Effect of tauroursodeoxycholate on actin filament alteration induced by cholestatic agents. A study in isolated rat hepatocyte couplets

367

Journal of Hepatology, 1993: 19:367-376 Elsevier Scientific Publishers Ireland Ltd.

HEPAT 01444

Effect of tauroursodeoxycholate on actin filament alteration induced by cholestatic agents. A study in isolated rat hepatocyte couplets

N a n c y T h i b a u l t a, Mich~le M a u r i c e b, Michel M a r a t r a t a, A n d r 6 C o r d i e r a, G 6 r a r d F e l d m a n n b a n d Fran~:ois Ballet ~ "Rh6ne-Poulenc Rorer S.A.. Drug Safe O' Department. 94403 Vitrv sur Seine. and hlNSERM U327. Facuht; tit" Mt;decbw Xavier Bichat. Universitt; Paris 7. Paris, France

(Received I April 1992)

The mechanism of the protective effect of ursodeoxycholic acid in cholestatic liver diseases remains unclear. Since there is evidence that alterations in the pericanalicular actin microfilament network play a major role in cholestasis, the aims of this study were (a) to determine the effect of the cholestatic agents, taurolithocholate (TLC) and erythromycin estolate (ERY), on F-actin distribution in isolated rat hepatocyte couplets and (b) to assess the effect of tauroursodeoxycholate (TUDC) and taurocholate on the modifications induced by these two compounds. F-actin was stained with fluorescein-isothiocyanate phalloidin and fluorimetric measurements were performed using a scanning laser cytometer ACAS 570. F-actin distribution was assessed in the couplets by the ratio of the pericanalicular area fluorescence/total couplet fluorescence (CF/TF). At non-cytotoxic concentrations, TLC (50, 100 tzM) and ERY (10, 50, 100 ~M) induced a significant accumulation of' F-actin around the bile canaliculus as indicated by increased fluorescence in the pericanalicular area and by the increased CF/TF ratio compared with the controls. Electron microscopy studies showed significant alterations in bile canaliculi microvilli in couplets treated with 100 #M TLC. Only a few canaliculi showed an increase in pericanalicular microfilaments after treatment with 100 #M ERY. As assessed by scanning laser cytometry, TUDC prevented changes in F-actin distribution when it was added to the medium with taurolithocholate or erythromycin estolate at equimolar concentrations. However, the morphological changes observed by electron microscopy after treatment with TLC were not modified by co-treatment with TUDC. Taurocholate was ineffective. We conclude that (a) abnormalities of pericanalicular F-actin microfilaments occur in two different models of cholestasis, (b) tauroursodeoxycholate prevents the accumulation of pericanalicular F-actin detected by scanning laser cytometry but not the morphological changes of the canaliculus observed by electronic microscopy. Therefore, in these experimental conditions, the protective effect of TUDC appears to be partial.

Key words." Actin; Bile acids and salts; Cholestasis; Erythromycin estolate; Hepatocyte couplets; Scanning laser cytometry; Taurolithocholate; Tauroursodeoxycholate

Several reports suggest that ursodeoxycholic acid could be a promising treatment for several chronic cholestatic liver diseases (1-7). It has also been shown that ursodeoxycholic acid and its taurine conjugate,

tauroursodeoxycholate (TUDC), prevent liver lesions in experimental cholestasis (8-12). However, the mechanism of this protective effect remains unclear. It is well known that alterations in pericanalicular

Correspon~h,nce to: Franqois Ballet M.D., Ph.D.. Rh6ne-Poulenc Rorcr. Drug Safety Department. BP 14. 94403 Vitry sur Seine Cedex. France. Ahhreviations: ACAS, anchored cell analyser system: CFDA, carboxylluorescein diacetate: DMSO, dimethyl sullbxide: ERY. erythromycin estolate: FCS, fetal calf serum: FITC, fluorescein isothiocyanate: IRHC, isolated rat hepatocyte couplet: PH, phalloidin: TC, taurocholate: TLC. taurolithocholate; TUDC, tauroursodeoxycholate; CF. pcricanalicular area Iluorescence: TF, tot:d couplet fluorescence.

368 microfilaments are present in several experimental models of cholestasis (13-21). Although hepatocytes in culture have been used to study the cytoskeletal modifications induced by cholestatic compounds, in this system the hepatocytes lose their structural and functional polarity. The isolated rat hepatocyte couplet (IRHC) is a new model developed by Oshio and Phillips (22) and Grafet al. (23). The IRHC represents a primary secretory unit that excretes bile into a closed canalicular space between two attached cells, and therefore maintains a highly polarized structure in short term culture. The aims of the present work were to use the IRHC model to study the effects of two well known cholestatic agents, taurolithocholate (TLC) (24) and erythromycin estolate (ERY) (13) on F-actin distribution and to assess the effects of TUDC and taurocholate (TC), two hydrophilic and non-cholestatic bile acids (24), on the modifications induced by TLC or ERY. A scanning laser cytometer was used to quantify Factin modification. This new technique provides accurate quantification of fluorescent probes in adherent cells. Morphological studies of the bile canaliculi were also performed by electron microscopy.

Materials and Methods

Preparation and culture of &olated rat hepatoc3'te couplets Isolated hepatocyte couplets were prepared from 200-250 g male Sprague-Dawley rats (Charles River, Saint Aubin les Elbeuf, France), fed ad libitum with laboratory pellets and tap water, by the non-recirculating collagenase perfusion method of Seglen (25) modified by Graf et al. and Gautum et al. (23,26). After anesthesia with ketamine hydrochloride (125 mg/kg body wt., Rh6ne Merieux, Lyon, France), the liver was perfused via the portal vein with 500 ml of a Ca2+/Mge+-free Hepes buffer solution at 37°C at a flow rate of 40 ml/min, and then dissociated with 250 ml of the same buffer supplemented with 0.05% (w/v) collagenase type I (Boerhinger Mannheim, Meckenheim, Germany), CaC12 10 mM and 0.8 units of trypsin inhibitor (Sigma Chemical Co, Saint Louis, MO) per unit of trypsin activ;ty in the collagenase at a flow rate of 20 ml/min. After collagenase perfusion, the liver was excised, minced in Leibovitz 15 (L-15) culture medium (Gibco Laboratories, Grand Island, NY), supplemented with 12 mM NaHCO3 and 10% fetal calf serum (FCS) (Gibco), filtered sequentially through two sizes of nylon mesh (60

N. THIBAULTet al. and 30 #m, Poly Labo, Strasbourg, France) and allowed to settle by gravity in a conical centrifuge tube (Nunc, Roskilde, Denmark) for three successive 10-rain periods to selectively sediment hepatocytes. Cell viability was assessed by the ability to exclude Trypan Blue (Flow Laboratories, lrvine, UK). The cells were finally suspended in an L-15 medium supplemented with 10% FCS. Cell suspensions were plated on Labtek chamber glass slides (Poly Labo) at a concentration of 100 000 cells/ml for fluorescence detection of F-actin and in 96 multiwell dishes (Nunc) at a concentration of 300 000 cells/ml (100 p,l in each well) for viability studies. The cells were then maintained in a 37°C, 5% CO_, incubator for 4 h.

Fluorescence microscopy The functional integrity of the organic anion transport system was qualitatively assessed by loading the cells with 5 #g/ml of carboxyfluorescein diacetate (CFDA) (Molecular Probes, Junction City, OR) in the medium following a pre-culture period of 4 h. After 15-30 min exposure, the cells were reincubated in dyefree medium for 10 min. Fluorescence and phase contrast microscopy were performed with a Zeiss Axiovert 405 M microscope (Carl Zeiss lnc, Thornwood, NY), equipped with an epi-illuminator filter for fluorescein fluorescence and phase contrast optics. Cytotoxicit), assessment The cytotoxicity of unlabelled phalloidin (Sigma), sodium taurocholate (Sigma), sodium taurolithocholate (Sigma), sodium tauroursodeoxycholate (Sigma) and erythromycin estolate (Sigma) was assessed by colorimetric MTT assay (27). Cells were exposed to the compounds in supplemented L-15 medium for a period of 2 h. TLC was dissolved in a final medium containing 1% dimethyl sulfoxide (DMSO, Sigma) and a control culture was established with the same DMSO concentration. Seven concentrations of each chemical were tested and 5 wells per concentration were analyzed. MTT (3(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide) (Sigma) was prepared as a 5-mg/ml stock solution in phosphate buffered saline (PBS) (Flow Laboratories) and added to the medium (10 p,I stock solution per 100 #1 medium). Plates were incubated at 37°C for 2 h. Finally, the medium was removed, 100/zl of DMSO were added to dissolve dark crystals (formazan precipitates), the plates were shaken, using a shaker (Flow Laboratories), for 10 rain and read at 570 nm on a Multiskan MCC spectrophotometer (Flow Laboratories). Results were expressed as percent of the control.

ACTIN A L T E R A T I O N A N D C H O L E S T A T I C AGENTS

Determhlation of F-acthl distribution in another series of experiments on I R H C prepared from 4 different isolations, unlabelled phalloidin (PH), TLC and ERY were added to the culture medium 4 h after plating and incubated at a final concentration of 10, 50 and 100 p.M for 2 h. In further experiments, IRHC from 4 different isolations were incubated in a medium containing 100 p.M T U D C , TC, TLC or ERY alone or 100 p.M T U D C or TC combined with 100 p.M of TLC or ERY for 2 h. After treatment, isolated hepatocytes were fixed at room temperature for 10 min in 3.75% formaldehyde in PBS. Cells were then incubated with 10 -6 M fluorescein isothiocyanate phalloidin (FITC phalloidin) (Molecular Probes) for 30 min for specific fluorescent staining of polymerized F-actin (cytoplasmic G-actin did not interact with the phallotoxin) (28). Unlabelled phalloidintreated hepatocytes required treatment with cold acetone (4°C) for 2 h before staining in order to remove unlabelled phalloidin from the binding sites for FITC phalloidin (20), and a control culture was performed with the same protocol. The slides were then washed 3 times with PBS to remove excess FITC phalloidin, the chambers were removed and 1 drop of glycerol/PBS mixture (9/I, v/v) was dispersed on each glass slide. Coverslips were mounted on the slides and sealed with nail polish. Fluorimetric measurements were performed on stained couplets using a Scanning Laser Cytometer ACAS 570 (Anchored Cell Analyser System) equipped with a 5W argon ion laser (Meridian Instrument, Okemos, MI). The optical system of the ACAS 570 is designed to excite a sample with a laser beam at a single wavelength and the fluorescence emitted by the probe is measured by photomultiplier tubes and pictured as a color coded digitized cell likeness on the computer screen. Consequently, FITC phalloidin can be measured with a single excitation at 488 nm and a single emission at 520 nm and above. Before each data collection, we verified system performance on a day to day basis: the laser was properly tuned, the beam centered, aligned and optimized. Then, a test curve was run with a fluorescent standard (fluorescent yellow slide) and compared with a stored calibration curve as follows: the difference between the area under the test curve and the area under the calibration curve is given by the computer and must be < 5%. Scan parameters were optimized to reduce bleaching: the laser power was decreased to 50 mW, the scan strength to 10%, the laser speed to 20 mm/s and the photomultiplier tube setting to 85°/,,. In these conditions, the maxi-

369 mum signal and the minimum background were detected. Bleaching was negligible because the exposure time was 0.2 s (control experiments on 20 couplets demonstrated a photobleaching < 1% during a scan). After data collection, all digitized cell images were processed identically. The cell autofluorescence was substracted by setting the background threshold at the fluorescence level of cells which were not stained by FITC-phalloidin (60 arbitrary units). To assess the total couplet fluorescence (TF) and the pericanalicular area fluorescence (CF), the whole couplet and the pericanalicular area were respectively enclosed with a mouse-driven cursor and subjected to image analysis. The pericanalicular area was defined, as described by Crawford et al. (29), as an ellipse with the junction of the two hepatocytes as its long axis and a transverse axis being a fifth of the transverse diameter of each cell (after fixation, all canaliculi collapse and the area measured represents only the pericanalicular ectoplasm). To avoid the subjectivity of this measurement, it was verified that the elliptical area designed as 'pericanalicular area' ocupied 10 ± 1.5% of the total couplet area (mean ± S.D.). In order to assess F-actin distribution, couplets from 4 separate isolations were examined: for each compound tested and for the controls, 20 couplets were randomly selected in a Labtek chamber and scanned for fluorescence. As previously shown (30) and as discussed further on (see Discussion), it was assumed that fluorescence was quantitatively related to the amount of F-actin in the cells. The CF/TF ratio was calculated for each couplet to evaluate the distribution of F-actin in the cells. This parameter indicates the relative amount of F-actin around the canaliculus. In our experience, this parameter provides a better estimation of F-actin distribution than the absolute value of fluorescence intensity which may vary significantly from one couplet to another (CF/TF: 1.45 ± 0.13; CF: 1190 ± 630 (arbitrary units); TF: 820 ± 401 (arbitrary units) in 100 non-treated couplets from 5 separate isolations).

Electron microscopy Cells were fixed in a solution of 2.5 % glutaraldehyde buffered with 0.1 M sodium phosphate buffer pH 7.4 for 30 min at 4°C, post-fixed in phosphate-buffered 1% osmium tetroxide solution for 30 min at room temperature, and embedded in epoxy resin. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with a Siemens Elemiskop IA electron microscope. Electron microscopy was performed on hepato-

370 cyte couplets incubated for 2 h with phalloidin (100 p,M) or with erythromycin (100 #M) or taurolithocholate (100 ~M) or tauroursodeoxycholate (100 /~M) or the mixture of taurolithocholate (100 #M) and tauroursodeoxycholate (100 #M). Untreated hepatocyte couplets were examined 4 h after plating.

Statistical methods Results are presented as mean .4- S.D.. Statistical analysis was performed using ANOVA.

N. THIBAULT et al. --O--

TUDC - ' • - "

Characterization of the model Hepatocyte viability, as measured by Trypan blue exclusion, was >85%. Sixty percent of the isolated hepatocytes attached to the glass slides. Cells cultured on glass remained spherical 8 h after isolation. Using 42 separate and successive isolations, couplets were counted after 4 h in culture: 28 .4- 7.5% cells had formed IRHC. Larger clusters of cells (3-6 cells) were also observed (less than 5%). Of the IRHC, 8 -4- 3% resealed a canalicular space of variable size (2-6 /~m in diameter). After 8 h in culture, the cells had flattened and the canalicular spaces collapsed or communicated with the surrounding media. Cellular uptake and biliary secretion of organic anion fluorescent dyes in I R H C were investigated. Nonfluorescent C F D A was rapidly taken into the cells, hydrolysed by hepatic esterases to form the fluorescent derivative carboxyfluorescein and secreted into the canalicular space. After incubation with CFDA, homogeneous cellular fluorescence was visualised within 2 min. Secretion of the fluorescent dye into the canalicular space was observed after 15-30 min indicating that the canalicular excretory system was functionally intact. Effects of unlabelled phalloidin, bile acids and erythromycin estolate on F-acth7 distribution In order to study the effect of unlabelled phalloidin, bile acids and erythromycin estolate at non-cytotoxic concentrations, the cytotoxic potential of these compounds was first assessed. In the range of concentrations studied, TUDC, TC and ERY were non-cytotoxic after incubation for 2 h. PH and TLC significantly impaired cell viability, with a CI 50 (concentration inducing 50% cell death) of about 750 and 260/,tM, respectively (Fig. 1). In subsequent experiments, TLC was tested at the maximal concentration of 100 0.M, and was shown to have no significant effect on cell viability (88 .4- 9% vs 100 .4- 9%; NS).

-.tl-.

EIRy - - • - -

TLC

-- +--

PH

140 120

O O

1OO

L O

80

~

so

~ m N

4O

~,, z~F-=Sr~_T T --*

--

T

-,--

T

. . . .

20

12s

Results

TC

2S0

SOO

lO0O

CONCEm'R~TION ( 1 ~ )

Fig. I. Dose-dependent effect of unlabelled phalloidin, bile acids and erythromycin estolate on hepatocyte viability (cells from 4 separate isolations). PH, T U D C , TC, TLC and ERY were added to the medium 4 h after plating and viability was measured after 2 h of exposure. Each point represents the percentage of viable cells compared with the controls. Range bars indicate means .4- S.D.

F-actin distribution was assessed for each compound on 20 couplets from 4 separate isolations and compared with untreated IRHC. Fluorescence was observed throughout the whole cell. The dense pericanalicular actin filament network was visualized as an intense fluorescent staining in the pericanalicular region (Fig. 2A). After administration of unlabelled phalloidin (designed to serve as a positive control), a significant increase in pericanalicular fluorescence, indicating a marked accumulation of F-actin around the canaliculi, was observed (Fig. 3). Treatment with TLC and ERY induced similar modifications (Fig. 2B,C). The increase was significant at 50 #M with TLC and 10 ~M with ERY (Table 1, Fig. 4) (the absolute fluorescence values of the pericanalicular area in 100 ~M TLC and ERY-treated couplets were 2136 .4- 731 units and 1864 ± 808 units, respectively vs. 1198 -4- 623 units for the controls. The absolute fluorescence values of the whole couplets were 885 -4- 358 units and 918 -4- 274 units, respectively, vs. 822 -4- 398 units for the controls). The effect of 100/,tM T U D C or TC alone or in combination with 100 ~M TLC or ERY is shown in Fig. 5. T U D C or TC alone did not induce any change in Factin distribution compared with the controls (controls, 1.47 -4- 0.15 vs TUDC, 1.52 .4- 0.1; TC, 1.49 -4- 0.13, NS). In the presence of T L C and .ERY, T U D C completely prevented the change in F-actin distribution. TC was ineffective (TLC, 2.42 .4- 0.15 vs. T U D C + TLC, 1.49 .4- 0.09, P < 0.001; TC + TLC, 2.39 + 0.18, NS; ERY, 2.05 -4- 0.12 vs. T U D C + ERY, 1.60 .4- 0.1, P < 0.001; TC + ERY, 1.98 .4- 0.16).

371

ACTIN ALTERATION A N D CHOLESTATIC AGENTS

~

kj

}} i 1~,~

+~

? } N2

• ~,,' •

,

,,;-

255? - 2'847 -

I 7

?i

,3

•

L.

Fig. 2. Visualization of F-actin in a normal (A). 100 #M ERY treated (B) and 100 t~M TLC treated (C) couplet obtained with the ACAS 570 (color coded digitized image). Treated hepatocytes were incubated for 2 h with 100 p.M TLC before fixing and staining. Note the pericanalicular accumulation of F-actin around the canaliculi in treated IRHC,

372

N. TH IB AU L T et al.

10

pM

~

80

pM

~

100

pM

l

4,00

l 10 pM

I-'--] 60 pM

l~

100 pM

4.00 A

3.20

O

,It'

w

4e

"r:

a.2o

O

O 410.

Ik IO

2.40

2.40

1.8o

:. i:

T

0.80

0.00

" '

i

0.00 PH

CONTROLS

CONTROLS

Fig. 3. Effect of PH on F-actin distribution in IRHC. Four hours after plating, 10, 50 and 100 #M PH were added to the medium and fluorescence staining of F-actin was performed 2 h after their addition. The total number ofcouplets examined were 20 for each dose and each isolation (4 separate isolations). The CF/TF ratio (mean + S.D., n = 8) represents F-actin accumulation around the canaliculus in the couplets. *Significant differences between treated and untreated IRHC (P < 0.001).

Electron microscopy In electron microscopy, unlike control hepatocyte couplets (Fig. 6a) where the ultrastructural appearance of bile canaliculi were similar to that observed in vivo, many bile canaliculi of phalloidin-treated hepatocytes (Fig. 6b) presented lesions characterized by a striking dilation of their lumina with a near complete loss of microvilli. There was a thickening of the microfilament network in the pericanalicular area. Cytoplasmic vesicles were visible at some distance from the plasma membrane of the bile canaliculi. In erythromycin-treated hepatocyte couplets (Fig. 6c), the appearance of bile canaliculi was similar to that observed with phalloidin with several differences: (a) the lesions were less pronounced, with some microvilli still visible in dilated canaliculi and the microfilament net-

tl'

Ii

"1"

1.8o

i,

0.80

411

T

TLC

ERY

Fig. 4 Effect of TLC and ERY on F-actin distribution in IRHC. IRHC were treated and fluorescence was measured as indicated in Fig. 3. Results are expressed as the CF/TF ratio (means ± S,D., n = 80). *Significant differences between treated and untreated IR H C (P < 0.001 ).

work less increased; (b) a few bile canaliculi were abnormal. In contrast, in taurolithocholate-treated hepatocyte couplets, almost all bile canaliculi were altered to some degree. There was dilation of the lumen and loss of microvilli in certain bile canaliculi, while in others, slim elongated microvilli were visible in the dilated lumina. In still others, the shape was dramatically altered (Fig. 6d), the microvilli were difficult to recognize and replaced by a lamellar transformation of the canalicular membrane. The thickening of the microfilament network was not easily discernible in any of these canaliculi. In hepatocyte couplets incubated with tauroursodeoxycholate

4.00 3.20 it ,m

14. I-IJ. O

,It,

2.40

1.60

TABLE I 0.80

Effect

of TLC

and

ERY

on

F-actin

distribution in

TLC Controls 5p.M 10 p.M 50 gM 100 #M

1.48 1.44 2.56 2.51

IRHC

ERY 1.50 4- 0.09 4-0.08 4- 0.10 ± 0.21" 4- 0.13"

1.51 2.28 2.17 2.06

0.00

4-0.11 4- 0.14" :t: 0.08* + 0.10"

Results are expressed as the CF/TF ratio (arbitrary fluorescence units) (mean ± S.D., n = 80). *Significant differences between treated and untreated IRHC (P < 0.001 ).

CONTROLS

TUDC

TC

TLC

TLC TLC • ÷ TUCK:: T C

ERY

ERY ERY + + TUOC T C

Fig. 5. Effect of 100 #M TU D C or TC on F-actin distribution when added to the medium alone or in combination with 100 #M of TLC or ERY. IRHC were treated and nuorescencc was measured as in Vig. 3. Results are expressed as the CF/TF ratio (means 4- S.D., n = 80). *Significant differences between treated and untreated IRHC (P < 0.001 ).

•

-'

~,,"

I ~ ,

+.~ '+.

,

•

•

.

+

.

.:.

~

.,:"~+~',~,

~,-~r'*. . : ~ - - -

•

"

."/I..+~.'~ " .

•

..

:+

, ,

/

,

3

-~'.

-+

..

< ]

41.

r'~'2,

'"'~

"

..._~..~..~.~:-,; ;~'::"~-.~'~:~'~:':

.: .:.-.

"

i

. " •,

"

'

~

~

"I~-"~'f~3"

.:i: .il. :::.. . . .

• • " :

<:

.

'

"

'"

•..

~,, ~,

.

•

v

~,'++

Fig. 6. Electron microscopic appearance of bile canaliculi from hepatocyte couplets either untreated (a) or treated with 100 #M phalloidin (b). 100 gM erythromycin (c), 100 p.M taurolithocholate (d), 100 ttM tauroursodeoxycholate (e) and 100 p.M taurolithocholate + 100 #M tauroursodeoxycholate (13. a, (control) the bile canaliculus exibits numerous microvilli and has the same appearance as bile canaliculi in vivo; b, phailoidin induces a dilation of the bile canaliculus, a loss of microvilli and the formation of a thick microfilament network (arrow) from which cytoplasmic organelles are excluded; c, erythromycin also induces a dilation of the canalicular lumen with a loss of microvilli and thickening of the microfilament network (arrow) but changes are less marked than with phalloidin; d, taurolithocholatc dramatically affects the shape of the bile canaliculus and of microvilli the thickening of the microfilament network is difficult to estimate; e. tauroursodeoxycholate induces a dilation of the bile canaliculus bul no changes in microvi]li and microfilaments - - note that the Golgi apparatus (GAI is also dilated; f. when lauroursodeoxycholate and taurolithocholate are added together to hepatocyte couplets, the shape of the bile canaliculus is altered as with taurolithocholate alone: the thickening of the microfilament network is also difficult to appreciate, × 12000• Bar, I tam.

374 (Fig. 6e), the main abnormality was a dilation of the bile caniculi lumina. A decrease in the number of microvilli was sometimes visible. The microfilament network was unchanged. In the cytoplasm of the cells, voluminous Golgi apparatus were visible. Finally, in hepatocyte couplets incubated with taurolithocholate and tauroursodeoxycholate, the ultrastructural appearance of the bile canaliculi was similar to that observed when the cells were incubated with taurolithocholate alone (Fig. 60.

Discussion

The alterations of pericanalicular F-actin induced by TLC and ERY were studied using the IRHC model. This model retains structural and functional polarity (31-35), and enables hepatocytes to maintain their secretory function as demonstrated by the canalicular accumulation of carboxyfluorescein (23,36). Although previous studies using standard histological techniques (21,37,38) and fluorescent dyes (15,20,39) showed the redistribution of F-actin in cholestasis, these modifications could not be quantified. The scanning laser cytometer permits a quantitative assessment of fluorescence distribution in adherent cells. Although it is not certain that FITC phalloidin fluorescence intensity is quantitatively related to F-actin, this is probably reasonable since phalloidin binds covalently to actin with a stoichiometry of 1:1 (40) and when saturating quantities of fluorescent phalloidin are used, fluorescence is quantitatively related to the amount of F-actin (28). Also, phalloidin, which has been shown to increase the amount of F-actin in the pericanalicular area in vivo (15,21,41) and in cultured hepatocytes (20), significantly increased canalicular fluorescence in the couplets. Finally, electron microscopy clearly showed an accumulation of microfilaments in the pericanalicular area. This modification is very similar to the lesions observed in vivo in bile canaliculi when phalloidin is administered in the animal (41,42). Most important, phalloidin acts very rapidly in vitro; alterations of bile canaliculi were obtained in only 2 h while it took at least 1 day to observe phalloidin effects in vivo (41,42). Watanabe et al. (19) observed changes in bile canaliculi in primary cultures of rat hepatocytes from animals pretreated 1 day before the culture. These findings provide further evidence that increased pericanalicular fluorescence is related to the accumulation of F-actin. The results obtained by scanning laser cytometry suggest that TLC and ERY, two compounds which induce cholestasis in vivo (13,24), induce an accumulation of F-

N. T H I B A U L T et al.

actin in the pericanalicular region. Indeed, the significant increase in the CF/TF ratio was secondary to an increase in the absolute value of fluorescence intensity in the pericanalicular area (CF). In order to better understand the modifications of perican~dicular F-actin detected with scannning laser cytometry, a morphological study of the bile canaliculi was performed by electronic microscopy. After TLC treatment, significant modifications of the canalicular morphology were observed with electron microscopy. These modifications are very similar to those reported in vivo in the rat (43-45) and recently, in the IRHC model (46). However, it must be emphasized that visible microfilament accumulation was not revealed in the pericanalicular area with electron microscopy. The reasons for this discrepancy are unclear but could be because only ultrathin sections of cells are studied with electron microscopy whereas with scanning laser cytometry, fluorescence intensity measurements correspond to the total thickness of the hepatocyte couplet. After treatment with erythromycin estolate, morphological modifications of the canaliculi were observed but thickness of the microfilament network was only increased in a few bile canaliculi. There are little data on the canalicular modifications observed in erythromycininduced cholestasis (47). An increase in pericanalicular microfilaments has been reported in one patient with cholestasis after treatment with erythromycin estolate (48). The absence of a constant increase in pericanalicular microfilaments using electron microscopy in our experimental conditions is unclear but again might be due to the difficulty to quantify actin accumulation on ultrathin sections. A marked increase in the pericanalicular microfilamentous network also has been demonstrated in familial cholestasis in certain North American Indian children (49). Studies on isolated hepatocytes treated with other cholestatic compounds including nor-ethandrolone (14), cytochalasin B (20) or chlorpromazine (50), show the same phenomenon. Untreated IRHC from common bile duct ligated rats also showed an increase in pericanalicular F-actin (work in preparation). This result supports previous observations (21,39), and indicates that the modifications observed in IRHC in vitro, after exposure to TLC, may be relevant to cholestasis in vivo. The relationship between this accumulation and bile flow impairment remains unclear. Actin filaments, present in relatively large amounts in the pericanalicular area (34,51), are thought to play a role in the contraction and maintenance of canalicular wall shape (17,52,53). Microfilaments have also be found at the tight junctions and they may play a role in

375

ACTIN ALTERATION AND CHOLESTATIC AGENTS

the modulation of tight junction permeability (54). Finally, microfilaments and microtubules may also be involved in intracellular vesicle transport, a mechanism recently demonstrated as playing a role in the cellular transport of bile acids (55,56). It is not clear whether the biochemical mechanism of the increase of F-actin around the bile canaliculus with TLC and ERY was a consequence of an inhibition of Factin depolymerization as demonstrated with phalloidin, an increased synthesis of G-actin or another higher reorganization of actin (i.e., redistribution, bundling or reorientation). In contrast to TLC and ERY, TUDC - - a hydrophilic and non-toxic bile acid (57) - - did not change F-actin distribution in the couplet but induced a dilatation of bile canaliculi and an increase in the volume of the Golgi apparatus. TUDC prevented the alteration in F-actin distribution studied by scanning laser cytometry when it was combined with TLC. TUDC did not prevent the morphological changes induced by TLC. This suggests that the increase in the number and size of the microvilli is not directely related to the amount of F-actin in the pericanalicular region. In this study, we used another hydrophilic bile acid, TC, in order to determine if this protective effect was specific to TUDC; TC alone did not change F-actin distribution in the couplets and contrary to TUDC, did not prevent F-actin alteration induced by TLC or ERY. These results are a strong argument for the specific effect of TUDC and suggest that TUDC acts against a mechanism which is common to TLC and ERY, e.g. intracellular calcium fluxes (58-61) or actin polymerization (62). In conclusion, our data indicate that the change in Factin distribution induced by TLC and ERY can be prevented by simultaneous administration of TUDC. Nevertheless, this effect appears to be partial since TUDC did not prevent the morphological changes of the bile canaliculus observed by electronic microscopy. Futher studies are needed to clarify the effects of bile acids on the pericanalicular F-actin.

2 3

4

5 6

7

8 9

l0

II 12

13 14 15

16

17 18

19 20

21

Acknowledgements We would like to thank Prof Raoul Poupon for his very useful comments during the study.

22 23 24 25

References 26 Poupon R, Chretien Y, Poupon RE, Ballet F, Calmus Y, Darnis F. Is ursodeoxycholate an effective treatment for primary biliary cirrhosis? Lancet 1987; i: 834-6.

27

Chazouilleres O, Poupon R, Capron JP, et al. Ursodeoxycholic acid for primary sclerosing cholangitis..I Hepatol 1990; I1: 120-3. Poupon RE, Eschw~ge E, Poupon R and UDCA-PCB Study Group. Ursodeoxycholic acid for the treatment of primary biliary cirrhosis. Interim analysis of a double-blind multicenter randomised trial. J Hepatol 1990; I1: 16-21. Leuchner U, Fisher H, Kurtz W, Hubner K. Ursodeoxycholic acid in primary biliary cirrhosis: results of a controlled double blind trial. Gastroenterology 1989; 87: 1268-74. Colombo C, Setchell KD, Podda M, et al. Effects of ursodeoxycholic acid therapy for liver disease associated with cystic fibrosis. J Pediatr 1990; 117: 482-9. Cotting J, Lentze M, Reichen J. Effects of ursodeoxycholic acid treatment on nutrition and liver function in patients with cystic fibrosis and longstanding cholestasis. Gut 1990; 31: 918-21. Hayashi H, Higuchi T, Ichimiya H, Hishida N, Sakamoto N. Asymptomatic primary sclerosing cholangitis treated with ursodeoxycholic acid. Gastroenterology 1990; 99: 533-5. Kitani K, Kanai S. Tauroursodeoxycholate prevents taurocholate induced cholestasis. Life Sci 1982; 30: 515-23. Sch61meric J, Kitamura S, Baumgartner U, Miyai K, Gerok W. Taurohyocholate, taurocholate and tauroursodeoxycholate but not tauroursocholate and taurodehydrocholate counteract effects of taurolithocholate in rat liver. Res Exp Med 1990; 190: 121-9. Sch61meric J, Baumgarmer U, Miyai K, Gerok W. Tauroursodeoxycholate prevents taurolithocholate-induced cholestasis and toxicity in rat liver. J Hepatol 1990; 10: 280-3. Schumcker DL, Ohta M, Kanai S, Sato Y, Kitani K. Hepatic injury induced by bile salts: correlation between biochemical and morphological events. Hepatology 1990; 12: 1216-21. Heuman DM, Mills S, Mc Call J, Hylemon PB, Pandak WM, Vlahcevic ZR. Conjugates of ursodeoxycholate protect against cholestasis and hepatocellular necrosis caused by more hydrophobic bile salts. In vivo studies in the rat. Gastroenterology 1991; 100: 203-11. Plaa GL, Priestly BG. Intrahepatic cholestasis induced by drugs and chemicals. Pharmacol Rev 1977; 28: 207-73. Phillips MJ, Oda M, Funatsu K. Evidence for microfilament involvement on norethandrolone-induced intrahepatic cholestasis. Am J Pathol 1978; 93: 729-44. Low RB, Low ES, Chaponnier C, Mitchell JW, Gabbiani G. Effect of phalloidin on liver actin distribution, content and turnover. J Cell Biochem 1982; 20: 393-407. Reichen J, Simon FR. Cholestasis. In: Arias l, Popper H, Schachter D, Shafritz DA, eds. The Liver: Biology and Pathobiology. New York: Raven Press, 1982; 785-80. Phillips MJ, Oshio C, Miyiari M, Watanabe S, Smith CR. What is actin doing in liver cells? Hepatology 1983; 3: 433-6. Phillips MJ, Oshio C, Miyiari M, Smith CR. lntrahepatic cholestasis as a canalicular motility disorder. Evidence using cytochalasin. Lab Invest 1983: 48:205-11. Watanabe S, Miyiari M, Oshio C, Smith CR, Phillips MJ. Phalloidin alters bile canalicular contractibility in primary monolayer cultures of rat liver. Gastroenterology 1983; 85: 245-53. Nickola i, Frimmer M. Effects of phalloidin and cytochalasin B on cytoskeletal structures in cultured rat hepatocytes. Cell Tissue Res 1986; 245: 635-41. Kachi K, Okanoue T, Morioka H, et al. immunoelectron microscopic localization of actin in normal and cholestatic rat hepatocytes. Gastroenterol Jpn 1989; 24: 523-7. Oshio C, Phillips MJ. Contractibility of bile canaliculi: implications for liver function. Science 1981; 212: 1041-2. 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: 6516-20. Palmer RH. Bile acids, liver injury and liver diseases. Arch Intern Med 1972; 130: 606-17. Seglen PO. Preparation of isolated rat liver cells. Methods Cell Biol 1976; 13: 29-83. Gautam A, Ng OC, Boyer JL. Isolated hepatocytes couplets in short-term culture: structural characteristics and plasma membrane reorganization. Hepatology 1987; 7: 216-23. Mosmann T. Rapid colorimetric assay for cellular growth and

376

28

29

30 31

32 33

34 35

36

37

38

39

40 41 42

43 44

45

N. THIBAULT et al. survival: application to proliferation and cytotoxicity assays. J Immunol Methods 1983; 65: 55-63. Wulf E. Deboben A, Bautz FA, Faulstich H, Weiland Th. Fluorescent phallotoxin, a tool for the visualization of cellular actin. Proc Natl Acad Sci USA 1979; 9: 4498-502. Crawford JM, Vinter DW, Golan JL, Taurocholate induces pericanalicular localization of C6-NBD-ceramide in isolated hepatocyte couplets. Am J Physiol 1991; 260: G119-32. Howard TH, Meyer WH. Chemotactic peptide modulation of actin assembly and locomotion in neutrophils. J Cell Biol 1984: 98: 1265-71. Boyer JL, Gautam A, GrafJ. Mechanisms of bile secretion: insights from the isolated rat hepatocyte couplet. Semin Liver Dis 1988; 8: 308-16. Graf J, Boyer JL. The use of isolated rat hepatocyte couplets in hepatobiliary physiology. J Hepatol 1990; 10: 387-94. Gautam A, Ng OC, Stazzabosco M, Boyer JL. Quantitative assessment of canalicular bile formation in isolated hepatocyte couplets using microscopic optical planimetry. J Clin Invest 1990; 83: 565-73. Nickola I, Frimmer M. Preservation of cellular polarity in isolated hepatocytes. Cell Tissue Res 1986; 243: 437-40. Weinmann SA, Graf J, Boyer JL. Voltage driven, taurocholatedependent secretion in isolated hepatocyte couplet. Am J Physiol 1989; 256: G826-32. Fentem JH° Foster B, Mills CO, Coleman R. Chipman JK. Biliary excretion of fluorescent cholephiles in hepatocyte couplets: an in vitro model for hepatobiliary and hepatotoxicity studies. Toxicol in vitro 1990; 4: 452-7. Prentki M, Chaponnier C, Jeanrenaud B, Gabbiani G. Actin microfilaments, cell shape, and secretory processes in isolated rat hepatocytes. J Cell Biol 1979: 81: 592-607. Yasuura S, Ueno T, Watanabe S, Hirose M, Namihisa T. Immunocytochemical localization of myosin in normal and phalloidin-treated rat hepatocytes. Gastroenterology 1989; 97: 982-9. Tamura K, Kuroda H, Watanabe S. Yokoi Y. Actin filaments of hepatocytes in experimental rat cholestasis. Observations using a fluorescent staining method by DACM labeled heavy meromyosin. Acta Histochem Cytochem 1981: 14: 661-9. Weiland T. Poisonous principles of mushrooms of the genus Amanita. Science 1968; 159: 946-52. Dubin M. Maurice M, Feldmann G. Erlinger S. Phalloidin induced cholestasis in the rat: relation to changes in microfilaments. Gastroenterology 1978; 75: 450-5. Gabbiani G, Montesano R, Tuchweber B, Salas M, Orci L. Phalloidin-induced hyperplasia of actin filaments in rat hepatocytes. Lab Invest 1975: 33: 562-9. Layden TJ, Schwarz J. Boyer JL. Scanning electron microscopy of the rat liver: studies of the effect of taurolithocholate and other models ofcholestasis. Gastroenterology 1975: 69: 724-38. Miyai K, Richardson AL, Mayr W, Javitt NB. Subcellular pathology of rat liver in cholestasis and choleresis induced by bile salts. Lab Invest 1977; 36: 249-58. Loosli H, Gardiol D. Gautier A. Experimental intrahepatic cholestasis. A comparative physiopathological, ultrastructural and cytochemical study of different types. Virchows Arch Cell Pathol 1981; 35: 213-28.

46

47 48

49

50 51 52 53

54

55

56

57

58 59

60

61 62

Watanabe N, Miyiari M, Suyama K, Oshio C, Oda M, Tsuchiya M. Effect of taurolithocholate on the bile canalicular motilities in isolated rat hepatocytes. J Clin Electron Microsc 1989; 22: 5-6. Zafrani ES, lshak KG, Rudski C. Cholestatic and hepatocellular injury associated with erythromycin esters: report of nine cases. Dig Dis Sci 1979; 24: 385-96. Adler M, Chung KW, Scaffner F. Pericanalicular hepatocytic and bile ductular microfilaments in cholestasis in man. Am J Pathol 1980; 98: 603-16. Weber AM, Tuchweber B, Youssef l, et al. Severe familial cholestasis in North American Indian children: a clinical model for microfilament dysfunction? Gastroenterology 1981; 81: 653-62. Kawahara H, Marceau N, French S. Effects of chlorpromazine and low calcium on the cytoskeleton and secretory function of hepatocytes in vitro. J Hepatol 1990; 10: 8-16. French SW, Davies PL. Ultrastructural localization ofactin-like filaments in rat hepatocytes. Gastroenterology 1975; 68: 765-74. Phillips MJ, Oshio M, Miyiari M, Katz H, Smith CR. A study of bile canalicular contractions in isolated hepatocytes. Hepatology 1982; 2: 763-8. Miyiari M, Oshio C, Watanabe S, Smith CR, Youssef IM, Phillips MJ. Taurocholate accelerates bile canalicular contractions in isolated rat hepatocytes. Gastroenterology 1984: 87: 788-92. Marada JL. Stafford J, Barenberg D, Carlson S. Functional coupling of tight junctions and microfilaments in T84 monolayer. Am J Physiol 1988; 254: G416-23. Lamry Y, Roda A, Dumont M, Feldmann G, Erlinger S. lmmunoperoxidase localization of bile salts in rat liver cells. Evidence for a role of the Golgi apparatus in bile salt transport. J Clin Invest 1988; 82: 1173-82. Kitamura T, Gatmaitan Z, Arias IM. Serial quantitative image analysis and con focal microscopy of hepatic uptake, intracellular distribution and biliary secretion of a fluorescent bile acid analog in rat hapatocyte doublets. Hepatology 1990; 12: 1358-64. Hofmann AF, Roda A. Physicochemical properties of bile acids and their relationship to biological properties: an overview of the problem. J Lipid Res 1984; 25: 1477-89. Anwer MS, Engelking LR, Nolan K, Sullivan D. Zimniak P, Lester R. Hepatotoxic bile acids increase cytosolic Ca ++ activity of isolated rat hepatocytes. Hepatology 1988; 8: 887-91. Combettes L, Dumont M, Berthon B, Erlinger S, Claret M. Release of calcium from the endoplasmic reticulum by bile acids in rat liver cells. J Biol Chem 1988: 263: 2299-303. Beuers U. Thiel M, Bardenheuer H, Paumgartner G. Tauroursodeoxycholate acid inhibits the Ca ++ increase in human neutrophils stimulated by formyl-methionyl-leucyl-phenylalanine. Biochem Biophys Res Commun 1990: 171:1115-21. Watanabe S, Phillips MJ. Ca -'+ causes active contraction of bile canaliculi: direct evidence from microinjection studies. Proc Natl Acad Sci USA 1984; 81: 6164-8. Tuchweber B, Desroches S, Yousef IM, Gicquaud C, Weber AM, Loranger A. Effects of bile acids on actin polymerization in vitro. Life Sci 1990: 48: 1299-307.