Vesicle targeting to the apical domain regulates bile excretory function in isolated rat hepatocyte couplets

GASTROENTEROLOGY1995;109:1600-1611

Vesicle Targeting to the Apical Domain Regulates Bile Excretory Function in Isolated Rat Hepatocyte Couplets JAMES L. BOYER and CAROL J. SOROKA Department of Medicine and Liver Center, Yale UniversitySchool of Medicine, New Haven, Connecticut

See editorial on page 1706. Background & Aims: Plasma membrane solute transport may be regulated in many epithelial cells by vesicle traffic to and from the site of residence of the transporter. The aim of this study was to determine if this phenomenon may also play a role in the regulation of canalicular transport of bile acids. Methods: Confocal microscopy and image analysis were performed to quantitatively assess changes in secretory capacity and vesicle targeting in isolated rat hepatocyte couplets that had been exposed to fluorescent bile acid after pretreatment with dibutyryl adenosine 3',5'-cyclic monophosphate (DBcAMP) and/or nocodazole. Resuits: DBcAMP stimulated bile acid secretion by 240% while significantly increasing canalicular circumference. Nocodazole decreased secretion by 410% and significantly decreased canalicular circumference. When DBcAMP was added to nocodazole-treated couplets, a slight but significant increase was found in both fluorescent bile acid secretion and canalicular circumference as compared with nocodazole alone. Finally, DBcAMP stimulated translocation of vesicles to the canalicular membrane as determined by immunocytochemical localization of a putative bile acid transporter, Ca 2÷, Mg2÷-ecto-adenosine triphosphatase. Conclusions: The findings support the view that apical membrane transport activity in the rat hepatocyte is highly regulated by the insertion of vesicles into this domain and that this process involves both microtubule-dependent and -independent mechanisms.

he hepatocyte is a highly polarized cell whose apical canalicular membrane functions as the excretory domain for the secretion of bile. Tight junctions form barriers between the canalicular membrane and the basolateral domain and seal the lumen of the bile canaliculus, into which the secretion of bile is elaborated.* A number of membrane proteins are localized to this apical pole of the cell, including several adenosine triphosphate-dependent transporters, the multiorganic anion transporter, the multidrug resistance gene product (P-170), a bile

T

acid transporter, and a CaZ+,Mg2+-ecto-adenosine triphosphatase (ATPase). 2 Ion transport proteins, including a CI-/HCO3- exchanger and an HCO3-/SO 4- exchanger, are also restricted to this domain. 3'4 Little is known about the regulation of these canalicular transporters, although proteins that reside there are believed to be targeted to this functional location by microtubule-dependent vesicle transport (transcytosis) after initially being directed to the basolateral plasma membrane after their synthesisJ -7 To date, all apical (canalicular) membrane proteins in hepatocytes (whose targeting has been assessed) use this indirect pathway. In a previous study, we have been able to show that the activity of one of these apical membrane proteins, the CI-/HCO3- exchanger, can be stimulated after intracellular alkalinization or by administration of an adenosine 3',5'-cyclic monophosphate (cAMP) analogue, dibutyryl adenosine 3',5'-cyclic monophosphate (DBcAMP), and that this enhanced activity is dependent on intact microtubule function.8 Under similar conditions, the canalicular membrane protein Ca2+,Mgg+-ecto-ATPase, a putative bile acid transport protein, was also increased at the apical membrane as determined by immunofluorescence studies. These findings led to the general hypothesis that bile secretory function might be regulated by movement of vesicles that contain these transport proteins into or out of the canalicular domain. 8'9 To test this hypothesis further, we have used a fluorescent bile acid to assess the excretory capacity of isolated rat hepatocyte couplets, a polarized bile secretory unit in cell culture, while altering the targeting of vesicles to the apical canalicular domain with microtubule inhibitors and DBcAMP. Changes in apical targeting were assessed by measurements of the canalicular membrane circumference using imaging techniques. The findings provide evidence that canalicular bile acid excretion is Abbreviationsused in this paper: CGamF,cholylglycylamidofluorescein; DBcAMP,dibutyryladenosine3',5'-cyclic monophosphate; IBMX, isobutylmethylzanthine. © 1995 by the AmericanGastroenterologicalAssociation 0016-5085/95/$3.00

November 1995

VESICLE TARGETING AND BILE ACID EXCRETION 1603.

r e g u l a t e d b y the insertion of vesicles c o n t a i n i n g bile acid

Functional (Bile Acid) Secretion

transporters into the apical d o m a i n .

M a t e r i a l s and M e t h o d s Collagenase (type D) was purchased from Boehringer Mannheim Biochemicals (Indianapolis, IN). Leibovitz's (L-15) medium, penicillin and streptomycin, and fetal calf serum were from Gibco (Grand Island, NY). DBcAMP, isobutylmethylzanthine (IBMX), and nocodazole were from Sigma Chemical Co. (St Louis, MO). All other chemicals were the highest purity available commercially. Hepatocyte couplets (viability, 8 5 % - 9 2 % ) were isolated from rat liver as previously described from this laboratory1° and cultured in Liebovitz's (L-15) medium containing 10% fetal bovine calf serum and penicillin and streptomycin. Cells were plated on glass coverslips at concentrations of 1.1 × i05 celts/cm 2 at 37°C in an air atmosphere.

Treatments Control cells were maintained in L-15 media alone for 4 hours. Inhibition of microtubule-mediated vesicular movement was accomplished by the addition of nocodazole ( 1 0 - 2 0 btmol/L) at the time of plating. Cells were treated for 4 hours except in the case of reversal studies in which the ceils were removed from the nocodazole after 2 hours and placed in L15 media. DBcAMP (100 btmol/L) and IBMX (500 btmol/L) were added to dishes after 2 hours of culture, and cells were incubated in their presence for an additional 2 hours.

Immunofluorescence The cellular distribution of the putative bile acid transporter Ca2+,Mgi÷-ecto-ATPase was determined by immunocytochemical techniques using a polyclonal antibody to this canalicular membrane protein. After 4 hours in culture, the cells were fixed with cold methanol for 10 minutes. Primary antibody was diluted in phosphate-buffered saline containing 0.25% Triton X-100 and incubated on the ceils for 2 hours at room temperature. After washing, secondary antibody conjugated to Texas Red was incubated for an additional hour and fluorescent localization detected using confocal microscopy. In addition, the effects of nocodazole on microtubule structure were assessed by an immunofluorescence assay using a polyclonal antibody to ~-tubulin (Amersham). In this case, better visualization of intact microtubules was accomplished by using frozen sections. The cultured ceils were fixed with 4% paraformaldehyde, scraped from the culture dish, and embedded in 10% gelatin before freezing with liquid nitrogen. Semithin ( 0 . 5 - 1 . 0 [.tm) sections were cut with a cryomicrotome (Reichert Ultracut E) and placed on poly-L-lysinecoated glass microscope slides. Immunocytochemistry was performed with the monoclonal antibody to ~-tubulin and a secondary anti-mouse antibody conjugated to fluorescein isothiocyanate.

The functional secretory capacity of the hepatocyte couplets was assessed by measuring the hepatic uptake and secretion of 1 btmol/L cholylglycylamido fluorescein (CGamF) into the canalicular space. CGamF was synthesized according to Schteinart et al. tl and kindly provided by Dr. Alan Hofmann (San Diego, CA). Control and treated cells were incubated in the presence of 1 ~tmol/L CGamF in HEPES buffer (in mmol/ L: KC1, 4.7; MgSO4, 1; CaCl2, 1.25; KI-I2PO4, 1.2; HEPES, 10; NaC1, 135; and glucose, 5) for 5 minutes at 37°C. Coverslips containing the ceils were then removed from the fluorescent bile acid, and the cells were washed in buffer and allowed to secrete for an additional 10 minutes before analysis. The percentage of couplets secreting the fluorescent-labeled bile acids into the canalicular lumen was quantitated initially by counting the number of couplets with expanded canaliculi that contained fluorescence using a Nikon Microphot epifluorescence microscope (Melville, NY). For the majority of the experiments, ceils were scanned using a Zeiss Axiovert microscope (Thornwood, NY) attached to a Biorad MRC-600 confocal microscope (Richmond, CA) with a krypton/argon laser. Confocal machine settings (gain, aperture, black level, and neutral density filters) were set to maximize the dynamic range between background and the more intense canalicular fluorescence and were identical for all experiments. W i t h these settings, the depth of focus was 0.8 ~tm and saturation of fluorescent intensity measurements was minimized and occurred only for canalicular fluorescence in experiments in which secretion was stimulated with DBcAMP. Ceils were excited with a 488-nm line of krypton laser, and emission wave lengths > 5 1 3 were collected and Kalman-filtered during collection for noise reduction. Couplets with expanded canaliculi were selected randomly as in previous studies 8'12 without knowledge of treatment using bright-field optics before fluorescent imaging. Images were saved to an optical disc, and quantitation of total cellular and canalicular fluorescence intensity was subsequently performed using COMOS software (Biorad). Measurements of the canaticular circumferences were also obtained in the same couplets that were used for quantitation of canalicular fluorescence by tracing the fluorescent canalicular lumens. Previously we have used optical planimetry and Nomarski optics to measure the change in canalicular surface membrane under various conditions.13 Similar results were obtained when the measurements were made on the confocal fluorescent images. Therefore, subsequent calculations were performed using the confocal fluorescent images, which permitted the dual assessment of the functional capacity of these isolated bite secretory units. The percent of CGamF excreted into canaliculi could then be plotted as a function of their respective canalicular circumference. In all of these experiments, couplets were excluded from analysis if the lumens were not expanded because a portion of isolated couplets either fail to seal their canalicular spaces and thus do not retain secretory polarity or the canalicular spaces

1602

BOYER AND SOROKA

GASTROENTEROLOGY Vol. 109, No. 5

have collapsed before the time of measurement as a result of continued secretion into the closed luminal space.~2'14

Statistics Data are expressed as the mean + SD and were considered significant when the P value was < 0 . 0 5 using the paired t test.

Results Figure 1 shows the effects of DBcAMP, nocodazole, and nocodazole plus DBcAMP on the canalicular membrane circumference measured approximately 4 hours after isolation of hepatocyte couplets. We have previously shown that, during this period, the canalicular membrane surface increases approximately 16-fold as a result of the targeting of membrane vesicles to the remaining apical domain. 13 Although canalicular circumference increased to a variable degree when maintained under control conditions for 4 hours, exposure to DBcAMP resulted in a significant increase in this measurement. These findings suggest that protein kinase A agonists stimulate vesicle targeting to the canalicular domain and are consistent with previous observations in intact isolated perfused liver preparations in which

horseradish peroxidase was used as a marker of vesicle transcytosis. 15 In contrast (Figure 1), the microtubule inhibitor nocodazole significantly reduced the size of the canalicular membrane circumference, consistent with the concept that microtubule-dependent transcytosis is required for the targeting of vesicles to the apical domain in hepatocytes) When DBcAMP was added to the cells that had been incubated in nocodazole for 2 hours, the canalicular perimeter size also significantly increased (P < 0.05) as compared with nocodazole alone (Figure 1). This finding suggests that there could be a pool of mem-

A

B

• 25

8

F° cell

!!i!!!~inii!ii:ii:iiiil F ° cell

_a lo0

I.-15

DB cAMP

N20

N20/cAMP

Figure 1. Canalicular circumference measurements in isolated rat hepatocyte couplets incubated in the presence of L-15 alone (1_-15), 100 pmol/L DBcAMP and 500 pmol/L IBMX (DB cAMP), 20 #mol/L nocodazole (N20), or 20 #mol/L nocodazole and DBcAMP and IBMX (N20/cAMP). Hepatocytes were cultured for 4 hours before fluorescent bile acid uptake. Couplets showing expanded canalicular spaces were randomly chosen using bright-field optics. Canalicular circumference measurements of the confocal fluorescent images were obtained with COMOS software. Values are mean _+ SD of 60 couplets per treatment obtained from six isolations. *DBcAMP is significantly increased over L-15 (P < 0.01) using paired analysis of Student's t test; +20 pmot/L nocodazole is significantly decreased as compared with L-15 alone (P < 0.001); #20 pmol/L nocodazole and DBcAMP and IBMX is significantly increased over 20 pmol/L nocodazole alone (P < 0.05).

% Secretion :

Uptake

:

F°can X 100 F°cell + F ° can F°cell + F°can pm z

Canalicular Circumference (pm) Figure 2. A fluorescent micrograph showing the excretion of CGamF into the canalicular lumen of an isolated hepatocyte couplet exposed to 1 #mol/L CGamF for 5 minutes, washed, and then incubated for an additional 10 minutes before placement of the cells on ice (bar = 10 pm).

November 1995

brane vesicles that are still capable of fusing with the apical membrane even when microtubule function is impaired.

Assessment of Bile Excretory Functional Capacity for Bile Acids When hepatocyte couplets cultured for the 4-hour experimental period are exposed to CGamF (1 btmol/L) in HEPES buffer for 5 minutes and then allowed to secrete for 10 minutes in the absence of the fluorescent bile acid, couplets that are secretory competent rapidly excrete the fluorescent bile acid into the canalicular space. Figure 2 shows the typical appearance of couplets that demonstrate this functional capacity. In the first series of experiments, when 20 couplets with expanded canaliculi in each treatment group were selected at random under bright-field optics and then examined with epifluorescent microscopy, 53% + 8% of 100 couplets from five separate isolations showed bright green fluorescence within their respective canalicular lumens. DBcAMP and IBMX markedly stimulated the canalicular excretion of the fluorescent bile acid so that 85% + 4% of the couplets now excreted the bile acid into their canalicular spaces. In contrast, when nocodazole-treated ceils were examined, only 12% + 10% of the hepatocyte couplets excreted the fluorescent bile acid into their canalicular lumens. Confocal microscopy and image analysis were then performed to obtain a more quantitative assessment of the secretory capacity in individual couplets and to adjust for the variability in the size of the ceils and their respective lumens. CGamF fluorescence in the canalicular lumen was measured for each couplet showing an expanded canaliculus, and was expressed as a percent of the total cellular (F° cell) plus canalicular fluorescence intensity (F° can), e.g., (F° can)/(F ° can) + (F° cell) X 100 = %. This approach avoided differences in uptake and excretion of the bile acid that occur between couplets based on heterogeneity in size and baseline transport function. The results of these findings are presented in Figure 3 and show that DBcAMP stimulated whereas nocodazole nearly completely prevented the accumulation of the fluorescent bile acids within the canalicular lumens. Because the fluorescent intensity of the canalicular lumen was saturated in some experiments in DBcAMP-treated couplets, the percent of the fluorescent organic anion that was excreted into the canalicular lumens during this period probably exceeded the recorded values. In addition, the confocal analysis showed that the addition of DBcAMP to nocodazole-treated couplets caused a significant increase in the percentage of CGamF that was

VESICLE TARGETING AND BILE ACID EXCRETION

20 1816=14o12_ ~10-

1603

T = t

I ........

J .........

ii!iii~!~iiiii~i!~!iiii!~i!i~i~iiil ~. :.:.:.:.:.:.:+:+:.:.:.:.:.:+:.:~

!ii~i~i!i~!~!~!!i~i!!ii!ii!!ii!!~iii!

o (n

::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::: !i!i~!i!i~i!ii#i!iiiii!i!i!i!~!}i . ....,..,..,........., ....

~iii¢i¢i¢i¢iii¢i¢i¢i¢; i:~:!:!:!:!:~:!¢~:~:!:~:i:!:i:!:i:! i!ii¢i¢i¢!¢iii¢i¢i¢!¢ i:~:i:[:~:~:i:!:!¢!:i:?:!:i:i:i:~:! !::::::::::::::::::::::::::::::::::: ................... ~i~i~i~i~i~i~i!~i~i~i~i~i~i!-

# +

ii!iii!iiiiiiiiiiiiiiii!i!ii!ii!!iii!_

L-15

DB cAMP

N20

N20/cAMP

Figure 3. Percentage of fluorescent bile acid excreted into the canalicular space in isolated rat hepatocyte couplets incubated in the presence of L-15 alone (L-15), DBcAMP and IBMX (DB cAMP), 20 p m o l / L nocodazole (N20), or 20 ~tmol/L nocodazole and DBcAMP and IBMX (N20/cAMP). Hepatocytes were exposed to 1 pmol/L CGamF for 5 minutes and allowed to secrete in the absence of fluorescent bile acid for an additional 10 minutes. Couplets were randomly chosen using bright4ield optics to detect expanded canalicular spaces. Fluorescent images were obtained by confocal fluorescent microscopy and analyzed with COMOS software by measuring the amount of fluorescence in the canalicular space. Percentage secretion was derived by F°can/(Pcell + F%an) × 100. Values represent the mean ± SD of 60 couplets per treatment obtained from six isolations. *DBcAMP increased secretion significantly as compared with L-15 alone (P < 0.01); +20 pmol/L nocodazole decreased secretion as compared with L-15 alone (P < 0.05); #20 btmol/L necodazole and DBcAMP and IBMX significantly increased secretion over 20 btmoi/L nocodazole (P < 0.05).

secreted compared with nocodazole alone despite the continued presence of nocodazole. To verify that nocodazole inhibition was indeed the result of the depolymerization of microtubules, the effects of removal of nocodazole were assessed on microtubule structure by immunofluorescence and on excretory function. Figure 4 shows immunofluorescent staining of microtubules in hepatocyte couplets that have been cultured for 4 hours in the presence or absence of two concentrations of nocodazole. Depolymerization of microrubules is complete when cells are treated with 20 btmol/L nocodazole, whereas a few residual, shortened microtubules (probably associated with the microtubule organizing center) are evident with 10 btmol/L nocodazole. Reformation of elongated, filamentous structures occurs when media containing 10 btmol/L nocodazole is replaced with L-15 media after 2 hours and the cells are allowed to recover for an additional 2 hours. The reversibility of the effects of nocodazole on the excretory capacity for fluorescent bile acids was assessed using confocal microscopy (Figure 5). Two hours after

1604

BOYER AND SOROKA

GASTROENTEROLOGY Vol. 109, No. 5

...... a : .... {~{{~{{

{i}{ } {

,

{ e>~}}{ 7{{ ~ {1 { :{ ,~

!~

{

=

>a<

~g ....

~: . . . . . . .

Figure 4. Immunofluorescent staining of microtubules in isolated rat hepatocyte couplets. (A) Control cells incubated in L-15 media. (B) Cells treated with 10 pmol/L nocodazole showing the absence of polymerized microtubules except for short segments presumably representing the microtubule organizing center. (C) Cells treated with 20 btmol/L nocodazole showing the complete absence of microtubules. (D) Cells treated with 10 ~tmol/L nocodazole for 2 hours, washed, and allowed to recover for an additional 2 hours, showing repotymerization of microtubules (bar = 20 ILtm).

removal of 10 Bmol/L nocodazole, and in parallel with the repolymerization ofmicrotubules as shown by i m m u nofluorescence, the percentage of CGamF secreted into the canalicular lumen recovered to levels similar to the untreated cells from the same isolation that were maintained for 4 hours in L-15 media (Figure 5). Similarly, the size of the canalicular circumference significantly increased in couplets allowed to recover for 2 hours from nocodazole treatment (P < 0.05). DBcAMP has been reported to increase Na+/taurocholate cotransport into rat hepatocyte suspensions. 16 To assess this possibility, the total fluorescence intensities per square micrometer in hepatocyte couplets were determined at the same time that measurements were obtained of the percent of CGamF excreted into the canalicular lumens, e.g., 10 minutes after an initial 5-minute exposure to 1 t.tmol/L CGamF. There were no significant differences in the values for total fluorescent intensity ex-

pressed per square m i c r o m e t e r in any of the t r e a t m e n t groups compared with values obtained in the cells in L-15 media alone (Figure 6). Thus, changes in the uptake of the fluorescent bile acid do not seem to account for the highly significant effects of the various treatments on the percent secreted into the canalicular lumens. Finally, when the circumference of each canalicular l u m e n (measured in the couplets that were analyzed for canalicular fluorescence intensity by confocal microscopy) was plotted as a function of the percent secretion for each respective canalicular lumen, excretion of C G a m F increased geometrically as a direct function of the canalicular circumference independently of the t r e a t m e n t administered (Figure 7). Because the surface area of the canalicular lumen is geometrically related to either its radius or cross-sectional circumference, these findings provide additional evidence that the

November 1995

VESICLE TARGETING AND BILE ACID EXCRETION 1605

functional capacity for CGamF excretion was closely related to the surface area of canalicular membrane present in any given isolated hepatocyte couplet.

Distribution of the Ca2÷,Mg2÷-Ecto-ATPase Because the fusion of transcytotic vesicles with the canalicular membrane should result in an increase in insertion of bile acid transporters as well as an increase in membrane area, we assessed this possibility using an antibody to a putative canalicular bile acid transporter, the Ca2+,Mg2+-ecto-ATPase, ~7'.8 and an immunofluo-

A

S 4sl A g 4oi

.

aol

.i:01 L-15 30

DB cAMP

NIO

rescent assay. Although it is unclear whether this protein represents the primary canalicular bile acid transporter, it has been shown to confer bile acid transport function to COS ceils in which it has been transfected. .8'19 In control couplets incubated in L-15 media alone, the ectoATPase was localized predominantly to the canalicular membrane and pericanalicular punctuate structures (Figure 8A). Peripheral plasma membrane staining was also noted in many couplets, albeit to a lesser degree. This has previously been described for the intact liver as well 2°'21 and probably represents the normal targeting of apical proteins first to the basolateral membrane. Distinct patches of plasma membrane containing large amounts of ecto-ATPase were also occasionally noted (Figure 8B and C; lower left of couplet). These patches are believed to be canalicular membrane remnants that have not undergone reorganization to the remaining canalicular domain in the isolated hepatocyte couplet model. 22 Electron microscopy has shown these areas to have a morphology consistent with the apical membrane and distinct from the surrounding plasma membrane. 22 When cells were stimulated with DBcAMP, there was a substantial decrease in the punctuate intracellular pattern (Figure 8B), suggesting fusion of transcytotic vesicles with the cana-

Reversal

g

25-

T I

.~_2 0 -

Q =

oQ 1 5 -

iiiiii!i!iiiiiii!ii

10

~

i

L-15 L-15

DB cAMP

N10

Reversal

Figure 5. Reversal studies showing recovery of (A) canalicular circumference and (B) percentage secretion in isolated hepatocyte couplets after treatment with 10 lamol/L nocodazole. Hepatocytes were incubated for 2 hours in the presence of 10 btmol/L nocodazole and then placed in L-15 alone for an additional 2 hours to allow recovery of the microtubules. Fluorescent bile acid uptake and microscopy was performed as described in previous figures. Values are mean _+ SD of 50 couplets per treatment obtained from five isolations. *Using paired Student's t test (P < 0,05). Although percentage secretion is not significantly increased in reversal couplets as compared with couplets treated for 4 hours with 10 btmol/L nocodazole, the trend was toward increased secretion (P < 0,09).

DB cAMP N20

N20/cAMP

Figure 6. Uptake of CGamF in isolated rat hepatocyte couplets incubated in the presence of L-15 alone (L-15), 100 btmol/L DBcAMP and 500 btmol/L IBMX (DB cAMP), 20 btmol/L nocodazole (N20), or 20 ,LtmoI/ L nocodazole and DBcAMP and IBMX (N20/cAMP). After 4 hours in culture, hepatocytes were exposed to 1 pmol/L CGamF for 5 minutes and then maintained in the absence of the fluorescent bite acid for an additional 10 minutes. Confocal images were analyzed using COMOS software by delineating the entire couplet and calculating the total fluorescence detected in the cytoplasm and canalicular space. This value was divided by the area to give fluorescence per square micrometer. Values represent the mean _+ SD of 60 couplets per treatment obtained from six isolations. No significant differences in CGamF uptake were observed between treatment groups using paired analysis of Student's t test.

1606 BOYERAND SOROKA

GASTROENTEROLOGYVol. 109, No. 5

totic vesicles with the canalicular membrane and insertion of more bile acid transport proteins.

g

Discussion 10-

o o U}

• ~

•e S

r =0.72, p
0

5

10 15 20 25 30 35 40 45 50

Canalicular circumference (gm) Figure 7.

Relationship between CGamF excretion and canalicular membrane circumference in rat hepatocyte couplets from all treatment groups combined. Note that the percent of total CGamF fluorescence that is secreted into the canalicular space is plotted as a log function as compared with the canalicular circumference expressed in micrometers. This finding is consistent with the mathematical relationship between canalicular membrane surface area and the luminal radius or circumference. The former is a geometric function of the luminal radius or circumference because surface area = 4/~r 2. It is the surface area of the canalicular membrane that is directly related to the transport capacity of the couplet and altered by the insertion or retrieval of pericanalicular membrane vesicles.

licular membrane. Indeed, when a blinded analysis was performed, one of us (J.L.B.) correctly identified the specific treatment in 81 of 100 couplets by this difference in the pattern of punctuate staining. An increase in staining was also noted at sites of remnant retention, suggesting that vesicle fusion can occur at these retained "apical" sites as well. In contrast, in nocodazole-treated cells, prominent fluorescent staining for ecto-ATPase was found not only at the canalicular membrane but at the peripheral plasma membrane and at intracellular sites distal to the canaliculus (Figure 8C). The presence of the ecto-ATPase at the canalicular membrane presumably represents protein that was retained from the original apical domains at sites of attachment of the two cells at the time of isolation as well as possibly from the microtubule-independent reorganization of the couplet plasma membrane system. The increased staining at peripheral sites and lack of pericanalicular punctuate staining is consistent with nocodazole blocking trafficking of the canalicular protein at the same time that it prevents transport of CGamF to the canalicular lumen. DBcAMP, on the other hand, stimulates the secretion of the fluorescent bile acid, decreases the pericanalicular punctuate staining of the ecto-ATPase, and increases the apical surface area, findings all consistent with fusion of transcy-

Previous studies have shown that isolated hepatocyte couplets are primary bile secretory units that retain cell polarity and normal secretory capacity when maintained in short-term culture. 1°'14'23'24 W i t h time in culture, secretion is elaborated into the canalicular lumen, which is sealed by tight junctions between two adjacent hepatocytes. Couplets also respond to the administration of bile acids with a prompt increase in fluid excretion within the enclosed canalicular space) 3 Their canalicular lumens undergo cycles of expansion and collapse either as a result of endogenous stimuli such as bile acids, increases in intracellular calcium, or because secretory pressure builds up and ruptures the tight junction seal. 1'14 Organic anions, including fluorescein, carboxyfluorescein, and 2',7'-bis(2-carboxyethyl)-5,(6)-carboxyfluorescein, are also transported into the canalicular lumen as detected by fluorescence microscopy, 23'25 and transcytotic vesicles labeled with a fluid phase marker (horseradish peroxidase) move to the canalicular region within 1 0 15 minutes, a time frame comparable with what has been observed in the intact liver. 24 However, in contrast to the normal liver, vesicle targeting to the apical domain in the isolated hepatocyte couplet seems to be greatly amplified. This occurs because all transcytotic vesicles as well as some of the hemicanalicular remnants from other regions of the cell membrane that are no longer in contact with neighboring hepatocytes after collagenase digestion are now targeted to the single remaining canalicular domain of the couplet. 13'24 Electron-microscopic studies indicate that this process results in an expansion in the apical membrane within 4 hours of isolation, which averages 16-fold over the size of the canalicular membrane at the time of isolation when couplets are maintained in a hormone-free L-15 basal media. 13 There is increasing evidence that regulation of plasma membrane transport function in many cells can occur by the recruitment and/or retrieval of cytoplasmic vesicles to and from the site where the protein functionally resides on the membrane. 26 For example, proton pumps are inserted into the luminal membrane of turtle bladders after acidification27; water channels are inserted in the surface membranes of amphibian urinary bladders after antidiuretic hormone stimulation, 28 and cystic fibrosis transmembrane regulator chloride channels are inserted on vesicles into the apical membrane of T-84 colonocytes after cAMP stimulation. 29'3° Hepatocytes are particularly interesting cells to examine such phenomenon because

November 1995

VESICLE TARGETING AND BILE ACID EXCRETION

1607

Figure 8. Immunofluorescence staining of rat hepatocyte couplets with antibody to the canalicular membrane Ca2+,Mg2+-ecto-ATPase.(A) A representative couplet maintained in L-15 media for 4 hours after isolation. (B) A couplet treated with DBcAMP and IBMX. (C) A couplet treated with 20 pmol/L nocodazole ( b a r = 10 Am).

their apical poles form the luminal membrane of the bile canaliculus and contain highly specialized transporters for the biliary excretion of a number of different organic and inorganic solutes. 2'31'32 In addition most, if not all, newly synthesized proteins that reside on the canalicular domain traffic there by an indirect route via microtubular-dependent transport vesicles after being sorted from the basolateral membrane. 5'6'33'34 In the present study, we examine the possibility that bile acid excretion, a major excretory function of the hepatocyte, is regulated in part by recruitment of pericanalicular cytoplasmic vesicles containing bile acid transporters to the canalicular domain. Several prior experimental observations provide indirect support for this hypothesis. Bile acid-induced choleretic responses correlate closely with canalicular membrane surface area in isolated rat hepatocyte couplets. 13 In addition, hypotonic stress induces fusion of vesicles with the canalicular domain in isolated perfused rat livers in parallel with changes in the transport maximum for bile acid excretion. 35 However, direct evidence for such a relationship is lacking. To provide such evidence, induction of changes in the surface area of the canalicular membrane should parallel changes in transport activity and be associated with translocation of the transport protein to this specific domain. By using hepatocyte couplets, we have been able to make direct measurements of changes in the apical membrane area from video images obtained with Nomarski optics or, as shown in this present study, by using fluorescent images obtained by confocal fluorescent microscopy (Figure 1). Measurements of the circumference of the canalicular space when obtained at its midpoint correlate closely with both volume and surface area mea-

surements (unpublished observation, September 1993) and thus are assumed in the present study to reflect the surface area of the canalicular membrane. The availability of a fluorescent analogue of bile acid (CGamF) provided us the opportunity to determine the functional capacity for canalicular bile acid excretion in the same hepatocytes in which calculations of the amount of apical membrane could be made. Finally, an antibody directed to a canalicular membrane Ca 2+, Mg2+-ecto-ATPase 36'3v has permitted us to assess the translocation of this putative bile acid transport protein from cytoplasm to apical membrane. Although it is unclear that this protein represents the primary canalicular bile acid transporter, when this ectoATPase is transfected into COS cells, it endows these cells with the capacity to secrete bile acids) 8'19 The results of these studies provide evidence that DBcAMP stimulates an increase in canalicular membrane surface area in association with an increase in the canalicular excretion of CGamF and at the same time affects translocation of a putative bile acid transport protein to the canalicular domain. In contrast, nocodazole treatment reduces the amount of surface membrane at the canaliculus, diminishes CGamF secretion, and results in a paucity of ecto-ATPase-containing vesicles in the pericanalicular region. Altogether, these combined studies suggest that bile acid excretion in rat hepatocytes is highly regulated by vesicle traffic to the canalicular domain. Although the transport protein(s) that perform the translocation of bile acids across the canalicular membrane may include proteins other than ecto-ATPase, which have not yet been purified or sequenced, their functional properties are readily assessed in these studies by the excretion of CGamF. We expressed this functional activity as the percent of the total fluorescence taken up

1608

BOYER AND SOROKA

into the cells to control for differences in uptake related to variation in hepatocyte size and because it is not possible to accurately quantitate the amount of bile acids taken up and excreted when using fluorescent markers. Differences in hydrophobicity and pH as well as a tendency for high concentrations of fluorescence compounds to self-quench contribute to this difficulty. In these experiments, conditions were standardized to minimize these potential problems. Because the activity of bile acid transporters on both sinusoidal and canalicular surfaces can also be regulated directly by phosphorylation-dephosphorylation reactions that are catalyzed by cAMP, the possibility that CGamF excretion into the bile lumens was also stimulated by phosphorylation of the transporter(s) cannot be excluded. 1<38'39 However, such effects seem unlikely to fully explain our results because CGamF accumulated in the canalicular lumens in a geometric fashion in parallel with linear changes in the surface membrane not only with DBcAMP but also in L-1 5 - and nocodazole-treated cells. This finding is most consistent with changes in insertion and/or retrieval of unit membrane containing the bile acid transporters at the canalicular domain. In addition, previous studies have also shown that bile acids stimulate fluid secretion into the canalicular lumen in rat hepatocyte couplets in direct proportion to differences in the surface area of their canalicular lumens when exogenous DBcAMP was not present. 13 Finally, the observation in the present study that pericanalicular punctuate structures containing the ecto-ATPase were observed less frequently in DBcAMP-stimulated cells provides additional evidence that vesicles containing this putative bile acid transporter were translocated to the apical domain. Quantitation of those results as well as verification o f vesicle movement await quantitative immunoelectron microscopic studies. Nocodazole treatment of ceils resulted in depolymerization of microtubules, a phenomenon that blocked transcytosis as previously reported] '24'4°'41 Inhibition of transcytosis had the anticipated effect of reducing the canalicular plasma membrane surface area, consistent with the known role for microtubules in the targeting of newly synthesized proteins to the canalicular domain. 5 That these effects were specific for microtubule-dependent vesicle movement is suggested by the ability to reverse this phenomenon when cells were allowed to reestablish apical targeting after removal of nocodazole from the media. A significant increase in canalicular circumference occurred after removal of nocodazole as well as an increase in CGamF secretion into the lumen. The increase in CGamF secretion did not reach statistical significance

GASTROENTEROLOGY Vol, 109, No. 5

(P < 0.09), probably because microtubule repolymerization was not fully recovered 2 hours later, when all measurements were made (Figure 5). The fnding that DBcAMP also increased canalicular circumference and excretion of CGamF in nocodazoleblocked hepatocytes is of additional interest. Although this phenomenon is not completely explained, it implies that a protein kinase A-responsive pool of vesicles that is not dependent on the function of microtubules may be present in hepatocytes, perhaps in the pericanalicular area or originating from a late subapical endosomal compartment. When phosphorylated by protein kinase A, these vesicles may be capable of fusion with the canalicular membrane. Previous studies have shown a well-organized subapical tubular-vesicular compartment in the pericanalicular region in hepatocyte couplets using electron-microscopic localization of the fluid phase marker horseradish peroxidase. 24 Studies in the isolated perfused rat liver that examined the transcellular movement and exocytosis of horseradish peroxidase indicated that some horseradish peroxidase-labeled vesicles are able to fuse with the canalicular membrane and discharge horseradish peroxidase into bile by microtubule-independent processes. 42 Apical targeting of vesicles from basolateral or early endosomes to a subapical or late endosomal compartment has been shown for receptor-mediated transcytosis of immunoglobulin A in MDCK cells and is dependent on intact microtubules. 43-45 I n contrast, transcytotic cargo that has already moved to subapical endosomes in MDCK cells can be discharged at the apical surface by mechanisms that are independent of the effects of nocodazole on microtubule function. 45 Furthermore, nocodazole has no effect on the ability of these vesicles to recycle between the apical membrane and a late endosomal compartment. 45 It is not known if vesicles recycle between the canalicular membrane and a late endosomal subapical compartment in hepatocytes, but the present effects of DBcAMP during nocodazole treatment are consistent with this possibility. These results might also be explained by the unique properties of the isolated hepatocyte couplet, which reorganizes (recycles?) apical proteins from canalicular remnants to the remaining apical pole of the couplet when placed in short-term culture after isolation from the intact liver. The apical beltlike canalicular membrane that normally completely encircles the hepatocytes in the intact organ is reorganized only to the remaining apical domain of the couplets located between the two adjacent cells and contributes in part to its progressive expansion. This retargeting of previously formed apical membrane is thought to occur largely by microfilament-dependent

November 1995

o

!D-7 Figure 9. Proposed model for the regulation of canalicular membrane bile acid transport. Vesicles containing the bile acid transporter(s) are endocytosed at the basolateral membranes of the hepatocyte couplet and are sorted in early endosomes, where they are then transported through the cell to the apical pole via transcytotic vesicles. This latter movement is microtubule dependent. In the basal state (L15 incubation), vesicles move to and fuse with the apical canalicular membrane. The possibility that they merge with a subapical late endosomal compartment cannot be excluded. Steady-state conditions allow visualization of ecto-ATPase-containing vesicles within the pericanalicular cytoplasm. Additions of DBcAMP (100 limol/L) and IBMX (500 limol/L) to the incubation media stimulates fusion of these vesicles with the canaticular membrane containing the ecto-ATPase and other putative bile acid transporters. Nocodazole disrupts the microtubules and prevents the movement of vesicles to the canalicuJar domain, resulting in a significant decrease in the canalicular surface area and ability to secrete bile acids. There seems to be a microtubule-independent population of vesicles, perhaps directly underlying the apical membrane, that remain capable of fusion in the presence of nocodazole in response to DBcAMP.

mechanisms because previous studies indicate that colchicine and cycloheximide have no effect on the apical relocation of a canalicular ATPase detected histochemically whereas cytochalasin D prevents this occurrence. 22 Intact microfilament function is also required to maintain vesicle recycling to and from the apical membrane and subapical endosomes in MDCK cells.46 Stimulation of apical targeting of vesicles containing transport proteins by cAMP-dependent mechanisms has also been shown for several other transport proteins in other epithelia, including the CFTR gene product in transfected cell lines, 29 vasopressin-induced water channels in the toad bladder, 28 insulin-induced Glut-4 expression, 47 and secretin-stimulated HCO3- excretion in the pig pancreas 48 and pig and rat bile duct epithelia, 49'5° among others, cAMP-dependent mechanisms also effect exocytosis in MDCK cells. 51'52 In addition, cAMP stimu-

VESICLE TARGETING AND BILE ACID EXCRETION 1609

lares formation of vacuolar apical compartments in subconfluent MDCK cellsJ 3 Vacuolar apical compartments are considered to be intracellular storage compartments consisting of apical plasma membrane that form when MDCK cells are prevented from forming cell-cell contacts. Intracellular canalicular vacuoles have also been described in single hepatocytes that no longer maintain a polarized apical surface membrane. Their formation is also stimulated by cAMPJ 4 Finally, the capacity of hepatocytes to regulate other canalicular transport proteins, including the apical C1-/ HCO3- exchanger 8 and the multiple organic anion transporter, 55 is also closely related to the ability of these cells to target vesicles to the canalicular domain. Together these findings provide growing support for the concept summarized in the model in Figure 9 that excretory function in hepatocytes is regulated by transcytotic vesicle movement and molecular events that control exocytosis and endocytosis at the apical membrane domain. On the basis of these i3ndings, we propose that the transcytotic pathway, formally believed to be "constitutive" in hepatocytes, has the capability to be regulated and therefore is susceptible to modification by a variety of signal transduction pathways.

References 1. Watanabe S, Smith CR, Phillips MJ. Coordination of the contractile activity of bile canaliculi: evidence from calcium microinjection of triplet hepatocytes. Lab Invest 1 9 8 5 ; 5 3 : 2 7 5 - 2 7 9 . 2. Arias IM, Che MX, Catmaitan Z, Leveille C, Nishida T, St Pierre M. The biology of the bile canaliculus. Hepatology 1993; 1 7 : 3 1 8 329. 3. Meier PJ, Knickelbein RG, Moseley RH, Dobbins JW, Boyer JL. Evidence for carrier-mediated chloride/bicarbonate exchange in canalicular rat liver plasma membrane vesicles. J Clin Invest 1985; 7 5 : 1 2 5 6 - 1 2 6 3 . 4. Bissig M, Hagenbuch B, Stieger B, Koller T, Meier PJ. Functional expression cloning of the canalicular sulfate transport system of rat hepatocytes. J Biol Chem 1 9 9 4 ; 2 6 9 : 3 0 1 7 - 3 0 2 1 . 5. Battles JR, Feracci HM, Stieger B, Hubbard AL. Biogenesis of the rat hepatocyte plasma membrane in vivo: comparison of the pathways taken by apical and basolateral proteins using subcellular fractionation. J Cell Biol 1987; 1 0 5 : 1 2 4 1 - 1 2 5 1 . 6. Schell M J, Maurice M, Stieger B, Hubbard AL. 5' Nucleotidase is sorted to the apical domain of hepatocytes via an indirect route. J Cell Biol 1992; 1 1 9 : 1 1 7 3 - 1 1 8 2 . 7. Hubbard AL, Stieger B, Bartles JR. Biogenesis of endogenous plasma membrane proteins in epithelial cells. Ann Rev Physiol 1989;51:755-770. 8. Benedetti A, Strazzabosco M, Ng OC, Boyer JL. Regulation of activity and apical targeting of the CI-/HC03- exchanger in rat hepatocytes. Proc Natl Acad Sci USA 1 9 9 4 ; 9 1 : 7 9 2 - 7 9 6 . 9. Boyer JL. The role of vesicle transport and exocytosis in bile formation and cholestasis: Influence of cell volume, pHi, hormones and bile acids. In: Gentilini P, Arias IM, Mclntyre N, Rodes J, eds. Cholestasis. New York: Elsevier, 1 9 9 4 : 6 9 - 7 8 . 10. Boyer JL, Phillips JM, Graf J. Preparation and specific application

1610

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26. 27.

28. 29.

BOYER AND SOROKA

of isolated hepatocyte couplets. In: Fleischer S, FleJscher B, eds. Methods in enzymology. New York: Academic, 1990:501-516. Schteingart DC, Eming S, Ton-Nu HT, Crombie DL, Hofmann AF. Synthesis, structure, and transport properties of fluorescent derivatives of conjugated bile acids. In: Paumgartner G, Stiehl A, Gerok W, eds. Bile acids and the hepatobiliary system. Boston: Kluwer Academic, 1994:177-183. Weinman SA, Graf J, Veith C, Boyer JL. Electroneutral uptake and electrogenic secretion of a fluorescent bile salt by rat hepatocyte couplets. Am J Physiol 1993; 264:G220-G230. Gautam A, Ng OC, Strazzabosco M, Boyer JL. Quantitative assessment of canalicular bile formation in isolated hepatocyte couplets using microscopic optical planimetry. J Clin Invest 1989; 83:565-573. Boyer JL. Isolated rat hepatocyte couplets: a model for the study of bile secretory function. In: Tavoloni N, Berk PD, eds. Hepatic transport and bile secretion: physiology and pathophysiology. New York: Raven, 1993:597-606. Hayakawa T, Bruck R, Ng OC, Boyer JL. DBcAMP stimulates vesicle transport and HRP excretion in isolated perfused rat liver. Am J Physiol 1990;259:G727-G735. Grune S, Engelking LR, Anwer MS. Role of intracellular calcium and protein kinases in the activation of hepatic Na+/taurocholate cotransport by cyclic AMP. J Biol t h e m 1993;268:1773417741. Sippel CJ, Ananthanarayanan M, Suchy FJ. Isolation and characterization of the canalicular membrane bile acid transport protein of rat liver. Am J Physiol 1990;258:G728-G737. Sippel CJ, Suchy FJ, Ananthanarayanan M, Pedmutter DH. The rat liver ecto-ATPase is also a canalicular bile acid transport protein. J Biol Chem 1993;268:2083-2091. Sippel CJ, McCollum MJ, Perlmutter DH. Bile acid transport by the rat liver canalicular bile acid transport/ecto-ATPase protein is dependent on ATP but not on its own ecto-ATPase activity. J Biol Chem 1994;269:2820-2826. Barr VA, Hubbard AL. Newly synthesized hepatocyte plasma membrane proteins are transported in transcytotic vesicles in the bile duct-ligated rat. Gastroenterology 1993; 105:554-571. Stieger B, Meier PJ, Landmann L. Effect of obstructive cholestasis on membrane traffic and domain-specific expression of plasma membrane proteins in rat liver parenchymal cells. Hepatology 1994; 20:201-212. Gautam A, Ng OC, Boyer JL. Isolated rat hepatocyte couplets in short-term culture: structural characteristics and plasma membrane reorganization. Hepatology 1987;7:216-223. Graf J, Gautam A, Boyer JL. Isolated rat hepatocyte couplets: a primary secretory unit for electrophysiologic studies of bile secretory function. Proc Natl Acad Sci USA 1984;81:65166520. Sakisaka S, Ng OC, Boyer JL. Tubulovesicular transcytotic pathway in isolated rat hepatocyte couplets in culture. Gastroenterology 1988; 95:793-804. Strazzabosco M, Sakisaka S, Hayakawa T, Boyer JL. Effect of UDCA on intracellular and biliary pH in isolated rat hepatocyte couplets and perfused livers. Am J Physio11991;260:G58-G69. Bradbury NA, Bridges RJ. Role of membrane trafficking in plasma membrane solute transport. Am J Physiol 1994;267:Cl-C24. AI-Awqati Q. Regulation of membrane transport by endocytic removal and exocytic insertion of transporters. Methods Enzymol 1989;172:49-59. Handler JS. Antidiuretic hormone moves membranes. Am J Physiol 1988; 255:F375-F382. Schwiebert EM, Gesek F, Ercolani L, Wjasow C, Gruenert DC, Karlson K, Stanton BA. Heterotrimeric G proteins, vesicle trafficking, and CFTR Cl- channels. Am J Physiol 1994;267:C272C281.

GASTROENTEROLOGY Vol. 109, No. 5

30. Bradbury NA, Jilling T, Berta G, Sorscher EJ, Bridges RJ, Kirk KL. Regulation of plasma membrane recycling by CFTR. Science 1992; 256:444-445. 31. Boyer JL, Graf J, Meier PJ. Hepatic transport systems regulating pHi, cell volume, and bile secretion. Ann Rev Physiol 1992;54:415-438. 32. Nathanson MH, Boyer JL. Mechanisms and regulation of bile secretion. Hepatology 1991;14:551-566. 33. Goltz JS, Woilkoff AW, Novikoff PM, Stockert RJ, Satir P. A role for microtubules in sorting endocytic vesicles in rat hepatocytes. Proc Natl Acad Sci USA 1992;89:7026-7030. 34. Maurice M, Schell MJ, Lardeux B, Hubbard AL. Biosynthesis and intracellular transport of a bile canalicular plasma membrane protein: studies in vivo and in the perfused rat liver. Hepatology 1994; 19:648-655. 35. Haussinger D, Hallbrucker C, Saha N, Lang, Gerok W. Cell volume and bile acid excretion. Biochem J 1992;288:681-690. 36. Lin SH. Localization of the ecto-ATPase (ecto-nucleotidase) in the rat hepatocyte plasma membrane. J Biol Chem 1989; 264:14403-14407. 37. Lin SH, Guidotti G. Cloning and expression of a cDNA coding for a rat liver plasma membrane ecto-ATPase. J Biol Chem 1989; 264:14408-14414. 38. Sippel CJ, Fallon RJ, Perlmutter DH. Bile acid efflux mediated by the rat liver canalicular bile acid transport/ecto-ATPase protein requires serine 503 phosphorylation and is regulated by tyrosine 488 phosphorylation. J Biol Chem 1994;269:19539-19545. 39. Divald A, Simpser E, Fisher SE, Karl PI. Vasopressin and phorbol12, 13-dibutyrate inhibit glucagon-or cyclic AMP-stimulated taurocholate uptake in isolated rat hepatocytes. Hepatology 1994; 20:159-165. 40. Kacich RL, Renston RH, Jones AL. Effects of cytochalasin D and colchicine on the uptake, transtocation, and biliary secretion of horseradish peroxidase and 14C sodium taurocholate in the rat. Gastroenterology 1983; 85:385-394. 41. Goldman IS, Jones AL, Hradek CT, Huling S. Hepatocyte handling of immunoglobulin A in the rat: the role of microtubules. Gastroenterology 1983; 85:130-140. 42. Hayakawa T, Ng OC, Ma A, Boyer JL. Taurocholate stimulates transcytotic vesicular pathways labelled by horseradish peroxidase in the isolated perfused rat liver. Gastroenterology 1990; 99:216-228. 43. Gruenberg J, Gdffiths G, Howell KE. Characterization of the early endosome and putative endocytic carrier vesicles in vivo and with an assay of vesicle fusion in vitro. J Cell Biol 1989;108:13011316. 44. Apodaca G, Katz LA, Mostov KE. Receptor-mediated transcytosis of IgA in MDCK cells is via apical recycling endosomes. J Cell Biol 1994; 125:67-86. 45. Barroso M, Sztul ES. Basolateral to apical transcytosis in polarized cells is indirect and involves BFA and Trimeric G. protein sensitive passage through the apical endosome. J Cell Biol 1994; 124:83-100. 46. Gottlieb TA, Ivanov IE, Adesnik M, Sabatini DD. Actin microfilamerits play a critical role in endocytosis at the apical but not the baso~ateral surface of polarized epithelial cells. J Cell Biol 1993;120:695-710. 47. Lienhard GE. Regulation of cellular membrane transport by exocytic insertion and endocytic retrieval of transporters. Trends Biochem Sci 1983;125-127. 48. Raeder MG. The origin of the subcellular mechanisms causing pancreatic bicarbonate secretion. Gastroenterology 1992; 103:1674-1684. 49. Buanes T, Grotmol T, Landsverk T, Raeder MG. Secretin empties bile duct cell cytoplasm of vesicles when it initiates ductular

November 1995

50.

51.

52.

53.

54.

HC03- secretion in the pig. Gastroenterology 1988;95:417424. Kate A, Gores CJ, LaRusso NF. Secretin stimulates exocytosis in isolated bile duct epithelial cells by a cyclic AMP-mediated mechanism. J Biel Chem 1992;267:15523-15529. Hansen SH, Casanova JE. GsA stimulates transcytosis and apical secretion in MDCK cells through cAMP and protein kinase A. J Cell Biol 1994;126:677-687. Pimplikar SW, Simons K. Activators of protein kinase A stimulate apical but not basolateral transport in epithelial Madin-Darby canine kidney ceils. J Biol Chem 1994;269:19054-19059. Brignoni M, Podesta EJ, Mele P, Rodriguez ML, Vega-Salas DE, Salas PJI. Exocytosis of vacuolar apical compartment (VAC) in Madin-Darby canine kidney epithelial cells: cAMP is involved as second messenger. Exp Cell Res 1993;205:171-178. Remy L, Reynier M-O, Hashieh IA, Gerolami A. Intracelluiar lumen in cultured hepatocytes of rats. Study of an original cytoplasmic infrastructure. CR Acad Sci Paris 1988;307:63-68.

VESICLE TARGETING AND BILE ACID EXCRETION 1611

55. Roelofsen H, Soroka C, Boyer JL. Cyclic AMP stimulates sorting of the canalicular multispecific organic anion transporter into the transcytotic bile secretory pathway in hepatocyte couplets (abstr). Hepatology 1994;20:174A.

Received January 20, 1995. Accepted May 16, 1995. Address requests for reprints to: James L. Boyer, M.D., Department of Medicine and Liver Center, Yale University School of Medicine, P.O. Box 208019, 333 Cedar Street, New Haven, Connecticut 06520-8019. Fax: (203) 785-7273. Supported by DK 25636 (to J.L.B.) and the hepatocyte isolation and cell culture core and morphology core from the Yale Liver Center (DK 34989). The authors thank Oi-Cheng Ng and John McGrath for technical assistance and Dr. Michael Nathanson for his helpful advice and analysis. Antibodies to the ecto-adenosine triphosphatase were a gift from Dr. F. Suchy.