apical processes of the RPE. The geometry of these processes has previously been assessed (Miller and Steinberg. 19 77b). There are about 10’ processes per cm” ; each can be approximated by a truncated cone with a volume of about 1.2 x 10~“’ cm”. The volume of the processes is therefore approximately 0.0012 cm” cm-“. The volume of the cell somata was calculated from the estimated height (15 /rm) of the epithelium as 0.001 5 cm” cm-“. Thus the total RPE cell volume was estimated as 0002 7 cm’ cm-:‘, and the intracellular (cytoplasmatic) volume to surface ratio as 0.5 5/0.002 7 = 0.001 5 cm” cm-‘]. For cells in steady state during perfusion with the control solution I,,, = 0.33 /tmol cm-” hr’. It should be noted that I,., is expressed relative to the exposed epithelial surface area. and not relative to the true membrane area. In the isolated frog KPE-choroid preparation net transepithelial Cl- absorptions between 0 19 and 0.74 Ltmol cm-’ hr’ have been measured by the Ussing-chamber technique (Miller and Steinberg, 1977a: DiMattio et al., 1983: Miller and Farber.

Cl

TRANSPORT

IN FROG

RPE

1984 ; Hughes et al., 1987). The rate of furosemidesensitive Na+,K+,Cll influx of 0.33 pmol cm-” hr-‘, estimated in the present study, is within this range of experimentally determined transepithelial Cl- absorptions. It is possible that 100 ,a~ furosemide did not completely inhibit Na+.K+,Cl- co-transport, and it cannot be excluded that other Cl- influx mechanisms exist in the retinal membrane, alongside the furosemide-sensitive Na+,K+.Cl- co-transport system. Electronrutrality

Retmo 3No+

Furosemfde

N *

2K+

K+

Na+

Cl-

of‘ Na+.K+.CI~ Co-transport

Na+,K+,Cl- co-transport is generally believed to be electroneutral (O’Grady et al., 198 7). Also in the frog RPE this co-transport seems to be electroneutral since furosemide administration inhibited Cl- influx across the retinal membrane, but produced only small changes in retinal membrane potential (Fig. 4). The depolarizations of the retinal membrane potential caused by the changes from low to high Cll (Fig. 3) were observed both in the absence and presence of 100 PM furosemide. It is thus not likely that they were caused by changes in Na+,K+,Cl- co-transport. The depolarizations were probably caused, at least in part, by liquid junction potentials, which would tend to make the retinal reference electrode more positive in low Cll solutions (Barry and Diamond. 1970). K+ Dependenceof‘ Na+.K*,CI- Co-transport The initial rates of decrease in the intracellular Cll activity, observed after administration of 100 ,uM furosemide, increased when the extracellular K+ concentration was increased between 0 and 10 mM. Michaelis-Menten analysis yielded half maximal effect at 0.5 mM extracellular K+ (Fig. 6). This is in good agreement with what is found for Na+,K+,Cll cotransport in other tissues (O’Grady et al., 198 7). The rate of furosemide-sensitive Na+,K+,Cl- co-transport was almost constant in the physiological range of K’ activities in the subretinal space (1-5 mM). Therefore it is unlikely that changes in Na+,K+,Cl- co-transport caused the changes in K+ and Cll transport that have been reported after physiological changes in the retinal extracellular K’ concentration (Miller and Steinberg. 1982; la Cour et al., 1986: Edelman, Miller and Hughes. 1988). Effects of cAMP on Na+.K+,CI- Co-transport Stimulation of the frog RPE with 1 mM CAMP and 0.5 mM IBMX has been shown to cause a 50% decrease in the transepithelial transport of Cl- in the retina-to-choroid direction (Miller and Farber, 1984 ; Hughes et al., 198 7). In the present study, CAMP and IBMX was found to cause a 26% decrease in rate of furosemide-sensitive CI- influx across the retinal membrane. Thus. a decrease in the rate of Na+,K+.Cll co-transport may contribute to the CAMP response in the frog RPE.

I--I T

Cl-

Choroid

FIG. 10. Model of Cl- transport in the frog retinal pigment epithelium. The retinal membrane incorporates a Na+,K+.CIco-transport system, which is furosemide-inhibitable. and through which Cl- enters the RPE cell. The Na+,K-.CI- cotransport system is energized by the Na+ gradient created by the Na’/K+-pump in the retinal membrane. The choroidal membrane incorporates a Cl- conductance through which Cl- exits the RPE cell.

Conductive Cl- Transport Across the Choroidnl Membrane A conductive exit pathway for Cl- in the frog RPE choroidal membrane (Fig. 10) has been suggested for some time (DiMattio et al.. 1983; Frambach and Misfeldt, 1983; Gallemore and Steinberg, 1989; Hughes and Steinberg, 1990: Miller and Edelman, 1990). In the present study, transepithelial currents that depolarized the choroidal membrane potential increased the intracellular Cl- activity: currents that hyperpolarized this membrane potential decreased the intracellular Cl- activity (Figs 7 and 8). These findings can be explained by conductive Cll transport through the choroidal membrane. The results are opposite to those expected for conductive Cll transport through the retinal membrane. While the retinal side of epithelium was perfused with low Cl solutions, positive currents still caused Cl- influx into the RPE cells. It is therefore unlikely that Cll transport through the retinal membrane was responsible for the currentinduced changes in the intracellular Cll activity. It is, in particular, unlikely that the Na+,K+,Cll co-transport system contributed to these changes because they were also seen while this co-transport was inhibited by 100 ,UM furosemide. It is assumed that the current-induced changes in the intracellular Cll activity were caused only by conductive transport through the choroidal membrane. The apparent conductance of the choroidal membrane, g, ,-, can then be calculated: g(.,- = aliF = 059 mS cm-?, where c1is the slope of the regression of da,Cll .dt-’ on AF,.,,,,.(Fig. 9), 11= 0001 5 cm is the

930

volume to surface ratio of the epithelium, and E’ is Faraday’s constant. The total conductance of the choroidal membrane has been estimated to be 2.2 mS cm -‘! (Miller and Steinberg, 1977b). Thus, the estimated apparent Cll conductance represents 270/, of the total membrane conductance in the choroidal membrane. For cells in steady state during perfusion with the control solution, the equilibrium potential for Clacross the choroidal membrane (ET!;-) was -47 mV. and the membrane potential across the choroidal membrane (F,.,,,,.)was - 74 mV. The conductive efflux of Cl- through the choroidal membrane can then be calculated : gc.,-(I?,.,,,,.- E;:;-) F-’ = 060 /Lrnol cm-” hr. ‘. This flux is well within the range of Cll absorptions of 01990.74 irmol cm-” hr-’ that has been measured across the isolated frog RPE-choroid preparation by the Ussing chamber technique (Miller and Steinberg, 1977a; DiMattio et al., 1983 ; Miller and Farber, 1984: Hughes et al., 1987). The estimated apparent Cl- conductance in the choroidal membrane is therefore large enough to account for the necessary Cll efflux through this membrane at least for cells in steady state under control conditions. In earlier electrophysiological studies 1 mM CAMP and 0 5 mM IBMX was found to cause a hyperpolarization of the transepithelial potential, a depolarization of the retinal membrane potential, an increase in the voltage divider ratio, and a decrease in the transepithelial resistance (Hughes, Miller and Farber, 19 8 7 : Hughes et al., 1988). In the present study similar electrophysiological observations were made. The changes in electrical parameters are compatible with a conductance increase in the choroidal membrane. and it has been suggested that CAMP induce an increase in the anion (putatively Cl-) conductance of this membrane (Hughes et al., 1987). In the present study administration of CAMP and IBMX increased the apparent Cl- conductance by 2 1 YO(Fig. 8). Since the apparent Cl- conductance only accounts for 27% of the total choroidal membrane conductance, this is not sufficient to explain the observed 72 “/” increase in the voltage divider ratio. Thus. the CAMP induced changes in the electrical parameters, can only be explained in part by the change in the apparent Cl- conductance in the choroidal membrane. CAMP stimulation decreased the rate of Cl- influx. via the Na+.K’,Cll co-transport system. It also increased the apparent Cl-- conductance in the choroidal membrane, and thus the rate of conductive Cl- efflux through this membrane. A decrease in the intracellular Cl- activity should therefore be expected after CAMP stimulation. Surprisingly only small and inconstant changes in the intracellular Cl- activity were observed (Fig. 7). Transepithelial flux measurements in Ussing chambers showed that CAMP stimulation decreased the rate of transepithelial Cll absorption by 50% (Miller and Farber, 1984; Hughes et al.. 1987). The net Cl- efflux through the choroidal membrane

M. LA COUR

should therefore be reduced after CAMP stimulation. Yet an increase in the conductive Cl. efflux through the choroidal membrane was estimated in the present study. It is suggested that besides affecting Na-.K’.C’l co-transport across the retinal membrane, and conductive Cll transport across the choroidal membrane. CAMP stimulation also affects some other Cl transport system in the frog RPE. Note Added in Proof After this paper was submitted. a paper from another group appeared (Joseph and Miller. 1991) in which electrophysiological evidence was presented for the existence of a large Cl conductance in the choroidal membrane of the isolated bovine RPE. Acknowledgements I am grateful to S. S. Miller. B. A. Hughes. B. Cherksey and T. Zeuthen for useful discussions: J. 1J.Prause for reading the manuscript; Strunk Rellim for useful comments on the proper use of the English language ; and T. Soland for technical assistance. This research was supported by the Danish Medical Research Council and Landsforeningen Bekzmpelse af Ojensygdomme og Blindhed.

til

References Adorante. J. S. and Miller, S. S. (1990). Potassium-dependent volume regulation in retinal pigment epithelium is mediated by Na.K.CI cotransport. 1. Gen. Physiol. 96. 1153-76. Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson. J. D. (1983). Molecular Biofogy of the Cell. Pp. l-l 146. Garland Publishing Inc.: New York. Barry, P. H. and Diamond, J. M. (1970). Junction potentials, electrode standard potentials, and other problems in interpreting electrical properties of membranes. /. Mrmhr. Biol. 3. 93-122. Christoffersen, G. R. J. and Skibsted. I,. H. (1975). Calcium ion activity in physiological salt solutions : influence of anions substituted for chloride. Camp. Bioclirni. Physiol. 52A. 3 17-22. DiMattio, J.. Degnan. K. G. and Zadunaisky. J. A. (1983). A model for transepithelial ion transport across the isolated retinal pigment epithelium of the frog. Exp. Eye Res. 37. 409-20. Edelman. J. L., Miller. S. S. and Hughes, B. A. ( 1988). Kegulation of chloride transport by frog retinal pigment epithelium (RPE). In Proc:erdirtgs o/’ tlie [ntcrnatianal Sockty for Eye Research, ~1. V. Abstracts. Eighth International Congress of‘ Eye Rewawh. San Francisco. CA. U.S.A. P. Sla. The International Society for Eye Research : Menlo Park, U.S.A. Frambach. D. A. and Misfeldt. D. S. (1983). Furosemidesensitive Cl- transport in embryonic chicken retinal pigment epithelium. A~J. 1. Ph;Jsiol. 244, F679-F685. Frambach, D. A., Valentine, J. L. and Weiter, J. J. (1989). Furosemide-sensitive Cl transport in bovine retinal pigment epithelium. h~vest. Ophthalmol. Vis. Sci. 30. 22714. Gallemore, R. P. and Steinberg, R. H. (1989). Effects of DIDS on the chick retinal pigment epithelium. I. Membrane potentials, apparent resistances, and mechanisms. 1. .Va.u-ash. 9, 1968-76.

Cl

TRANSPORT

IN FROG

RPE

Geck. P. and Heinz. E. (1986). The Na-K-2CI cotransport system. J. Mrntbr. Biol. 91, 97-105. Hughes. B. A.. Adorante, J. S.. Miller, S. S. and Lin, H. (1989). Apical electrogenic NaHCO, cotransport. A mechanism for HCO, absorption across the retinal pigment epithelium. I. Gen. Physiol. 94, 125-50. Hughes, B. A.. Miller. S. S. and Farber, D. B. (1987). Adenylate cyclase stimulation alters transport in frog Am. I. Physiol. 252, retinal pigment epithelium. C385-('395.

Hughes, B. A., Miller, S. S.. Joseph, D. P. and Edelman, J. L. ( 1988). CAMP stimulates Na+-K’ pump in frog retinal pigment epithelium. Atn. 1. Physiol. 254. c84-c98. Hughes. B. A. and Steinberg. R. H. (1990). Voltage-dependent currents in isolated cells of the frog retinal pigment epithelium. 1. Physiol. 428. 27 3-97. ]oseph. 1). t’. and Miller, S. S. (1991). Apical and basal membrane ion transport mechanisms in bovine retinal pigment epithelium. /. Physiol. 435, 439-63. Kennedy. B. G. ( 1990). Na+-K--Cl cotransport in cultured cells derived from human retinal pigment epithelium. Att~. 1. Ph~jsiol. 259. C29-C33. la Cour. M. (19891. Rheogenic sodium-bicarbonate cotransport across the retinal membrane of the frog retinal pigment epithelium. J. Physiol. 419, 539-53. la Cour, M. ( 199 1). Kinetic properties and Na+ dependence of rheogenic Na--HCO, co-transport in frog retinal pigment epithelium. 1. Physiol. 439. 59-72. la Cour. M.. Lund-Anderson, H. and Zeuthen. T. (1986). Potassium transport of the frog retinal pigment epithelium: autoregulation of potassium activity in the subretinal space. /. Physiol. 375, 461-79.

931

Miller. S. S. and Edelman. J. L. (1990). Active ion transport pathways in the bovine retinal pigment epithelium. 1. Physiol.

424,

283-300.

Miller. S. S. and Farber, D. (1984). Cyclic AMP modulation of ion transport across frog retinal pigment epithelium. Measurements in the short circuit state. I. Gcn. Physiol. 83.853-74.

Miller. S. S. and Steinberg, R. H. (1977a). Active transport of ions across frog retinal pigment epithelium. Esp. E,qr~ Res. 25. 235-48. Miller. S. S. and Steinberg, R. H. (19 77b). Passive ionic properties of frog retinal pigment epithelium. /. Men&. Bid. 36. 337-72. Miller. S. S. and Steinberg, R. H. (1982). Potassium transport across the frog retinal pigment epithelium. 1. Memhr. Bid. 67. 199-209.

Musch. M. W. and Field, M. (1989). K-independent Na-Cl cotransport in bovine tracheal epithrlial cells. Ant. I. Phgsiol. 256, C6.584665.

O’Grady. S. M.. Palfrey, H. C. and Field. M. (198i). Characteristics and functions of Na-K-Cl cotransport in epithelial tissues. Ant. 1. PhNsiol. 253. Cl i/-C192. Tsuboi. S.. Manabe, R. and Iizuka, S. (1986). Aspects of electrolyte transport across isolated dog retinal pigment epithelium. Am. 1. Physiol. 250, F78 I-F784. Wiederholt. M. and Zadunaisky, 1. A. (1985). Decrease of intracellular chloride activity by furosemide in frog retinal pigment epithelium. Cur. Eye RPS. 3. 673-5. Zeuthen, T. ( 19801. How to make and use double-barreled ion-selective microelectrodes. In Currmt Topics in Menbrcms md Trmsport. Vol. 1 3. (Ed. Boulpaep. E. I,.). Pp. 31&f7. Academic Press: London.