Potential conflict of interest: Nothing to report.
Canalicular bile is modified along bile ducts through reabsorptive and secretory processes regulated by nerves, bile salts, and hormones such as secretin. Secretin stimulates ductular cystic fibrosis transmembrane conductance regulator (CFTR)–dependent Cl− efflux and subsequent biliary HCO3− secretion, possibly via Cl−/HCO3− anion exchange (AE). However, the contribution of secretin to bile regulation in the normal rat, the significance of choleretic bile salts in secretin effects, and the role of Cl−/HCO3− exchange in secretin-stimulated HCO3− secretion all remain unclear. Here, secretin was administered to normal rats with maintained bile acid pool via continuous taurocholate infusion. Bile flow and biliary HCO3− and Cl− excretion were monitored following intrabiliary retrograde fluxes of saline solutions with and without the Cl− channel inhibitor 5-nitro-2-(3-phenylpropylamino)-benzoic acid (NPPB) or the Cl−/HCO3− exchange inhibitor 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS). Secretin increased bile flow and biliary excretion of HCO3− and Cl−. Interestingly, secretin effects were not observed in the absence of taurocholate. Whereas secretin effects were all blocked by intrabiliary NPPB, DIDS only inhibited secretin-induced increases in bile flow and HCO3− excretion but not the increased Cl− excretion, revealing a role of biliary Cl−/HCO3− exchange in secretin-induced, bicarbonate-rich choleresis in normal rats. Finally, small hairpin RNA adenoviral constructs were used to demonstrate the involvement of the Na+-independent anion exchanger 2 (AE2) through gene silencing in normal rat cholangiocytes. AE2 gene silencing caused a marked inhibition of unstimulated and secretin-stimulated Cl−/HCO3− exchange. In conclusion, maintenance of the bile acid pool is crucial for secretin to induce bicarbonate-rich choleresis in the normal rat and that this occurs via a chloride–bicarbonate exchange process consistent with AE2 function. (HEPATOLOGY 2006;43:266–275.)
Secretin is known to induce bicarbonate-rich hydrocholeresis in many animal species.1–7 Its interaction with a G-protein–coupled receptor selectively localized to the epithelial bile duct cells8 results in increased intracellular levels of cyclic adenosine monophosphate (cAMP) [cAMP]i7, 9, 10 and protein kinase A activation.11, 12 Phosphorylation and opening of a cAMP-dependent Cl− channel, the cystic fibrosis transmembrane conductance regulator (CFTR),13 causes Cl− efflux to the ductular lumen. This appears to stimulate an apical Na+-independent Cl−/HCO3− anion exchange (AE),14 with HCO3− efflux and Cl− influx, that is facilitated by the outside to inside transmembrane gradient of Cl− at relatively high intracellular HCO3− concentration.10–12, 15, 16 Several bicarbonate transporters, most of them encoded by the SLC4 and SLC26 gene families,17 have been described to exert AE activity. A decade ago, we localized one of those polypeptides, the SLC4A2 or AE2,18 to the apical membrane in cholangicytes. More recently, AE2 was described to colocalize in cholangiocytes with CFTR and the water channel aquaporin-1 in intracellular vesicles, which co-redistribute to the apical membrane upon both cAMP and secretin stimulation.19 Although AE2 appears to be a good candidate as the effector of the Na+-independent Cl−/HCO3− exchange in cholangiocytes, the possible role of the remaining transporter polypeptides has never been ruled out, and the putative physiological role for AE2 is awaiting definite demonstration.
The ability of secretin to increase bile flow and bicarbonate excretion in the normal rat and the mechanistic role of AE activity in HCO3− excretion remain uncertain. Normal rats express secretin receptors in the biliary epithelium,8 and isolated normal rat bile duct units and cholangiocytes both respond to secretin with a cAMP-dependent Cl− efflux and Cl−/HCO3− exchange activity.10, 11, 13, 15, 16, 20 However, in vivo response to secretin could not be shown in normal rats, although these animals effectively respond to the hormone after induction of bile ductular cell hyperplasia.1, 19, 21–24 In the model of isolated bivascularly perfused normal rat liver (IPRL),25 secretin (via the hepatic artery) had no influence on the net bile flow, but it did increase biliary concentration of bicarbonate.25 This effect was partially blocked by infusion of the CFTR inhibitor 5-nitro-2-(3-phenylpropylamino)-benzoic acid (NPPB), whereas it was unchanged after infusion of the anion exchanger inhibitor 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS). It was therefore suggested that secretin might not act through Cl−/HCO3− exchange in the normal rat, but rather via the CFTR, which would directly regulate bicarbonate efflux.25 However, this possibility is difficult to concur with aforementioned studies of isolated bile duct units and isolated cholangiocytes.10, 11, 15, 16, 20
Our findings in live normal rats indicate that, within the context of a maintained bile acid pool, secretin stimulates bicarbonate-rich choleresis through apical Cl−/HCO3− exchange. Furthermore, we provide in vitro evidence that AE2 is the main effector of Cl−/HCO3− exchange activity in normal rat cholangiocytes.
Direct Biliary Monitorization Model in Infused Normal Rat.
Live rats were cannulated to allow for intravascular and intrabiliary administration of different compounds and bile collection, according to a protocol (Fig. 1) approved by the Animal Care Committee of the University of Navarra. Following midline abdomen incision in normal male Wistar rats (250 g; anesthetized with intraperitoneal Na+-pentobarbital 50 mg/kg bw), the common bile duct was cannulated with a Clay-Adams PE-50 catheter (Parsippany, NJ). A 24G × 19 mm Venisystem cannula (Abbott Laboratories, Abbott Park, IL) was then introduced into the iliac vein for continuous infusion (2 mL/h) with Na+-taurocholate (Sigma-Aldrich, St. Louis, MO) in 0.9% NaCl solution (or with just saline solution). Fifteen minutes after starting taurocholate infusion, intrabiliary solutions (0.2 mL) with either 0.1 mmol/L NPPB (Calbiochem, La Jolla, CA) or 0.5 mmol/L DIDS (Sigma-Aldrich) were given via retrograde fluxes using a 1-mL syringe connected to a 24G × 19 Venisystem cannula (which was within the winder of the bile duct PE-50 catheter). Solutions with inhibitors—initially dissolved in dimethylsulfoxide, with a final 1:200 dilution in saline—or just diluted dimethylsulfoxide as control, were let to stand in the bile duct for 20 minutes. Five minutes after retrograde fluxes, a 0.12-mL bolus of saline solution with or without different doses of the secretin (Bachem, Torrance, CA) was administered via the iliac vein. Most experiments with intrabiliary blockers were performed using a dose of 40 nmol of secretin. Fifteen minutes after secretin infusion, the cannula connected to the syringe was removed from the PE-50 catheter, and bile was collected in 1.5-mL tubes (three 5-min collections followed by two 10-min collections). The biliary concentration of HCO3− and Cl− was determined with a Beckman Synchron CX3 analyzer (Beckman, Albertville, MN). Biliary bile acids were measured with a Randox kit (Crumlin Co., Antrim, UK).
Construction of Adenoviral Vectors for AE2 Gene Silencing.
For rat AE2 gene silencing through RNA interference,26 we selected the sequence GGTGTGGACGAGTACAACG (its AE2 specificity being assessed with the NCBI-BLASTN Program). An adenoviral vector was designed for the intracellular expression of small hairpin RNA (shRNA-1) toward this C-terminal sequence, which is common to all rat and human N-terminal AE2 variants.27, 28 As a negative control, we used an adenoviral vector for intracellular expression of shRNA-2 toward the human-specific AE2 sequence TCATCCTCACAGTGCCGCT (underlined nucleotides are not conserved in the rat). First, shRNA plasmids expressing the shRNA molecules under the H1 promoter were constructed to test their silencing potential in transiently transfected HepG2 cells. Annealing of complementary sets of synthetic oligonucleotides (Table 1) was followed by phosphorylation of resultant small double-stranded DNA fragments and directional ligation into BglII-HindIII digested pSUPER vector (OligoEngine, Seattle, WA). Once their silencing potential was confirmed, shRNA expression stretches were released from the pSUPER constructs with KpnI and NotI and subcloned into pShuttle2 vector using the BD Adeno-X Expression System kit (Clontech, Palo Alto, CA). Prior to subcloning, pShuttle2 was released from its cytomegalovirus promoter (through digestions with MluI and BsaI followed by religation) and from its poly A signal (through further digestions with AflII and EcoRI and religation). Resultant shRNA constructs in pShuttle2 were digested with I-CeuI and PI-SceI to produce shRNA adenoviral constructs (following the BD Adeno-X Expression System protocol), these being amplified in E. coli, purified, and linearized with PacI. HEK293 cells were transfected with linearized recombinant adenoviral constructs (5 μg) using FuGene 6 (Roche, Basel, Switzerland). Adenoviruses were propagated in HEK293 cells, being collected after three consecutive freeze–thaw cycles, and purified in a CsCl isopycnic banding step. Viral solutions were dialyzed for 2 hours against 10 mmol/L Tris (pH 8), being further titered with the Adeno-X Rapid Titer Kit (Clontech). Ad-LacZ (also produced with the BD Adeno-X Expression System kit) was a gift from Dr. Lasarte (CIMA, University of Navarra).
Table 1. Oligonucleotides Used for shRNA Constructs and Quantitation of mRNAs
NOTE. Expression of mRNAs was measured via reverse transcription of total RNA followed by quantitative real-time polymerase chain reaction.
Adenoviral Infection of Normal Rat Cholangiocytes and Assessment of AE2 Gene Silencing.
We isolated normal rat cholangiocytes (NRCs) in our laboratory, essentially as described.29 These cells, which stained for cytokeratin 7 (RCK105; Santa Cruz Biotechnology, Santa Cruz, CA), were routinely grown on the top of rat-tail collagen with enriched DMEM-Ham's F-12 medium.30 NRCs with 10-15 passages were seeded on 12-mm glass coverslips in quiescent medium (DMEM-Ham's F-12 medium with 3% fetal bovine serum and penicillin/streptomycin) to maintain cells in a healthy/nonmitogenic state. To determine the optimal multiplicity of infection of NRCs, different concentrations of Ad-LacZ were tested, staining the cells 2 and 3 days after infection with X-Gal. A multiplicity of infection of 4,000 during 9 hours could infect the NRCs located in the periphery of cell clusters, while a multiplicity of infection of 10,000 during 9 hours was necessary to infect most cells (including those in the center of NRC groups).
Assessment of AE2–messenger RNA (mRNA) silencing was performed in NRCs seeded onto 24-mm collagen-coated glass coverslips (Becton-Dickinson, Bedford, MA) and infected with recombinant adenoviruses (multiplicity of infection 10,000) in 1 mL of quiescent medium for 9 hours. Cells were then maintained in fresh quiescent medium for 2 and 3 days. Isolation of total RNA was followed by reverse transcription and real-time polymerase chain reaction (iCycler iQ Apparatus; Bio-Rad Laboratories, Hercules, CA) using specific primers (Table 1) for rat AE2 complementary DNA and the GAPDH normalizing control.
Measurement of AE Activity in NRCs.
To determine the physiological effects of AE2 gene silencing, infected NRCs were assessed for their AE activity. Cells seeded onto 12-mm glass coverslips were infected for 9 hours with 0.3 mL of quiescent medium with recombinant adenovirus (multiplicity of infection 4,000). AE activity was measured in the periphery of the NRC clusters at days 1 to 3 after infection by assessing the acid–base fluxes induced by established maneuvers.16 Intracellular pH (pHi) was measured via microfluorimetry in cells loaded with 12 μmol/L BCECF-AM (Molecular Probes, Eugene, OR) using Hamamatsu fluorescence detection/analysis equipment. Rates of pHi change (δpHi/δt) were measured in cells perfused with Krebs-Ringer bicarbonate and Krebs-Ringer bicarbonate–propionate buffers. AE measurements were also performed in noninfected cells preincubated with and without the anion exchanger inhibitor DIDS (0.5 mmol/L), as well as in the presence and absence of stimuli—either 50 nmol/L secretin or a stimulation mixture that increases [cAMP]i.16 After AE measurements, cells were fixed with methanol for detection of adenoviral infection rate with the Adeno-X Rapid Titer kit (Clontech). Staining with mouse anti-hexon antibody plus rat anti-mouse antibody (HRP conjugate) indicated that virtually 100% of the peripheral cells analyzed for AE activity were infected.
Results of continuous variables are expressed as the mean ± SD. For statistical analyses, nonparametric paired (Wilcoxon) and unpaired (Mann-Whitney) tests were used.
Effect of Secretin on Bile Flow in Live Normal Rat.
Bolus administration with different doses of secretin to normal rats infused with 20 mmol/L Na+-taurocholate (2 mL/h; protocol chart in Fig. 1) led to dose-dependent increases in bile flow (Fig. 2A). For most experiments, we used 40 nmol of secretin. This caused higher bile flow compared with rats receiving saline only (Figs. 2C, 3A,C), which was diminishing over time. Interestingly, the effect of secretin changed with different doses of infused taurocholate (Fig. 2B), with no effect being observed in rats without taurocholate (Fig. 2B-C). In rats without taurocholate, the biliary bile acid excretion rate was found to be decreased by 35 minutes after bile duct cannulation, dropping from 0.785 ± 0.168 μmol/min (in 5-min bile fraction collected just after bile duct cannulation) to 0.573 ± 0.075 μmol/min (in 5-min experimental fraction collected 35 minutes later; P < .001; n = 5).
To analyze the contribution of anion exchangers to secretin-stimulated bile flow in taurocholate-infused rats, animals were given 0.2 mL intrabiliary solution with 0.5 mmol/L DIDS. The presence of DIDS in the biliary tract for 20 minutes led to 48% inhibition of initial secretin-induced increase in bile flow (Δbile flow) compared with Δbile flow values in secretin-stimulated rats without DIDS (9.2 ± 1.8 vs. 17.6 ± 4.2 μL/min; P < .05) (Fig. 3B, white and black bars). In unstimulated rats, intrabiliary DIDS had no effect on bile flow (Fig. 3A). The contribution of the chloride channel CFTR to secretin-induced Δbile flow was then estimated by using the inhibitor NPPB. Intrabiliary NPPB (0.1 mmol/L) resulted in a pronounced inhibition (90%) of initial secretin-induced Δbile flow compared with rats receiving vehicle only during secretin infusion (1.8 ± 4.5 vs. 17.6 ± 4.2 μL/min; P < .01) (Fig. 3D, gray and black bars). In the absence of secretin, intrabiliary NPPB produced no reduction in bile flow (Fig. 3C).
Effect of Secretin on Biliary Bicarbonate and Chloride Excretion in Live Normal Rat.
In taurocholate-infused rats, biliary bicarbonate excretion was significantly increased by secretin (Figs. 2C, 4A,C). Similar to the effect on the bile flow, the effect of secretin on bicarbonate excretion decreased over time since the opening of the bile duct cannula (Fig. 4). In addition, the effect of secretin on the biliary bicarbonate excretion was dependent on the maintenance of the bile acid pool, no secretin effect being observed in rats without taurocholate infusion (Fig. 2C).
Secretin-induced increase in bicarbonate excretion (ΔHCO3−) was sensitive to DIDS. Thus, intrabiliary DIDS resulted in a 45% reduction in the first secretin-induced ΔHCO3− (0.34 ± 0.02 vs. 0.62 ± 0.04 mEq/min; P < .05) (Fig. 4B, white and black bars). Because of parallel effects of DIDS on secretin-induced Δbile flow and ΔHCO3−, no change occurred in the biliary concentration of bicarbonate. On the other hand, intrabiliary NPPB caused an 85% reduction in the initial secretin-induced ΔHCO3− (0.09 ± 0.17 vs. 0.62 ± 0.04 mEq/min, P < .01) (Fig. 4D, gray and black bars). Similarly to DIDS, NPPB had no effect on the biliary concentration of bicarbonate.
In taurocholate-infused rats, secretin showed a stimulatory effect on the biliary excretion of chloride which matched those on the biliary bicarbonate excretion (Figs. 2C, 5). This effect was dependent on the bile acid pool and was not observed in rats not given taurocholate (Fig. 2C).
Secretin-induced increase in chloride excretion (ΔCl−) was very sensitive to intrabiliary NPPB (0.1 mmol/L), causing 96% reduction in initial values (0.04 ± 0.24 vs. 0.99 ± 0.20 mEq/min in the first 5-min bile collection; P < .01) (Fig. 5E, gray and black bars), but had no effect on biliary chloride concentration. In contrast to the marked NPPB effect on secretin-induced ΔCl−, this parameter was not significantly affected by intrabiliary DIDS in secretin-stimulated rats (Fig. 5C). Because DIDS lowered secretin-induced Δbile flow without inhibiting biliary chloride excretion, the presence of this blocker resulted in increased biliary concentration of chloride (Fig. 5A). Such a DIDS-associated increase, together with the absence of inhibitory effect of DIDS on secretin-induced ΔCl− and the marked blockade of this parameter by NPPB (Fig. 5C,E, respectively), strongly suggests that CFTR activity may account for most of the secretin-stimulated biliary excretion of chloride and minimizes the role of DIDS-sensitive Ca2+-dependent chloride channels in secretin-induced events. On the other hand, the fact that DIDS inhibits secretin-induced ΔHCO3− (Fig. 4B) rather than secretin-induced ΔCl− (Fig. 5C) concurs with the hypothesis that stimulation of bicarbonate excretion by secretin occurs through Cl−/HCO3− anion exchange activity.
Effect of Secretin on the DIDS-Sensitive Cl−/HCO3− Exchange Activity in NRCs.
To further investigate the effect of secretin on Cl−/HCO3− anion exchange, in vitro experiments were performed using NRCs which, as previously reported,16 express DIDS-sensitive Cl−/HCO3− anion exchange (Fig. 6). Thus, the AE activity displayed by these cells upon established maneuvers in basal conditions (i.e., without hormone stimulation) was reduced by 40% in the presence of 0.5 mmol/L DIDS during 20 minutes (P < .001) (Fig. 6). Similar to what has been previously reported in isolated bile duct units,31 the AE activity of NRCs was stimulated by 50 nmol/L secretin (P < .01) (Fig. 6). DIDS caused a 44% reduction in the Cl−/HCO3− exchange stimulated by secretin (P < .001) (Fig. 6). Such a DIDS effect in NRCs (both with and without secretin) is in the range of the inhibition reported for AE2 expressed in Xenopus oocytes when using a similar dose of DIDS for a similar length of time.32
AE2 Gene Silencing Blocks Cl−/HCO3− Exchange Activity in NRCs.
To determine whether the main alleged candidate AE2 is in fact implicated in the secretin-responsive and DIDS-sensitive Cl−/HCO3− exchange activity, experiments of AE2 gene silencing were performed in adenovirus-infected NRCs. Two recombinant adenoviruses were designed and produced for the expression of shRNA: one that targets rat AE2 mRNA (shRNA-1) and another specifically directed at human AE2 mRNA (shRNA-2). These vectors were used to infect cultured NRCs. The levels of rat AE2 mRNA were determined 1, 2, and 3 days after infection via real-time quantitative polymerase chain reaction. At days 1 and 2 after infection, there were no significant differences—although at day 2, the mean values in shRNA-1–infected cells were 35% lower than those in cells infected with the negative control shRNA-2 (Fig. 7). However, at day 3, NRCs infected with shRNA-1 showed an 89% reduction of rat AE2-mRNA levels compared with the control (P < .001) (Fig. 7). AE2-mRNA levels in control shRNA-2–infected NRCs were not significantly higher than those observed in noninfected cells maintained in the same medium for 3 days (data not shown).
To determine the functional impact of AE2 gene silencing on the Cl−/HCO3− exchange activity, adenovirus-infected NRCs were tested under unstimulated conditions and upon stimulation with 50 nmol/L secretin (Fig. 8A). AE2 gene silencing 3 days after infection with shRNA-1 resulted in a strong reduction of AE activity compared with NRCs infected with shRNA-2. The functional silencing involved both unstimulated AE activity (JOH− 1.37 ± 1.86 vs. 10.20 ± 2.31 mmol/L/min; P < .001) (Fig. 8A, lower panel) as well as secretin-stimulated activity (JOH− 1.47 ± 2.09 vs. 14.73 ± 6.56 mmol/L/min; P < .001) (Fig. 8A, lower panel). Nevertheless, none of these functional inhibitions was significant 2 days after shRNA-1 infection (Fig. 8A, upper panel), which is in line with the yet incomplete silencing of AE2-mRNA expression at this time point (cf. previous paragraph and Fig. 7).
Interaction of secretin with its receptor results in increased [cAMP]i. Intracellular cAMP is known to cause stimulation of AE activity in many cell types,33 cholangiocytes included.10, 11, 16 Thus, AE activity of silenced NRCs was also analyzed in the presence of the [cAMP]i stimulation mixture (cf. Fig. 8B). Although this mixture produced a potent stimulation of AE activity in shRNA-2–infected cells, it failed to stimulate NRCs infected with adenovirus shRNA-1. As shown in Fig. 8B, the inhibitory effect was seen 3 days (but not 2 days) after shRNA-1 infection. At day 3, the JOH− value after [cAMP]i stimulation was 1.46 ± 2.40 in AE2-silenced cells, whereas it was 17.35 ± 7.20 mmol/L/min in control cells (P < .001) (Fig. 8B, lower panel). Stimulated AE activity in control shRNA-2–infected NRCs was not significantly higher than that in noninfected cells maintained in the same medium for 3 days (data not shown). Altogether, these findings demonstrate that AE2 is the main effector of the Cl−/HCO3− exchange activity displayed by normal rat cholangiocytes in response to secretin or to a direct accumulation of its second messenger cAMP.
Our in vivo data demonstrate that secretin effects in the normal rat are dependent on the maintenance of the bile acid pool and provide strong evidence for the secretin-stimulated bicarbonate secretion occurring through biliary AE activity. Moreover, our in vitro findings demonstrate that AE2 is the main effector of such an activity in normal rat cholangiocytes.
Our minimally invasive in vivo procedure with direct biliary monitorization preserved the autonomic innervation within the liver,34 which may play an important role in intrahepatic hemodynamic and bile flow regulation.34–38 Although the enterohepatic circulation was interrupted, the bile acid pool was maintained by continuous infusion of Na+-taurocholate at a nonsaturating rate.39 Because previous studies have shown minimal effects of secretin in normal rats,21, 23 in contrast with the significant effects observed in rats with bile ductular cell proliferation,23, 24 high doses of the hormone were tried (Fig. 2A). A dose of 40 nmol (administered in 0.12-mL bolus) was selected for most secretin stimulation experiments and resulted in significant initial increases in bile flow and excretion rates of HCO3− and Cl− (Figs. 2-5). These findings, apparently in conflict with the poor response reported by others,7, 21, 23, 24, 40 might be related not only to the high dose of secretin employed, but also to the maintained bile acid pool. This was considered of relevance because of findings in previous experiments with dogs.5, 6 Taurocholate has been described to stimulate secretin-induced cAMP response and Cl−/HCO3− exchange activity in freshly purified rat cholangiocytes.41 Although there is one study in which the effects of secretin administration and taurocholate infusion were tested in the normal rat, such an infusion was interrupted before secretin administration.23 In the present study, it is remarkable that the effects of secretin (40-nmol bolus) are dependent on the presence of taurocholate infusion and change with the bile acid dose (Fig. 2). Concerning the effective doses of secretin in taurocholate-infused rats, although they are seemingly above the physiological range, it should be noted that the hormone is given as a bolus and that the half-life of exogenously administered secretin in the bloodstream is relatively short (≈5 min) because of rapid degradation and inactivation.40, 42 Thus, when trying to analyze the effects of secretin in the normal rat, our experimental model might be an appropriate alternative to the rat models with bile duct proliferation.
A bivascularly perfused IPRL model was proposed as another alternative to bile duct proliferation models to analyze the in vivo effects of secretin in the normal rat.25 In the IPRL model, secretin infusion via the hepatic artery led to increases in biliary HCO3− concentration and net biliary HCO3− excretion, but not bile flow. The lack of secretin effect on bile flow in this experimental setting is in contrast with our findings and could be related to the interruption of functional innervation in the IPRL model. Interestingly, whereas acetylcholine infusion induced a transient decrease in bile flow associated with portal pressure disturbances, such a decrease was not observed after acetylcholine-secretin coinfusion.25 Moreover, the role of the neurotransmitter is supported by in vitro experiments in which acetylcholine was found to increase the stimulatory effect of secretin on the Cl−/HCO3− exchange activity in isolated bile duct units and cholangiocytes isolated from normal rats.20
In our live rat model, intrabiliary NPPB (an inhibitor of CFTR and other Cl− channels) completely blocked the effects of secretin—not only the increase in Cl− excretion, but also the increases in HCO3− excretion and bile flow. On the other hand, intrabiliary DIDS (a nonspecific inhibitor of Cl−/HCO3− exchange) reduced secretin-induced increases in HCO3− excretion and bile flow, but not Cl− excretion. Furthermore, the values of biliary Cl− concentration paradoxically increased in secretin-stimulated rats with intrabiliary DIDS, as sustained net biliary excretion of Cl− was associated with DIDS inhibition of secretin-stimulated bile flow. Thus, although HCO3− permeability through activated CFTR has been shown in several cell systems,43, 44 our present results with intrabiliary inhibitors indicate that secretin-induced efflux of Cl− and efflux of HCO3− each occur via interrelated but different flux proteins.
Although in the IPRL model DIDS had no inhibitory effect on the secretin-induced increase in biliary HCO3− excretion,25 such a discrepancy with our findings may be related to differences in the route of administration (intrabiliary in our model and intravascular in IPRL). Intravascularly administered DIDS (an impermeant drug that does not cross cell membranes)45 may have difficulty reaching the duct lumen to interact with the apical Cl−/HCO3− exchanger. This possibility is not ruled out by the observed inhibition of acetylcholine-stimulated increase in biliary HCO3− excretion by intravascular DIDS.25 In a context of possible acetylcholine-induced vascular disturbances in the liver, impermeant DIDS may manage to reach bile duct lumen and interact extracellularly with different carriers such as Ca2+-activated Cl− channels,46 anion exchangers, and other HCO3− transporters. Alternatively, DIDS may inhibit basolateral factors presumably involved in acetylcholine signaling (which differs from secretin signaling).
Although the role of AE2 protein as an effector of biliary Cl−/HCO3− exchange was previously assumed,16, 18, 47–49 this had yet to be demonstrated. Recent in vivo and in vitro data on colocalization and secretin-responsive apical co-redistribution of flux proteins in bile duct cells have given strong support to such a role.19 In our live rat model, intrabiliary exposure to 0.5 mmol/L DIDS for 20 minutes resulted in 45% inhibition of secretin-induced increase in HCO3− excretion, which is in the range of DIDS inhibition reported for AE2 expressed in Xenopus oocytes under similar dosage and time.32 Likewise, DIDS blocked 40%-44% of the Cl−/HCO3− exchange activity in our NRCs, which was also in the range expected for AE2.
To ascertain the role of AE2 in the AE activity displayed by rat cholangiocytes, gene silencing experiments were performed by infecting NRCs with the adenovirus shRNA-1. Three days after infection, a specific and highly efficient silencing (89%) of AE2-mRNA expression was observed in parallel with a blockade (87%-90%) of unstimulated and secretin-stimulated AE activity. Activity silencing was profound, as it could not be overridden by increasing intracellular levels of cAMP. These findings clearly show that AE2 is pivotal in the Cl−/HCO3− exchange activity displayed by these cholangiocytes, both under basal and secretin-stimulated conditions. Nevertheless, extrapolation of results obtained with cultured cholangiocytes to the in vivo physiology has obvious limits that circumscribe the conclusions to suggestions rather than definite demonstrations.
In conclusion, our present data fully concur with the most widely assumed hypothesis on the mechanisms involved in secretin-induced bicarbonate-rich hydrocholeresis and reveal the importance of a maintained bile acid pool for this to occur in normal rat. Increased intracellular cAMP levels after secretin–secretin receptor interaction7, 9, 10 may lead to apical relocation of vesicles with CFTR, AE2, and aquaporin-1 flux proteins.19 Subsequently, increased cAMP activates the opening of CFTR and efflux of Cl− to the duct lumen,13, 50 which in turn is exchanged with HCO3− through AE2. Resultant increased osmotic forces and relocated aquaporin-1 might further facilitate the passive movement of water and net hydrocholeresis.
We thank J. Pérez-Vizcaíno, L. Martínez, and J. Salas for valuable help, and J.J. Lasarte and R. Agami for aiding with silencing vectors.