METHODS

Materials

Cholyl-lysyl-fluorescein (CLF) was synthesised according to Mills and colleagues.^{38, 39} Collagenase (type A) was purchased from Gibco (Paisley, Scotland) and β-MC from Steraloids Inc (Newport, USA). Bovine serum albumin was purchased from Winlab (Maidenhead, UK), TUDC, TLC, staurosporine (SP), 1,2-bis-(o-aminophenoxy)-ethene-N,N,N`,N`-tetra-acetate tetra-(acetomethyl)ester (BAPTA/AM), H7, DBcAMP, KT2057, Leibovitz-15 tissue culture medium (L-15), and other fine chemicals from Sigma (Poole, UK).

Animals

Male Wistar rats (230–240 g body weight) had free access to a standard laboratory diet (41B maintenance diet; Pilsbury, Birmingham UK) and tap water ad libitum. Surgery was performed between 8 and 9 am to minimise circadian variations. Anaesthesia was achieved with ketamine hydrochloride 6 mg/100 g body weight and medetomidine 25 μg/100 g body weight. Protocols were approved according to the Animals Scientific Procedures Act 1986.

Isolation, enrichment, and culture of hepatocytes

Hepatocyte couplets were obtained from rat liver according to Wilton and colleagues,⁴⁰ adapted from Gautam and colleagues.⁴¹ The final preparation had a viability >90% (trypan blue exclusion) and contained 25–35% couplets. It was enriched using centrifugal elutriation to 70–80% couplets, with viability >95%. Couplets were plated in L-15 medium containing penicillin/streptomycin in 35 mm plastic culture dishes (2 ml per dish) at 0.5×10⁶ units/ml, and incubated at 37°C for 4.5 hours prior to experimentation.

Assessment of canalicular vacuolar accumulation (cVA)

Couplets were incubated for 15 minutes at 37°C with 2 μM CLF (final concentration). The fluorescent bile salt was then rapidly washed away twice with 2 ml L-15 medium and couplets were immediately examined using an inverted fluorescence microscope (IMT2-RFL; Olympus, London, UK). The number of couplets in four randomly selected regions of the plate (150–250 in each experiment) accumulating CLF into a fluorescent canalicular vacuole, judged by visual inspection as a proportion of the total couplets analysed, was then determined. This parameter is referred to as “cVA of CLF”, and is an indication of overall biliary secretion—that is, uptake, transcellular transport, canalicular secretion, and intracanalicular retention. Examples of a fully functional couplet preparation (control), a TLC inhibited preparation, and a preparation protected from TLC by TUDCA are shown in fig 1. The couplets counted as positive are those showing enough accumulation to be visualised (thresholded by the eye); within this population there may be a range of different accumulating abilities representing different original positions in the acinus, different expression of transporters, etc. Change in the proportion of accumulating couplets in the population represents a shift in the range of the couplet abilities. This semiquantitative approach enables rapid examination of the secretory behaviour of a mixed population of hepatocyte couplets and therefore is more likely to reflect what happens to the whole liver than analysis of the performance of an individual couplet. Furthermore, it provides similar information to other more quantitative but time consuming alternatives, using image analysis. Indeed, quantitative estimation of the amount of CLF retained in the canalicular vacuole (estimated as the product between mean canalicular fluorescence intensity and canalicular area corrected by total couplet fluorescence) showed a close correlation with semiquantitative estimation of the proportion of canalicular vacuoles displaying a visible amount of CLF.⁴²

Estimation of the influence of BAPTA/AM, H7, SP, and KT5720 on cVA of CLF

Couplets were incubated for 30 minutes with 10 μl of dimethyl sulphoxide (DMSO) (vehicle control), 20 μM BAPTA/AM (a Ca²⁺ chelator) (concentrations represent final concentrations in the incubation medium), 5 μM KT5720 (a specific PKA inhibitor), 100 μM H7 (a PKC/PKA inhibitor), and 1 μM SP (a pan-specific PKC inhibitor), as appropriate. CVA of CLF was analysed as described above. Concentrations of these compounds were chosen by making a balance between their effectiveness and specificity, based on Km and Ki inhibitory concentrations for their respective substrates, as described by the manufacturers.

Estimation of the influence of inhibitors on the cholestatic effect of TLC

Couplets were preincubated with 10 μl of DMSO (vehicle control), 20 μM BAPTA/AM, 100 μM H7, 1 μM SP, or 5 μM KT5720 (as appropriate) for 15 minutes. Then, TLC (2.5 μM) was added, and cells were incubated for another 30 minutes; cVA of CLF was then analysed as described above.

Estimation of the effect of TUDC, β-MC, and DBcAMP on TLC induced cholestasis

Couplets were incubated for 30 minutes with 10 μl of DMSO (vehicle control), TLC (2.5 μM), TLC (2.5 μM)+β-MC (2 μM), TLC (2.5 μM)+TUDC (2 μM), or TLC (2.5 μM)+DBcAMP (10 μM), as appropriate. All concentrations were derived from preliminary studies (data not shown). CLF (2 μM) was then added and incubated at 37°C for 15 minutes before washing twice with 2 ml of L-15; couplets accumulating CLF were counted as described above.

Estimation of the effect of inhibitors on hepatoprotection with TUDC and β-MC in TLC induced cholestasis

Couplets were preincubated with 10 μl of DMSO (vehicle control), 20 μM BAPTA/AM, 100 μM H7, 1 μM SP, or 5 μM KT5720 for 15 minutes. Then, TLC (2.5 μM), TLC (2.5 μM)+TUDC (2 μM), or TLC (2.5 μM)+β-MC (2 μM), as appropriate, was added, and cells were incubated for another 30 minutes. cVA of CLF was then analysed as described above. A schematic representation depicting the mechanism by which TUDC, DBcAMP, and the different signal modulators used in this study would exert modulatory effects on both signal transduction pathways and recycling of canalicular transporters from the subapical compartment (SAC) to the canalicular membrane, is shown in fig 2.

Estimation of the effects of inhibitors on hepatoprotection with DBcAMP in TLC induced cholestasis

Couplets were preincubated with 10 μl of DMSO (vehicle control), 20 μM BAPTA/AM, 1 μM SP, 5 μM KT5720, or BAPTA/AM (20 μM)+KT5720 (5 μM) for 15 minutes. Then, TLC (2.5 μM) alone, or together with DBcAMP (10 μM), was added, and cells were incubated for another 30 minutes. cVA of CLF was then analysed as described above.

Statistical analysis

Results are expressed as mean (SD). Statistical comparisons were made using ANOVA. Differences were considered significant when p<0.05.

RESULTS

TLC induced a significant decrease in cVA of CLF (from 57 (3)% to 25 (7)%, p<0.001). Significant protection from this impairment was given by TUDC (increase from cVA of 25 (7)% to 41 (9)%; p<0.001), β-MC (increase from cVA of 25 (4)% to 42 (5)%; p<0.001), and DBcAMP (increase from cVA of 25 (7)% to 41 (13)%; p<0.001). These represented relative increases in comparison with the unprotected TLC induced level of 64%, 68%, and 64% for TUDC, β-MC, and DBcAMP, respectively.

When both TUDC and DBcAMP were used together, their effect was not greater than for each compound used alone (from cVA of 25 (7)% to 36 (11)%; p<0.01); a 44% increase compared with controls (table 1).

The protective effect of TUDC was significantly impaired by the presence of the Ca²⁺ chelator BAPTA/AM, the PKC/PKA inhibitor H7, and the PKC inhibitor SP. The specific PKA inhibitor KT5720 had no significant effect however on hepatoprotection with TUDC (fig 3).

In the absence of the hepatoprotective compounds, PKC and PKA inhibitors did not have any significant independent effect on cVA of CLF (57 (3)%, 56 (7)%, 55 (3)%, and 58 (10)% for BAPTA/AM, H7, SP, and KT5720, respectively, compared with 61 (7)% in controls; NS). Similarly, in the absence of the hepatoprotective compounds, these signalling modulators did not influence the ability of TLC to impair cVA of CLF in hepatocyte couplets (25 (7)%, 24 (14)%, 22 (5)%, 28 (3)%, and 22 (6)% for TLC+BAPTA/AM, TLC+H7, TLC+SP, and TLC+KT5720, respectively, compared with 25 (7)% in TLC alone; p<0.05) (tables 2, 3).

The protective effect of β-MC was also reduced, although to a lesser extent, by the Ca²⁺ chelator BAPTA/AM, the PKC/PKA inhibitor H7, and the PKC inhibitor SP. Again, the PKA inhibitor KT5720 did not affect the improvement in cVA of CLF caused by β-MC (fig 4).

BAPTA/AM, but not SP or KT5720, significantly impaired the protective effect of DBcAMP. When both BAPTA/AM and KT5720 were administered together, their effects were not different from those of BAPTA alone (fig 5).

DISCUSSION

Stimulation of hepatocellular secretion by UDC and its conjugates via selective activation of different signalling pathways has been proposed to play a key role in their anticholestatic effects,^{2, 20, 22, 43, 44} but the hypothesis has yet to be tested in experimental models of cholestasis. In this study, we examined the putative mediation of three signal transduction cascades—that is, PKA, PKC, and Ca²⁺ dependent signalling pathways—in the hepatoprotective effects of two bile salts, TUDC and β-MC, against TLC induced cholestasis.

The study revealed:

1. TUDC was not unique in preventing TLC induced cholestasis as β-MC, another 7β-hydroxylated bile salt, was as effective as TUDC in exhibiting anticholestatic properties;

2. BAPTA/AM, a chelator of intracellular Ca²⁺, significantly decreased the protective effect of both bile salts, supporting the contention that the mechanisms by which they exert their hepatoprotective effects involve Ca²⁺ mobilisation;

3. TUDC and β-MC induced improvement of fluorescent bile salt accumulation was abolished by the PKC inhibitors SP and H7, but not by the specific PKA inhibitor KT5720, suggesting that activation of PKC is involved in the anticholestatic effect of both 7β-hydroxylated bile salts;

4. DBcAMP also attenuated TLC induced impairment of cVA of CLF, but its beneficial effect was only counteracted by the Ca²⁺ chelator BAPTA/AM and not by the PKA inhibitor KT5720, or the PKC inhibitor SP, indicating that mobilisation of Ca²⁺, but not PKA or PKC activation, is involved in this hepatoprotective action of DBcAMP.

TLC induces cholestasis by inhibiting both bile salt dependent and independent fractions of bile flow.⁴⁵ The mechanism by which TLC impairs bile salt secretion remains unclear. An alteration in the bile salt export pump, the canalicular transporter involved in bile salt secretion into the canalicular space, is likely as this represents the limiting step in overall hepatic bile salts transport.¹² Roma et al have shown that TLC induced cholestasis selectively impairs canalicular transfer of the mrp2 substrate sulphobromophthalein both in vivo and in isolated hepatocytes, without affecting its uptake.⁴⁶

Thus TUDC may stimulate vesicle mediated insertion of canalicular carriers waiting to be delivered into their membrane domain, and thus counteract reversal of this process induced by TLC exposure. Our finding that TUDC exerts its anticholestatic effect by Ca²⁺ dependent PKC activation fits well with this contention. Indeed, PKC activation with the phorbol ester phorbol dibutyrate stimulates exocytosis of previously accumulated HRP, a marker of transcytotic vesicular delivery, in isolated perfused liver.²³ Also, a similar stimulation of exocytic activity occurs by translocation/subsequent activation of PKC from cytosol to membranes in rat pancreatic β cells.^{47, 48} A subcellular fraction enriched in microtubule associated vesicles containing both canalicular export pumps and the transcytotic marker polymeric immunoglobulin A receptor, has been recently characterised; this vesicular pool may represent, or be functionally related to, a subapical vesicle pool involved in delivery of canalicular carriers.⁴⁹ Roelofsen et al showed that phorbol ester induced PKC activation favours recycling (following endocytosis induced by hepatocyte isolation) of canalicular transporters back to the plasma membrane in isolated hepatocytes⁵⁰ This is in line with the finding that PKC stimulates efflux of the mrp2 substrate dinitrophenyl-S-glutathione,⁵¹ and the organic cation tri-n-buthylmethylamonium,⁵² in isolated hepatocytes. Furthermore, phorbol ester stimulated PKC-α translocation, insertion of vesicles into the Golgi apparatus, and increased vesicular transport occurs in HepG2 cells.⁵³ In contrast, inhibition by PKC of spontaneous and DbcAMP stimulated transcytotic targeting of canalicular carriers was observed, both following restoration of couplet polarity post-isolation²¹ and in DbcAMP stimulated targeting of the Cl⁻ /HCO₃⁻ exchanger to the canalicular membrane in isolated liver cells.³¹ Taken together, these results suggest that distinct pathways exist for vesicular trafficking of transporters to the canalicular region which are differentially regulated by signal modulators, with the one stimulated by PKC being activated by TUDCA and β-MC. Whether this anticholestatic action is mediated by PKC activation itself or by other effectors of the PKC activated signalling cascade downstream of PKC remains to be answered. The involvement of PKC mediated activation of mitogen activated protein kinases (MAPK), suggested to mediate TUDC induced stimulation of bile salt secretion, is unlikely as simultaneous inhibition of PKC did not block TUDC induced MAPK activation⁵⁴; this suggests that other factors able to activate MAPKs, such as cytosolic Ca²⁺,⁵⁵ are involved.

Our results show that Ca²⁺ dependency for TUDC and β-MC induced hepatoprotection was indirect, via activation of Ca²⁺ dependent PKC isoforms, presumably PKC-α, activated by TUDC.²² This conclusion arises from the present studies showing that the anticholestatic effects of the hepatoprotective bile salts were fully counteracted by the two PKC inhibitors as well as by the Ca²⁺ chelator BAPTA/AM (see fig 2). However, our finding that DBcAMP exerted its effects exclusively via Ca²⁺ elevation—that is, in a PKA and PKC independent way (see fig 5), support the contention that, in addition to PKC activation, Ca²⁺ elevation per se affords hepatoprotection against TLC induced cholestasis. These apparently discrepant results are not necessarily in contradiction as under conditions of PKC activation (as occurs following TUDC and, presumably, β-MC, administration), many DBcAMP stimulated apically directed vesicular transport processes have been reported to be inhibited, including: increased targeting of canalicular transporters into their canalicular domain^{21, 31}; HRP biliary excretion²⁸; and sorting of sphingolipids to the apical pole.⁵⁶ Ca²⁺ mediated hepatoprotection may therefore not be operating when PKC is simultaneously activated by administration of TUDC or β-MC. Alternatively, Ca²⁺ mediated hepatoprotection may require the coadjuvant effect of other signalling pathways activated by DBcAMP, such as PKA activation, to induce its beneficial effect.

In conclusion, our results show that both TUDC and β-MC can protect disruption of biliary secretory function in hepatocyte couplets. PKC activation seems to play a crucial role in this beneficial effect. DBcAMP, which shares with TUDC some properties as a signal molecule, also exhibits a similar anticholestatic effect. This effect occurs via Ca²⁺ elevation but, in contradistinction, is independent of PKC activation, which is crucial to TUDC and β-MC afforded hepatoprotection in this model. Further studies of the mechanisms underlying the anticholestatic effect of this family of hepatoprotective bile salts should provide an insight into the events that, subsequent to Ca²⁺ elevation and/or PKC activation, mediate their hepatoprotective functions.