ERK1/2 and p38 MAPKs Are Complementarily Involved in Estradiol 17ß-d-Glucuronide-Induced Cholestasis: Crosstalk with cPKC and PI3K

Objective The endogenous, cholestatic metabolite estradiol 17ß-d-glucuronide (E217G) induces endocytic internalization of the canalicular transporters relevant to bile formation, Bsep and Mrp2. We evaluated here whether MAPKs are involved in this effect. Design ERK1/2, JNK1/2, and p38 MAPK activation was assessed by the increase in their phosphorylation status. Hepatocanalicular function was evaluated in isolated rat hepatocyte couplets (IRHCs) by quantifying the apical secretion of fluorescent Bsep and Mrp2 substrates, and in isolated, perfused rat livers (IPRLs), using taurocholate and 2,4-dinitrophenyl-S-glutathione, respectively. Protein kinase participation in E217G-induced secretory failure was assessed by co-administering selective inhibitors. Internalization of Bsep/Mrp2 was assessed by confocal microscopy and image analysis. Results E217G activated all kinds of MAPKs. The PI3K inhibitor wortmannin prevented ERK1/2 activation, whereas the cPKC inhibitor Gö6976 prevented p38 activation, suggesting that ERK1/2 and p38 are downstream of PI3K and cPKC, respectively. The p38 inhibitor SB203580 and the ERK1/2 inhibitor PD98059, but not the JNK1/2 inhibitor SP600125, partially prevented E217G-induced changes in transporter activity and localization in IRHCs. p38 and ERK1/2 co-inhibition resulted in additive protection, suggesting complementary involvement of these MAPKs. In IPRLs, E217G induced endocytosis of canalicular transporters and a rapid and sustained decrease in bile flow and biliary excretion of Bsep/Mrp2 substrates. p38 inhibition prevented this initial decay, and the internalization of Bsep/Mrp2. Contrarily, ERK1/2 inhibition accelerated the recovery of biliary secretion and the canalicular reinsertion of Bsep/Mrp2. Conclusions cPKC/p38 MAPK and PI3K/ERK1/2 signalling pathways participate complementarily in E217G-induced cholestasis, through internalization and sustained intracellular retention of canalicular transporters, respectively.


Introduction
Bile formation is a highly regulated process that depends upon the coordinated transport activity of basolateral and canalicular carriers on the hepatocyte. Complex signalling pathways regulate the density and activity of these carriers, and its imbalance impairs secretory function, leading to retention of biliary components both in liver and blood [1].
Estradiol 17ß-D-glucuronide (E 2 17G) is an endogenous metabolite of 17ß-estradiol that induces an acute, reversible cholestasis in rats [2]. Since its level build up during pregnancy, it has been suggested to be relevant to the pathogenesis of intrahepatic cholestasis in pregnant, susceptible women [2].
The mechanisms involved in E 2 17G induces cholestasis seems to be multifactorial. Trans-inhibition by E 2 17G of Bsep-mediated canalicular transport of bile salts [3] and increase in paracellular permeability leading to dissipation of plasma-to-bile osmotic gradients [4] have been shown to be causal factors. In addition, our group showed that endocytic internalization of both Bsep [5] and Mrp2 [6,7], two canalicular transporters crucial for bile formation, is also a key cholestatic mechanism. In addition, our group also demonstrated that E 2 17G activates both classical (Ca 2+dependent) protein kinase C (cPKC) isoforms [8] and the phosphoinositide 3-kinase (PI3K)-Akt signalling pathway [9], and that these signalling events are cooperatively involved in the changes in Bsep/Mrp2 localization status. Whereas cPKC triggers endocytic internalization of these transporters, PI3K-Akt retains them into intracellular, vesicular compartments [8,9].
Since the biological effects of cPKC and PI3K are often mediated through downstream protein kinases, we speculated that other signalling mediators are involved, provided they can crosstalk with cPKC or PI3K. Likely candidates are the mitogen activated protein kinases (MAPKs). Both cPKC [10,11] and PI3K/Akt [12,13] often act as upstream signalling activators of MAPKs in hepatocytes.
The main vertebrate MAPKs are the extracellular signalregulated kinases 1 and 2 (ERK1/2), c-Jun amino-terminal kinases 1 and 2 (JNK1/2), and p38 kinase (isoforms a and b in hepatocytes). ERK1/2 is preferentially activated by growth factors, while JNK1/2 and p38 are more responsive to stress stimuli [14]. After recognition of these effectors by surface receptors, MAPKs are activated by three-tiered, sequential phosphorylations mediated by small GTP-binding proteins (e.g., Ras, Rap) and two protein kinases (MAPKKK and MAPKK), which act as dual-specificity enzymes that activate a selective MAPK type. Non-canonical activation of MAPKs involves MAPK autophosphorylation, or direct MAPK phosphorylation by alternative protein kinases, such as Src or ZAP70 [15].
These findings have prompted us to study the role for MAPKs in E 2 17G-induced cholestasis. In particular, we ascertained which MAPK types contribute to E 2 17G-induced cholestasis both in isolated rat hepatocyte couplets (IRHCs) and in the isolated, perfused rat liver (IPRL). Our results demonstrate that both p38 and ERK1/2 contribute to the impairment of localization and function of Bsep and Mrp2 induced by E 2 17G, by acting in a complementary manner, and downstream of cPKC and PI3K, respectively.

Ethics Statement
All animals received humane care according to the criteria outlined in the ''Guide for the Care and Use of Laboratory Animals'' Eight Edition (National Academy of Sciences, 2011).
Experimental procedures were carried out according to the local Guideline for the Use of Laboratory Animals (Resolution Nu 6109/012) established by the institutional Bioethical Committee for the Management of Laboratory Animals and approved by the Faculty of Biochemical and Pharmaceutical Sciences of the National University of Rosario. A system of local committeebased regulatory control offers an equivalent level of regulatory control to that exercised by the systems in Canada. The guidelines for the use of laboratory animals has been approved by our Faculty in 2002, based on CCAC (Canadian Council on Animal Care) guidelines documents. The document has been updated since then according to current and emerging issues for the research community and to advances in laboratory animal care.

Animals
Adult female Wistar rats weighing 250-300 g and bred in our animal house as described [24], were used in all studies. Treatment were carried out under urethane anesthesia (1 g/kg intraperitoneally), and maintained thus throughout. When necessary, body temperature was measured with a rectal probe and maintained at 37uC.

Isolation and Culture of IRHCs
To obtain a preparation enriched in IRHCs, livers were perfused according to the two-step collagenase perfusion procedure, and further enriched by centrifugal elutriation [25]. Cell viability, assessed by trypan blue exclusion, was greater than 90%. To allow for restoration of their polarity, IRHCs were plated onto 24-well plastic plates at a density of 0.5610 5 U/mL in L-15 culture medium, and cultured for 5 h.

Function and localization of Bsep and Mrp2 in IRHCs
Functional changes in Bsep and Mrp2 were evaluated by assessing the canalicular vacuolar accumulation (cVA) of CGamF, a fluorescent Bsep substrate [26,27], or glutathione methylfluorescein (GS-MF), a fluorescent Mrp2 substrate derived from CMFDA; this latter compound diffuses passively, and is intracellularly metabolized by esterases and glutathione S-transferases to render GS-MF [28].
After a 20-min equilibration period, the p38 inhibitor SB203580 (250 nM) or the ERK1/2 inhibitor PD98059 (5 mM) or their solvent (DMSO; 370 mL/L) were added to the reservoir. Fifteen min later, a 5-min basal bile sample was collected, followed by administration of E 2 17G (3 mmol/liver; single intraportal injection, over a 1-min period), or its solvent in controls [DMSO/10% BSA in saline (4:96)]. Bile was then collected at 5-min intervals for another 30-min period. Experiments were considered valid only if the initial bile flow was greater than 30 mL/min/kg of body weight. Liver viability was evaluated by monitoring lactate dehydrogenase release into the perfusate outflow [33]; experiments exhibiting activities over 20 U/L were discarded.
DNP-SG content in bile was measured by high-performance liquid chromatography, as described [34]. Biliary bile salt concentration was determined by a modification of the Talalay's method [35].
At the end of the perfusion period, a liver lobe was excised, frozen in isopentane precooled in liquid nitrogen, and stored at 280uC for further immunofluorescence and confocal microscopy analysis of Mrp2 and Bsep intracellular localization. F-actin staining was carried out to demarcate the limits of the canaliculi, as described [5,8]. Liver sections were fixed and stained as described [6,7], followed by overnight incubation with the specific antibodies against Bsep or Mrp2, and 1 h incubation with the appropriate cyanine 2-conjugated secondary antibodies, and Alexa Fluor 568 phalloidin (1:100, 1 h) for F-actin. To ensure comparable staining and image capture performance for the different groups belonging to the same experimental protocol, liver slices were prepared on the same day, mounted on the same glass slide, and subjected to the staining procedure and confocal microscopy analysis simultaneously. Quantification of the degree of Bsep and Mrp2 endocytic internalization was performed on confocal images using ImageJ 1.34 m (National Institutes of Health), as described [5].

Statistical Analysis
Results are expressed as mean 6 standard error of the mean (SEM). Statistical analysis was performed with one-way analysis of variance followed by the Newman-Keuls test. The variances of the densitometric profiles of Bsep and Mrp2 localization were compared with the Mann-Whitney U test. P values,0.05 were considered to be statistically significant. representative Western blottings of phospho (p)-p38, p-ERK1/2, p-JNK1/2 and total forms of all these MAPK types were obtained from whole cellular lysates of primary-cultured rat hepatocytes incubated with E 2 17G (200 mM) for 10 to 60 min, or with E 2 17G (200 mM) for 20 min in cells pretreated with the PI3K inhibitor wortmanin (WM, 100 nM) or with the cPKC inhibitor Gö 6976 (Gö , 1 mM) for 15 min. A (right panel), and B and C panels show phosphorylation status of all MAPK types evaluated (calculated as the p-MAPK to total MAPK ratio for each experimental condition). An arbitrary value of 100 was assigned to the band of highest densitometric intensity in every Western blot before the ratio was calculated. The results are shown as mean 6 SEM (n = 5). *P,0.05 vs. control (cells treated only with DMSO), and # P,0.05 vs. E 2 17G (20 min). doi:10.1371/journal.pone.0049255.g001

E 2 17G activates MAPKs
Western blots of phospho(p)-p38 ( Fig. 1; A, right panel), p-ERK1/2 ( Fig. 1; B) and p-JNK1/2 ( Fig. 1; C) showed that E 2 17G increased the amount of each p-MAPK in a time-dependent manner, with increments becoming apparent as soon as 10 min after E 2 17G administration. Pretreatment with the cPKC inhibitor Gö6976 (1 mM) or the PI3K inhibitor WM (100 nM) selectively prevented the increase in p-p38 and p-ERK1/2, respectively ( Fig. 1), indicating that activation of p38 depends upon cPKC whereas activation p-ERK1/2 depends upon PI3K. Instead, pretreatment with either WM or Gö6976 prevented the increase in JNK2 phosphorylation induced by E 2 17G; although a clear trend towards prevention of JNK1 phosphorylation was also observed, this prevention only achieved statistical significance for Gö6976.

E 2 17G-induced impairment of Bsep and Mrp2 transport function and localization in IRHCs involves p38-and ERK1/2-dependent, additive mechanisms
Functional studies in IRHCs revealed that both the p38 inhibitor SB203580 and the ERK1/2 inhibitor PD98059 significantly prevented E 2 17G-induced impairment in cVA of both the Bsep and the Mrp2 substrates (CGamF and GS-MF, respectively; Fig. 2). Contrarily, the JNK1/2 inhibitor SP600125 was without effect, suggesting that JNK1/2 activation does not play a causal role in E 2 17G-induced cholestasis.
The effect of E 2 17G on Bsep and Mrp2 transport activity was accompanied by changes in the localization status of these transporters (Fig. 3, top panels). In control IRHCs, the carrierassociated fluorescence was localized mainly in the canalicular vacuoles, whereas in the E 2 17G-treated group, there was extensive relocalization of the fluorescence from the canalicular zone to the cellular body, indicating endocytosis of the canalicular carriers. This phenomenon was markedly prevented by either p38 or ERK1/2 inhibition (Fig. 3, top panels). This was confirmed by densitometric analysis, which showed a flatter Bsep and Mrp2 fluorescence profile in E 2 17G-treated IRHC (Fig. 3, lower panels). ERK1/2 or p38 inhibition prevented partial or totally this relocalization, as densitometric curves were statistically different from that of E 2 17G alone (Fig. 3, lower panels).
The preventive effects of PD98059 and SB203580 on CGamF and GS-MF secretory failures were additive in nature (Fig. 2), suggesting that ERK1/2 and p38 act through different but complementary mechanism. However, additivity of effects can only be assumed when recorded at concentrations of the inhibitors producing maximal effects individually. This was actually the case, since the protective effects of each of them remained virtually the same at concentrations 5-time higher than those used here in additivity studies (data not shown).
cPKC-p38 and PI3K-ERK1/2signalling pathways are involved in E 2 17G-induced canalicular secretory failure in an additive manner We have previously demonstrated that the cholestatic effect of E 2 17G is partially prevented by the selective inhibition of cPKC [8] and of PI3K [9], and that these kinases play a partial and complementary role in E 2 17G-induced cholestasis. Since we demonstrated here that cPKC and PI3K are selectively involved in p38 and ERK1/2 activation, respectively (Fig. 1), we tried to reproduce these selective dependencies in the functional field (Fig. 2). As expected, there was a lack of additivity in the protective effects when Gö6976 (cPKC inhibitor) and SB203580 (p38 inhibitor) were added together, and the same holds true when WM (PI3K inhibitor) was added together with PD98059 (ERK1/2 inhibitor). On the other hand, adittivity was observed when crosscombinations of these inhibitors were used (i.e., Gö6976 plus PD98059, or WM plus SB203580). Besides, the pretreatment of IRHCs with these same combinations of inhibitors markedly prevented the E 2 17G-induced internalization of canalicular carriers (Fig. 3, upper panels), resulting in control-like densitometric curves of Bsep/Mrp2 localization (Fig. 3, lower panels). This supports the fact that cPKC-p38 and PI3K-ERK1/2 signalling pathways act in parallel and in a complementary manner to account for E 2 17G cholestatic effects.

p38-ERK1/2 co-inhibition prevents the colocalization of Bsep/Mrp2 with the endosomal protein Rab11a induced by E 2 17G
In E 2 17G cholestasis, canalicular carriers are rapidly endocytosed from the canalicular membrane, and can return back to this membrane, presumably from the apical recycling endosomes (ARE) [1]. To visualize this phenomenon, and its prevention by MAPK inhibitors, we evaluated in IRHCs the colocalization status of canalicular carriers with the endosomal protein Rab11a, a marker of ARE that has been shown to colocalize with apically endocytosed proteins in hepatocytes [36]. Using this approach, no colocalization was observed between Rab11a (red) and Bsep or Mrp2 (green) in controls (Fig. 4). Contrarily, E 2 17G induced colocalization of canalicular carriers with Rab11a (yellow/orange merge). IRHCs pretreated with either SB203580 or PD98059 showed lack of colocalization, as in controls; this indicates that either or both the carrier internalization towards ARE was prevented or the carrier reinsertion from ARE has been accelerated.
p38 is involved in the initial impairment of bile secretory function induced by E 2 17G, whereas ERK1/2 blocks its recovery in the IPRL model The IRHC model revealed the existence of two complementary mechanisms accounting for E 2 17G-induced cholestasis, which differentially depends upon the cPKC/p38 and PI3K/ERK1/2 pathways (see above). The nature of these two cholestatic mechanisms cannot however be ascertained by using this approach. Using the IPRL model, which unlike IRHCs allow for the dissection of mechanisms occurring separately in time, we had shown a differential role for cPKC and PI3K in E 2 17Ginduced cholestasis: whereas cPKC is involved in the initial reduction in bile flow due to transporter endocytosis, PI3K (via Akt) blocks the otherwise spontaneous reinsertion of the endocytosed transporters [9]. We therefore used this model to assess whether p38 and ERK1/2 reproduce the cholestatic mechanisms induced by their respective upstream activators, cPKC and PI3K.
The bolus administration of E 2 17G induced a 61% decrease in bile flow within 10 min, which did not recover throughout the perfusion period (Fig. 5, upper panel). This was accompanied by a decrease in the biliary excretion of the Mrp2 and Bsep substrates DNP-SG and taurocholate, respectively (Fig. 5, middle and lower panels). Whereas the p38 inhibitor SB203580 (250 nM) prevented this initial drop, the ERK1/2 inhibitor PD98059 (5 mM) had little, if any, effect on this event. In contrast, PD98059 accelerated the recovery of both bile flow and biliary excretion of Mrp2 and Bsep substrates from 15 minutes of E 2 17G administration onwards. SB203580 and PD98059, when added alone, did not induce any changes in these parameters (data not shown).
Prevention of the biliary secretory failure induced by E 2 17G by SB203580 and PD98059 was accompanied by the recovery of the normal canalicular localization of Bsep and Mrp2 at the end of the perfusion period (Fig. 6). In control livers, transporter-associated fluorescence was confined to the canalicular space (from 21 mm to +1 mm). In E 2 17G-treated livers, relocalization of intracellular fluorescence associated with both carriers from the canalicular space to the pericanalicular area was apparent, as indicated by the decrease in the fluorescence intensity in the canalicular area together with the increased fluorescence at a greater distance from the canaliculus (P,0.001 vs. controls); this indicates endocytic internalization of the carriers. Whereas SB203580 and PD98059 itself did not induce any changes in transporter localization (data not shown), both inhibitors extensively prevented the internalization of Bsep and Mrp2, as illustrated by a control-like pattern of the Bsep and Mrp2 distribution profiles in livers pretreated with the inhibitors, as compared with the E 2 17G-treated group (P,0.005 vs. E 2 17G; n = 20-50 canaliculi per experimental group, from three independent experiments). This supports our contention that p38 contributes to E 2 17G-induced cholestasis by retrieving canalicular carriers from their membrane domain in a cPKC-dependent manner. On the other hand, ERK1/2 would hinder the spontaneous retargeting of the endocytosed transporters by acting downstream PI3K.

Discussion
This study provides further insights into the intracellular signalling pathways involved in E 2 17G-induced cholestasis.
The activation of these two signalling pathways by E 2 17G accounted for two independent and complementary cholestatic mechanisms; whereas the cPKC-p38 signalling pathway was essential for the acute cholestatic effect induced by E 2 17G, a phenomenon largely attributed to the endocytic internalization of Mrp2/Bsep, PI3K-ERK1/2 promoted the intracellular retention of these transporters during the cholestatic phenomenon (see Fig. 7). A similar complementarity had been demonstrated by us for their respective upstream activators, i.e. cPKC and PI3K/Akt [9].
The mechanism by which p38 mediates endocytosis results elusive because of our poor knowledge on the molecular mechanisms that trigger this process, but several reports in other cell models provide some hints. p38 accelerates clathrin-mediated endocytosis of several ligand receptors, such as epidermal growth factor receptor (EGFR) [42], the opioid receptors m and d [43], and the glutamate receptor [44], and there are several similarities between the internalization process of these receptors and that of Bsep. Like them, Bsep is internalized in a clathrin-dependent manner through a mechanism requiring the adaptor protein AP-2 [45], and AP-2 phosphorylation by p38 is essential for cargo recruitment to clathrin-coated pits, as shown for EGFR endocytosis [46]. Another putative p38 target is the Rab GDP dissociation inhibitor Rab:GDI. This protein stimulates the membranes/cytosol recycling of Rab5, a protein that coordinates formation of clathrin-coated vesicles and their fusion with early endosomes, thus increasing endocytic rate [47]. p38 increases Rab:GDI activity by direct phosphorylation at serine-121 [48]. Whether similar regulations for the endocytic process apply to Bsep, and whether Mrp2 shares with Bsep clathrin-mediated endocytic mechanisms remains to be ascertained. The latter is however probable, since endocytosis of Mrp2 in E 2 17G-induced cholestasis is a microtubule-independent event [7], and the same applies for clathrin-dependent endocytosis [49].
Our results also show that the PI3K/ERK1/2 pathway is involved in the intracellular retention of the endocytosed transporters, which otherwise would rapidly return to the canalicular membrane in a microtubule-dependent manner [7]. The nature of this vesicular trafficking is speculative at this point, but presumably involves the latest, microtubule-dependent exocytic step of the transcytotic route, which occurs via apical recycling endosomes (ARE). This is supported by the facts that: 1) Bsep and Mrp2 colocalize with the ARE marker Rab11a (see Fig. 4), 2) Mrp2 and Bsep colocalize on the same vesicles with both the microtubule motor dynein and polymeric immunoglobulin receptor (pIgR), a surface protein that is transcytosed via ARE from the basolateral to the apical membrane [50], 3) insertion of Bsep back to the canalicular membrane from ARE is a microtubule-dependent process [51], and 4) ARE is interconnected in a microtubule-dependent manner with apical early endosomes (AEE), the first station on the endocytic pathway involved in apical endocytosis [49]. Due to the microtubuledependent nature of the exocytic process, microtubule-based motor proteins are a likely target for ERK1/2 modulation of canalicular transporter exocytosis. In line with this, ERK1/2 impairs binding activity of the microtubule motor kinesin-1 [52]; this protein is enriched in both AEE and ARE in liver (40% and 45% of total kinesin content, respectively) [53], and may therefore play a key role in the trafficking of canalicular transporters from AEE to ARE and in their exocytosis from ARE, which are both microtubule-dependent events [49].
The dependency of E 2 17G-induced cholestasis on both p38 and ERK1/2 results somewhat paradoxical with respect to other reports showing that both MAPK types are involved in choleretic phenomena, such as those induced by the bile salt tauroursodeoxycholate (TUDC) [54,55,56] and cAMP [57]. For example, TUDCA stimulates biliary excretion of the choleretic bile salt taurocholate via an integrin-dependent dual signalling pathway that involves both Ras/Raf/MEK/ERK1/2 and Src/p38 signalling pathways [54,55,56,58]; of note, activation of ERK1/2 by TUDC depends on PI3K [59] but not on PKC [55], in agreement with our results (see Fig. 1). The p38-mediated stimulation of taurocholate excretion by TUDCA was due to an enhanced Bsep trafficking from the Golgi compartment to the canalicular membrane [56], and the further insertion of Bsep into the apical membrane [54], in a microtubule-dependent manner. The reason for the paradoxical pro-cholestatic and pro-choleretic effects of MAPKs is unclear at present, but several possibilities arise. MAPK activation is a highly compartmentalized process, which requires activation and localization in the organelle of scaffold proteins to allow for the regional recruitment of MAPK cascade components [60]. In quiescent hepatocytes, Raf-1 and MEK (MAPKK of ERK) are restricted to early endocytic compartments [61]. Therefore, rapid effects on endosomal trafficking of a MAPK modulating agent like E 2 17G may be limited to this early endosomal pathway, whereas the influence on Golgi/post-Golgi pathway would require further localization of specific MAPK Figure 7. Schematic representation of the signalling events involved in E 2 17G-induced cholestasis by endocytic internalization and further retention of canalicular transporters relevant to bile formation (Bsep, Mrp2). p38, acting downstream of cPKC, triggers endocytic internalization of the apical carriers presumably towards apical early endosomes (AEE), the first intracellular endosomal compartment receiving internalized proteins from the apical membrane, in a microtubule-independent manner (solid arrow). These transporters traffic to, and accumulate into, apical recycling endosomes (ARE), from where they can be retargeted to the apical membrane during the recovery of the cholestatic process, in a microtubule-dependent manner (dashed arrows). Activation of the PI3K/Akt/ERK1/2 signalling pathway halts this latter process, thus explaining the increased colocalization of Bsep/Mrp2 with Rab11a, an ARE marker. This prolongs the cholestatic effect of E 2 17G by impeding the fast, spontaneous retargeting of intracellular transporters that would lead to a rapid recovery from the cholestatic injury. doi:10.1371/journal.pone.0049255.g007 scaffold proteins at this site [62]. In addition, E 2 17G and TUDC may activate different signalling pathways apart from MAPKs that, by operating in concert with MAPKs, result in opposite final effects. For example, TUDC-induced acceleration of the Golgi/ post-Golgi exocytic pathway depends upon dual activation of p38 and novel PKC isoforms [56], and we have shown that E 2 17G does not activate the novel PKC isoform e [8]. Alternatively, in a cholestatic context, TUDC-induced MAPK-dependent choleretic mechanisms may not be operative. Actually, MAPKs do not mediate TUDCA anticholestatic effects in TLC-induced cholestasis [23]. Finally, mediation of choleretic and cholestatic effects by p38 may reflect a differential capability of cholestatic and choleretic compounds to evocate p38 isoforms with opposite effects. Only a and ß p38 isoforms are present in liver, and they have been shown to exert opposite effects in other cell types [63]. Importantly, stimulation of Mrp2 apical translocation by cAMP, which prevents E 2 17G-induced internalization of canalicular carriers [6], involves exclusively p38a [57]. It would be of interest to find out whether p38a activity is decreased by E 2 17G, and whether cAMP reverses this decrease as part of its anticholestatic action.
While further studies will be required to assess the specific molecular mechanisms mediated by ERK1/2 and p38 to impair the localization status of canalicular transporters in estrogeninduced cholestasis, our results strengthen the idea that there is a clear interplay between signalling cascades and intracellular trafficking in this cholestasis, and that both MAPKs are key players in its ethiology. Drugs that inhibit selectively MAPKs have been used in preclinical and clinical studies with reasonable success [64]. As far as our results correlate with clinical cholestatic situations in humans, they should help to envisage new, feasible therapeutic strategies to treat them.