Regulation of Ca²⁺ Signaling in Rat Bile Duct Epithelia by Inositol 1,4,5-Trisphosphate Receptor Isoforms

Abstract

Cytosolic Ca²⁺ (${\mathbf{Ca}}_{\mathbf{i}}^{2+}$) regulates secretion of bicarbonate and other ions in the cholangiocyte. In other cell types, this second messenger acts through Ca²⁺ waves, Ca²⁺ oscillations, and other subcellular Ca²⁺ signaling patterns, but little is known about the subcellular organization of Ca²⁺ signaling in cholangiocytes. Therefore, we examined Ca²⁺ signaling and the subcellular distribution of Ca²⁺ release channels in cholangiocytes and in a model cholangiocyte cell line. The expression and subcellular distribution of inositol 1,4,5-trisphosphate (InsP₃) receptor (InsP₃R) isoforms and the ryanodine receptor (RyR) were determined in cholangiocytes from normal rat liver and in the normal rat cholangiocyte (NRC) polarized bile duct cell line. Subcellular Ca²⁺ signaling in cholangiocytes was examined by confocal microscopy. All 3 InsP₃R isoforms were expressed in cholangiocytes, whereas RyR was not expressed. The type III InsP₃R was the most heavily expressed isoform at the protein level and was concentrated apically, whereas the type I and type II isoforms were expressed more uniformly. The type III InsP₃R was expressed even more heavily in NRC cells but was concentrated apically in these cells as well. Adenosine triphosphate (ATP), which increases Ca²⁺ via InsP₃ in cholangiocytes, induced Ca²⁺ oscillations in both cholangiocytes and NRC cells. Acetylcholine (ACh) induced apical-to-basal Ca²⁺ waves. In conclusion, Ca²⁺ signaling in cholangiocytes occurs as polarized Ca²⁺ waves that begin in the region of the type III InsP₃R. Differential subcellular localization of InsP₃R isoforms may be an important molecular mechanism for the formation of Ca²⁺ waves and oscillations in cholangiocytes. Because ${\mathbf{Ca}}_{\mathbf{i}}^{2+}$ is in part responsible for regulating ductular secretion, these findings also may have implications for the molecular basis of cholestatic disorders.

Cytosolic Ca²⁺ (${\mathrm{Ca}}_{\mathrm{i}}^{2+}$) regulates a wide range of functions in the liver. One way in which ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ simultaneously regulates multiple cell functions is through spatial and temporal ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling patterns, such as ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves and oscillations. For example, localized subplasmalemmal increases in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ direct secretion,^1-3 whereas ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ gradients direct cell motion.^4-7 Temporal ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling patterns such as oscillations also can regulate cell functions, including gene expression^8,9 and secretion.¹⁰ Although ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ oscillations can be induced in cholangiocytes,¹¹ it is unknown whether ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves or other subcellular ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling patterns also occur in this cell type.

The formation of ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves in polarized epithelia is believed to depend largely on intracellular Ca²⁺ stores rather than plasma membrane Ca²⁺ channels, because neither the orientation nor the speed of such waves is altered in Ca²⁺-free media.¹² The 2 intracellular Ca²⁺ channels principally responsible for ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling are the inositol 1,4,5-trisphosphate (InsP₃) receptor (InsP₃R) and the ryanodine receptor (RyR). It was previously believed that RyRs regulate ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling only in myocytes, whereas InsP₃Rs regulate ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling in nonexcitable epithelia. However, many cell types express both RyR and InsP₃R,^13,14 including polarized epithelia.^15,16 Moreover, there are multiple isoforms of both of these receptors, each of which displays distinct functional properties.^17-20 Thus, the formation of ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves likely depends not only on the expression and subcellular distribution of RyRs and InsP₃Rs, but on the distribution of individual isoforms of each of these receptors as well. Here we examine the expression and subcellular distribution of these Ca²⁺ release channels in cholangiocytes from rat liver and in normal rat cholangiocyte (NRC) cells, a polarized rat cholangiocyte cell line used as a model for cholangiocyte function.²¹

Materials and Methods

Animals and Materials

Male Sprague-Dawley rats (250-300 g; Camm Research Lab Animals, Wayne, NJ) were used for all animal studies. Acetylcholine (ACh), adenosine triphosphate (ATP), propidium iodide, deoxyribonuclease, hyaluronidase, bovine serum albumin, penicillin-streptomycin, and insulin were obtained from Sigma Chemical Co. (St. Louis, MO). Fluo-4 and fura-2 in acetoxymethyl ester form, Pluronic F-127, and rhodamine-conjugated phalloidin were obtained from Molecular Probes (Eugene, OR). SuperScript II ribonuclease H⁻ reverse transcriptase, colcemid, minimal essential medium (MEM), α-MEM, Liebowitz 15 (L-15), gentamicin, and other tissue culture reagents were from Gibco BRL (Life Technologies AG, Basel, Switzerland). Collagen-coated flasks were from Becton Dickinson, Bedford, MA. RQ1 ribonuclease-free deoxyribonuclease was from Promega (Wallisellen, Switzerland). Pronase was from Calbiochem (La Jolla, CA).

Antibodies

Each InsP₃R isoform was labeled using isoform-specific antibodies. Type I InsP₃R antibodies were from affinity-purified specific rabbit polyclonal antiserum directed against the 19 C-terminal residues of the mouse type I InsP₃R²² and were produced by Research Genetics (Huntsville, AL). Type II InsP₃R antibodies were from affinity-purified specific rabbit polyclonal antiserum directed against the 18 C-terminal residues of the rat type II InsP₃R²³ and were kindly provided by Richard Wojcikiewicz (SUNY, Syracuse, NY). A commercially available monoclonal antibody was used to label the N-terminal region of the human type III InsP₃R¹⁷ (Transduction Laboratories, Lexington, KY).

Cell Isolation and Cell Culture

Cholangiocytes and isolated rat bile duct units were isolated from normal rat liver as described previously.¹¹ Briefly, rats were anesthetized with pentobarbital sodium (50 mg/kg body wt intraperitoneally) and then their livers were perfused with Ca²⁺- and Mg²⁺-free Hank's buffer/0.019% ethylenediaminetetraacetic acid followed by Hank's buffer containing 0.05% collagenase (Boehringer Mannheim Biochemicals, Indianapolis, IN). The portal tissue residue was then either cut into strips for ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ measurements or processed further to obtain isolated cholangiocytes. For ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ studies, the long wavelength lipophilic dye 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine perchlorate (DiD; Molecular Probes) was first injected into the common bile duct to facilitate identification of cholangiocytes within portal strips (see below). To isolate cholangiocytes, the tissue was finely minced, digested further, and then filtered through nylon screens (Tetko, Lancaster, NY). The suspension was subjected to centrifugation in a Percoll density gradient.¹¹ Cells banding at densities 1.060 to 1.075 were collected, elutriated, and then used for Western blot analysis. This resulted in more than 10⁷ cells per liver, ~70% of which stained positive for γ-glutamyl transpeptidase and more than 90% of which were viable by trypan blue exclusion.

The NRC cell line was maintained in culture for separate studies. This polarized cell line is derived from normal rat cholangiocytes.²¹ Cells were maintained in culture on collagen in cholangiocyte growth medium.²¹

Immunoblot Analysis

Protein concentration of cell homogenate was determined as described.²⁴ Cholangiocyte or NRC cell proteins were separated by electrophoresis using a 5% polyacrylamide gel and transferred to nitrocellulose membranes. The membranes were blocked at 4°C overnight and probed for 2 hours with the primary antibody. The antibody against InsP₃R isoform I was used at a dilution of 1:60, the antibody against InsP₃R isoform II was used at a dilution of 1:100, and the antibody against InsP₃R isoform III was used at a dilution of 1:1,000. Peroxidase-conjugated secondary antibodies were from Pierce (Rockford, IL). The membranes were washed, incubated for 1 hour with peroxidase-conjugated secondary antibody immunoglobulin G anti-rabbit (1: 1,000) or anti-mouse (1:3,000), and shown by enhanced chemiluminescence (Amersham, Arlington Heights, IL). For quantitative analyses, immunoblotting was performed on samples of rat cholangiocytes or NRC cells along with positive controls containing known concentrations of InsP₃R isoforms.²³ For each immunoblot, type I, II, and III InsP₃R immunoreactivity was then quantified using a BioRad GS-700 imaging densitometer (Richmond, CA). InsP₃R isoform protein expression in cholangiocytes and controls was compared directly to yield the relative abundance of each isoform in cholangiocytes.²³ Specifically, the ratio of cholangiocyte or NRC cell band intensity divided by control cell band intensity was multiplied by the InsP₃R isoform concentration in the control cells or tissue to obtain the InsP₃R isoform concentration in cholangiocytes or NRC cells. Cerebellum was used as a standard for type I InsP₃R (129.6 ng/10 μg protein), whole liver was used for type II InsP₃R (0.65 ng/10 μg protein), and the RIN cell line was used for type III InsP₃R (14.1 ng/10 μg protein).²³ The relationship between band intensity and amount of InsP₃R loaded onto gels was linear over the range seen in the present study.

Confocal Immunofluorescence Microscopy

Immunochemistry was performed on 4-μm-thick frozen sections of rat liver hilum containing bile ducts or on NRC cells grown on coverslips. Specimens were fixed by cold acetone, methanol, or 3% formaldehyde followed by methanol permeabilization. After blocking steps, specimens were labeled with primary antibody, rinsed with phosphate-buffered saline, and incubated with Alexa 488 – or Alexa 568 – conjugated secondary antibody (Molecular Probes). Bile ducts also were stained with rhoda-mine-conjugated phalloidin to label actin, which is most concentrated beneath the plasma membrane,^12,15,25 or with propidium iodide to stain the nucleus.²⁶ For negative control studies, tissue was incubated with secondary antibodies but anti-InsP₃R (primary) antibodies were omitted. Specimens were examined using a Zeiss LSM 510 confocal microscope equipped with a krypton/argon laser (Thornwood, NY). To ensure specificity of staining, images were obtained using confocal machine settings that detected no fluorescence in negative controls. NRC cells were examined using either a Zeiss Axioskop microscope or a BioRad Microradiance AG-2 confocal system.

RNA Isolation and Reverse Transcription

Total RNA was extracted as described.²⁷ The concentration of RNA collected from the cholangiocytes or NRC cells was determined by absorbance at 260 nm, and the quality was controlled by running an aliquot on a 1% agarose form-aldehyde gel.

Polymerase Chain Reaction Analyses

Two types of polymerase chain reaction (PCR) analyses were performed. Reverse transcription (RT) PCR was performed to determine whether cholangiocytes or NRC cells express RyR, whereas real-time quantitative PCR was performed to examine relative amounts of InsP₃R isoforms in cholangiocytes and NRC cells. For RT-PCR, 5 μg of total RNA was treated with RQ1 ribonuclease-free deoxyribonuclease and reverse transcribed by the SuperScript II ribonuclease H⁻ reverse transcriptase with an oligo(dT) primer. Two control complementary DNA reactions were performed; RNA but no reverse transcriptase was added (RNA control) in one, and no RNA was added (DNA control) in the other. PCR amplification was performed using a PTC-100 automated thermocycler (MJ Research, Inc., Watertown, MA) using 2 μL of the first strand complementary DNA reaction, 150 pmol of each degenerate primer, 50 μmol/L deoxynucleoside triphosphates, 2.5 U AmpliTaq DNA polymerase in 10 mmol/L Tris-HCl (pH 8.3) containing 50 mmol/L KCl, and 2.5 mmol/L MgCl₂ in a total volume of 100 μL. Following hot start (2 minutes at 94°C), the samples were subjected to 30 cycles of 45 seconds at 94°C, 1 minute at 50°C, and 1 minute at 72°C. This was followed by a final extension at 72°C for 10 minutes. The degenerate primers were designed to amplify a 530 – base pair product from the 3′ region of all 3 known RyRs.¹⁵ In selected experiments, primers were designed to probe for each of the 3 RyR isoforms. For those studies, TCAAGCGGAAGGTTCTGGA was the sequence used for rRyR1-20-5′, TGACGGATGAAAGGATGGTG was used for rRyR1-352-3′, CATGGATAAGGTCGCACTGGA was used for rRyR2-75-5′, CGAGCTGTTTGCCATTATGG was used for rRyR2-364-3′, TTGGACAAAAATGCCCTGGA was used for rRyR3-88-5′, and GGTCTTAAAGCCCATGGCAATA was used for rRyR3-324-3′. Each PCR product was analyzed by agarose gel electrophoresis.

Real-time quantitative PCR analysis of InsP₃R iso-forms and glyceraldehyde-3-phosphate dehydrogenase was performed with a PE Applied Biosystems 7700 Sequence Detector (Perkin Elmer, Shelton, CT) as described previously.²⁸ Complementary DNA was amplified in a 50 μL volume containing 25 μL of the 2× TaqMan Universal PCR Master Mix (Perkin Elmer), 100 nmol/L probe, and 300 nmol/L of each primer. After a denaturing step of 10 minutes at 95°C, 40 cycles were performed: 95°C for 15 seconds and 60°C for 1 minute. The mathematical analysis of the results was performed as recommended by the manufacturer.

${\mathit{Ca}}_i^{2+}$ Measurements

Confocal microscopy was used to measure ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ in cholangiocytes because cells were within thick segments of the bile duct or in isolated bile duct units. Epifluorescence microscopy was used to monitor ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ in NRC cell monolayers. For confocal imaging, DiD (25 μmol/L) and the Ca²⁺ dye fluo-4/acetoxymethyl ester (18 μmol/L) were coinjected into the common bile duct, and then bile duct segments were isolated as previously described. The bile duct segments were observed using a BioRad MRC 1024 confocal imaging system. Tissue was excited at 647 nm and observed at more than 680 nm to identify DiD-labeled cholangiocytes, and then ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ was monitored by exciting the specimen at 488 nm and detecting emission signals greater than 515 nm. Images were obtained at a rate of 1 image per second to identify hormone-responsive cells,²⁶ and then confocal line scanning^29,30 was performed to identify the speed and direction of ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves. Cells were observed using a 20×, 0.75 NA objective (zoom factor, 5), resulting in a spatial resolution of 0.20 μm/pixel and a temporal resolution of 6 milliseconds. Increases in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ were expressed as percent increase in fluo-4 fluorescence intensity.^11,26,29 Velocities of ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves were determined from the rate at which fluorescence increases moved along the scan line.^29,30

For NRC cell studies, cells were loaded with fura-2 (5 μmol/L) and then observed using a Zeiss Axiovert 135 inverted microscope. Measurements of ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ were performed using a Zeiss Attofluor imaging system. Cells were excited at 334 and 380 nm, and emission signals greater than 520 nm were observed with an intensified CCD camera. ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ was calculated from the excitation ratio R according to the following equation³¹:

$${{\mathrm{Ca}}_{\mathrm{i}}}^{2+}=(\mathrm{R}-{\mathrm{R}}_{\min })∕({\mathrm{R}}_{\max }-\mathrm{R})\times {\mathrm{K}}_{\mathrm{d}}\times (\mathrm{Sf}2∕\mathrm{Sb}2).$$

Calibrations were performed using standards containing Ca²⁺-saturated and Ca²⁺-free fura-2, respectively. This yielded R_max, R_min, Sf2, and Sb2. A value of 224 nmol/L was used for K_d. NRC cells were measured before confluence, when the cells were present in large (>20 cells) islands. The traces shown represent values from regions of interest covering approximately one cell each.

Statistics

Results are expressed as mean ± SD. Comparisons between groups were made using Student's t test, and a P value less than .05 indicates a significant difference.

Results

Expression of InsP₃Rs and RyR in Cholangiocytes and NRC Cells

Expression of all 3 InsP₃R isoforms was detected in cholangiocytes by quantitative measurement of messenger RNA (mRNA). Using this approach, 40.8% of InsP₃R mRNA was for the type I isoform, 38.0% was for type II, and 21.2% was for type III (Fig. 1). Expression of all 3 InsP₃R isoforms was also detected in NRC cells by this approach. The relative mRNA distribution of the InsP₃R in NRC cells was 20% for the type I isoform, 13% for type II, and 67% for type III (Fig. 2). This finding is consistent with the observation that expression of the type III InsP₃R is increased in cell lines relative to what is observed in the corresponding native tissue.³² In contrast to the InsP₃R, RyR expression could not be detected in cholangiocyte mRNA using RT-PCR (Fig. 3). RT-PCR was also performed with NRC cell mRNA using primers specific for each of the 3 known isoforms of the RyR (Fig. 4). RyR-2 and RyR-3 were absent from NRC cells. A faint band for RyR-1 was detected after 35 cycles, suggesting that this isoform also is not expressed or is expressed only minimally. Together, these findings suggest that the InsP₃R is the principal intracellular Ca²⁺ release channel in cholangiocytes as well as NRC cells.

Fig. 1.

InsP₃R isoforms in cholangiocytes measured by real-time quantitative PCR using primers and probes specific for each isoform. Each of the 3 InsP₃R isoforms was expressed; 40.8% of the InsP₃ receptor mRNA was for the type I isoform, 38.0% was for type II, and 21.2% was for type III.

Fig. 2.

InsP₃R isoforms in NRC cells measured by real-time quantitative PCR. Examination of the relative mRNA distribution for each InsP₃ receptor isoform showed that the type III isoform is most abundant in NRC cells, representing 67%. Isoforms I and II represented 20% and 13%, respectively.

Fig. 3.

RT-PCR fails to detect RyR in cholangiocytes. Using isoformspecific primers, a single 530 – base pair product common to all known RyR isoforms was identified in cardiac RNA (positive control) but not in cholangiocyte RNA or in negative controls. (A) Cardiac RNA after (1) 20, (2) 25, or (3) 30 PCR cycles. (B) Cholangiocyte mRNA after (1) 20, (2) 25, or (3) 30 PCR cycles. (C) Control reactions: (1) DNA control, (2) RNA control, and (3) actin band (500 base pairs), showing that the cholangiocyte RNA is intact.

Fig. 4.

RT-PCR assessment of RyR in NRC cells. Using isoform-specific primers, single products were identified in the positive controls, rat skeletal muscle (RyR-1), rat heart (RyR-2), and rat diaphragm (RyR-3). A faint RyR-1 band was seen in NRC cells after 35 PCR cycles, and no band was seen in NRC cells for RyR-2 or RyR-3. A single band is seen for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in NRC cells, showing that the NRC cell RNA is intact.

Immunoblots were performed to determine whether cholangiocytes express all 3 InsP₃R isoforms at the protein level (Fig. 5). Determination of the relative protein level of each isoform is complicated by the fact that each iso-form-specific antibody has a different affinity for its epitopes. Therefore, to estimate relative protein levels, quantitative immunoblots for each isoform were performed along with blots from tissues that express known concentrations of each InsP₃R isoform.²³ Densitometric quantification of immunoreactivity that comigrated with each positive control showed that cholangiocytes expressed 1.37 ± 0.17 ng of the type III InsP₃R per 10 μg cell protein (mean ± SD) but only 0.12 ± 0.12 ng/10 μg protein of type I and only 0.15 ± 0.07 ng/10 μg protein of type II (Fig. 5D). Thus, most InsP₃Rs in cholangiocytes (84%) are the type III isoform, with lesser amounts of types I (7%) and II (9%). This confirms that all 3 InsP₃R isoforms are expressed in cholangiocytes; however, in contrast to the findings of our PCR studies, these results suggest that the type III InsP₃R is the predominant isoform at the protein level. Immunoblots were also performed to determine whether NRC cells express all 3 InsP₃R isoforms at the protein level. NRC cells expressed 7.6 ± 0.5 ng/10 μg protein of the type III InsP₃R, but neither the type I nor type II isoform could be detected (Fig. 6). This suggests that, at the protein level, nearly all InsP₃Rs in NRC cells are the type III isoform.

Fig. 5.

Cholangiocytes express all 3 InsP₃R isoforms. (A) Western analysis using a type I–specific InsP₃R antibody identifies a single band of the same size in lysates from cholangiocytes (60 μg) and the positive control, rat cerebellum (8.6 μg). (B) Type II–specific InsP₃R antibody CT2 identifies a single band of the same size in lysates from cholangiocytes and the positive control, rat liver (60 μg each). (C) A monoclonal type III–specific InsP₃R antibody identifies a single band of the same size in lysates from cholangiocytes and the positive control, RIN-5F cells (60 μg each). (D) Quantitative densitometric analysis of InsP₃R isoforms in cholangiocytes. Expression of the type III isoform is significantly greater than either of the other 2 isoforms (*P < .001). Blots were quantified in their linear range, and results are the mean ± SD of n = 5 blots for InsP₃R-1, n = 4 for InsP₃R-2, and n = 3 for InsP₃R-3.

Fig. 6.

NRC cells express only the type III InsP₃ receptor. (A) A type I–specific InsP₃R antibody identifies a single band in lysates from the positive control, rat cerebellum (8.6 μg), but not in lysates from NRC cells (40 μg). (B) Type II–specific InsP₃R antibody CT2 identifies a single band in lysates from the positive control, rat liver (60 μg each), but not in NRC cell lysates (40 μg). (C) A monoclonal type III–specific InsP₃R antibody identifies a single band of the same size in lysates from NRC cells (40 μg) and the positive control, RIN-5F cells (60 μg). (D) Densitometric analysis of InsP₃R isoforms in NRC cells. Expression of the type III isoform is greater than the amount observed in cholangiocytes. Blots were quantified in their linear range, and results are the mean ± SD of triplicate blots for each isoform (*P < .001).

Subcellular Localization of InsP₃R Isoforms

The subcellular distribution of the type I, II, and III InsP₃Rs was investigated in cholangiocytes and NRC cells by confocal immunofluorescence histochemistry. Rat liver sections were labeled with isoform-specific InsP₃R antibodies and colabeled with rhodamine-conjugated phalloidin to identify the actin network beneath apical and basolateral membrane or propidium iodide to identify the nucleus of cholangiocytes (Fig. 7). Type I and II InsP₃Rs were uniformly distributed throughout the cytosol (Fig. 7A and B). Although type III InsP₃R immunofluorescence was also seen throughout the cytosol, labeling was most intense in the apical region (Fig. 7D and E). No labeling was seen in negative control tissues stained with secondary but not primary antibodies (Fig. 7C and F). The mean apical and basolateral fluorescence intensity of each isoform was determined to quantify these observations (Fig. 7G). The ratio of apical/basolateral fluorescence was 1.10 ± 0.04 for the type I InsP₃R and 1.14 ± 0.04 for type II but 2.63 ± 0.31 for type III (P < .0002 relative to the type I and II isoforms; n = 8 measurements for each group). This suggests that the subcellular pattern of distribution of the type III InsP₃R differs from that of the type I and II InsP₃Rs in cholangiocytes. Confocal immunofluorescence was also used to examine the subcellular distribution of the type III InsP₃R in NRC cells. Optical sections obtained near the apical membrane showed a diffuse pattern of InsP₃R labeling (Fig. 8A-C), whereas sections obtained in the basolateral region showed minimal labeling (Fig. 8D-F). As with cholangiocytes, NRC cells thus express the type III InsP₃R to a greater extent in the apical region. Thus, cholangiocytes and NRC cells share certain key features in the expression and subcellular distribution of InsP₃R isoforms, although there are differences as well.

Fig. 7.

Subcellular distributions of InsP₃R isoforms in cholangiocytes as determined by confocal immunofluorescence labeling of rat liver. (A) Distribution of the type I InsP₃R (green). The specimen is colabeled with rhodamine-phalloidin (red) to identify actin, which outlines the plasma membranes of cholangiocytes. InsP₃R labeling is seen throughout each cholangiocyte. (Scale bar, 10 μm.) (B) Distribution of the type II InsP₃R (green) plus rhodamine-phalloidin (red) to identify actin, which is predominantly along the plasma membrane. Labeling of this isoform is also seen throughout each cholangiocyte. (C) Double-labeled image of a separate liver section stained with rhodamine-phalloidin (red) plus the anti-rabbit secondary antibody used to counterstain the primary antibodies for the type I and type II InsP₃R (green). The lack of nonspecific staining by the secondary antibody is evident. (D) Distribution of the type III InsP₃R(green) plus rhodamine-phalloidin (red) to identify actin, which is near the plasma membrane. Labeling of this isoform is diffusely distributed throughout the cytosol but is most intense along the apical membrane of the cholangiocyte. (E) Distribution of the type III InsP₃R (green) plus propidium iodide (red) to identify the nucleus. This double-labeled image also shows that the type III InsP₃R is most concentrated near the apical membrane. (F) Double-labeled image of a separate liver section stained with propidium iodide (red) plus the anti-mouse secondary antibody used to counterstain the primary antibody for the type III InsP₃R (green). The lack of nonspecific staining by this secondary antibody is evident as well. (G) Quantification of InsP₃R immunofluorescence. For each isoform, apical and basolateral immunofluorescence was quantified and the ratio was obtained. The apical and basolateral regions are labeled to a similar extent for both the type I and type II InsP₃R, but fluorescence labeling is more than twice as intense in the apical region than in the basolateral region for the type III receptor (*P < .0002).

Fig. 8.

Subcellular distribution of the type III InsP₃R in NRC cells as determined by confocal immunofluorescence. Optical sections were obtained in NRC monolayers double labeled with antibodies directed against the type III InsP₃R (green) and the tight junction protein ZO-1 (red), which identifies the apical region. Type III InsP₃R staining is both intense and diffusely distributed in optical sections near (A-C) the apical membrane but is minimal in optical sections near (D-F) the basolateral membrane.

Ca²⁺ Signaling in Cholangiocytes and NRC Cells

To examine ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling, cells were stimulated with ATP, which is known to increase ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ through activation of InsP₃Rs in cholangiocytes.¹¹ NRC cells were exposed to ATP apically, because only the apical side of these cells is exposed and because both cholangiocytes³³ and NRC cells³⁴ express apical P2Y receptors that link to InsP₃ formation. Apical exposure of NRC cells to extracellular ATP (1-100 μmol/L) elicited an immediate increase in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ followed by ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ oscillations (Fig. 9A and B). These oscillations terminated quickly on removal of the agonist. For comparison, the effect of ATP on ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling was examined in individual cholangiocytes within isolated bile duct units (Fig. 9C). ATP induced ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ oscillations in cholangiocytes, similar to what has been observed in cholangiocytes previously¹¹ and in NRC cells here. Because of technical constraints, the current studies compared ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals induced by stimulation of apical ATP receptors in NRC cells with basolateral ATP receptors in cholangiocytes. Although the response of cholangiocytes to ATP is attenuated when the cells are stimulated basolaterally rather than apically, cholangiocytes nonetheless express the same subtypes of P2Y ATP receptors on their apical and basolateral membrane.³³ These findings thus show that there are functional as well as structural similarities in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling between cholangiocytes and NRC cells.

Fig. 9.

Ca²⁺ oscillations in (A and B) NRC cells and (C) cholangiocytes. NRC cells were loaded with fura-2 and then exposed apically to ATP and examined by fluorescence microscopy, whereas cholangiocytes were loaded with fluo-4 and then exposed basolaterally to ATP and examined by confocal microscopy. (A) ATP (100 μmol/L) triggered a large increase in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ followed by ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ oscillations. The ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signal stopped with removal of the agonist. (B) A lower concentration of ATP (1 μmol/L) elicited only a small initial increase in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$, but this was followed by ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ oscillations of the same magnitude and frequency as induced by 100 μmol/L ATP. (C) ATP (10 μmol/L) also induces ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ oscillations in cholangiocytes. This tracing was obtained by monitoring a single cholangiocyte within an isolated bile duct unit using confocal microscopy. The result is representative of that seen in 15 cells.

Subcellular Ca²⁺ Signaling Patterns in cholangiocytes

To determine the relationship between the distribution of InsP₃R isoforms and subcellular ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling patterns, fluo-4–loaded cholangiocytes were examined by confocal line scanning microscopy.^29,30 This imaging approach was used because it is well suited to detect subcellular ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals, which typically occur on a millisecond timescale.^35,36 Because confocal line scanning microscopy of polarized epithelia requires the apical-to-basal axis of the cells to be within the focal plane,^29,30,37 we examined ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling in cholangiocytes within microdissected segments of intrahepatic bile ducts. However, ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling in NRC cells could not be examined by this approach because their apical-to-basal axis is perpendicular to the plane of focus. Individual cholangiocytes within intrahepatic bile duct segments were double labeled by retrograde injection of fluo-4 plus DiD before microdis-section to facilitate identification of the cells (Fig. 10). Bile duct segments then were stimulated with either ACh or ATPγS, both of which cause an InsP₃-mediated increase in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ in this cell type.¹¹ The nonhydrolyzable form of ATP was used here because nucleotides otherwise are rapidly hydrolyzed at the basolateral membrane of cholangiocytes.³³ ACh (100 μmol/L) caused an increase in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ throughout the cholangiocyte. The increase in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ began in the apical region and then spread rapidly (22.5 ± 9.3 μm/s; n = 40) to the basolateral region (Fig. 11). Both the direction and speed of this ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ wave were unchanged in Ca²⁺-free medium (21.5 ± 7.9 μm/s; n = 8; P > .35). Progressive, apical-to-basal increases in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ also were observed in cholangiocytes stimulated with ATPγS (Fig. 11E). These findings suggest that InsP_3-mediated ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals in cholangiocytes begin in the apical region and then spread as polarized, apical-to-basal ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves.

Fig. 10.

Identification of live cholangiocytes within strips of portal tracts. The common bile duct was coinfused with the Ca²⁺ dye fluo-4 and the lipophilic dye DiD, and then strips of portal tracts were isolated and examined by confocal microscopy. The left panel shows cholangiocytes and adjacent interstitium stained with DiD. The middle panel shows staining with fluo-4 (scale bar, 20 μm). The right panel shows the merged (superimposed) image.

Fig. 11.

ACh-induced ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves in cholangiocytes visualized by confocal line scanning microscopy. (A) Confocal image of a portal tract segment. The cells have been loaded with DiD to facilitate identification. (B) Confocal image of the same segment excited at a lower wavelength to show loading with fluo-4. The line used for line scanning is shown in white. Note that the line crosses the apical-to-basal axis of one of the cells. This and the subsequent image are pseudocolored according to the color scale shown below. (C) Confocal line scanning image of the cell shown in B. The tissue was stimulated with ACh (100 μmol/L) as confocal fluorescence along the line was collected every 6 milliseconds. An increase in fluorescence occurs throughout the cholangiocyte, which begins in the apical region and then spreads rapidly to the basolateral region, representing a ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ wave that crosses from the apical to the basolateral pole of the cell.

(D) Tracing of subcellular ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals in the same cell. The apical increase in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ (left arrow) precedes the basolateral ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ increase (right arrow) by approximately 100 milliseconds, reflecting an apical-to-basal ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ wave. Notice that the final increase in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ is similar in the apical and basolateral poles, reflecting the fact that the ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ wave does not diminish as it crosses the cell. The result is representative of that seen in 40 cells. (E) ACh-induced ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves are not affected by extracellular Ca²⁺. Results are mean ± SEM of 40 cells in Ca²⁺-containing medium and 8 cells in Ca²⁺-free medium. (F) Tracing of subcellular ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals in a separate cell stimulated with ATPγS (100 μmol/L). A regenerative, apical-to-basal ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ wave is seen (left arrow indicates onset of apical Ca²⁺ signal, right arrow indicates onset of basolateral signal), similar to what is observed in cells stimulated with ACh. The result is representative of that seen in 5 cells.

Discussion

There are 3 isoforms of the InsP₃R, and it was previously reported that the type III isoform is expressed in cholangiocytes.²⁸ Here we report that cholangiocytes actually express all 3 isoforms of the InsP₃R but in different amounts and subcellular distributions. Specifically, the type III isoform was most concentrated apically, whereas the other 2 isoforms were distributed more uniformly throughout the cytosol. Although quantitative PCR studies suggested that each isoform is expressed to a similar extent, immunoblotting instead suggested that the type III isoform is expressed most heavily. There are several possible reasons for this apparent discrepancy. First, mRNA levels do not always reflect protein levels. Second, cholangiocyte mRNA could have been contaminated with mRNA from hepatocytes, which predominantly express the type I and II isoforms.^23,28 Finally, errors may have been introduced by our method of estimating InsP₃R protein levels. However, our immunofluorescence examination of bile ducts plus the quantitative PCR studies of NRC cells together support the conclusion that the type III InsP₃R is indeed the most heavily expressed iso-form. It was previously reported based on quantitative PCR studies that cholangiocytes lose the type III InsP₃R but gain type II InsP₃R after bile duct ligation.²⁸ In hepatocytes, the type II InsP₃R comprises 80% of the total pool of InsP₃R and the remainder is the type I isoform.²³ Therefore, the profile of InsP₃R isoforms in cholangiocytes becomes similar to hepatocytes after bile duct ligation. Together, these findings suggest that the expression and subcellular distribution of InsP₃Rs in cholangiocytes is more similar to what is found in NRC cells than in cholangiocytes following bile duct ligation.

The current findings also show that ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals begin in the apical region of cholangiocytes, even though InsP₃ presumably is generated basolaterally when cells are stimulated with ACh. There are several observations that may explain this finding. First, the range of InsP₃ is up to 25 μm in cytosol,³⁸ so it is feasible for InsP₃ to act at a site different from where it is produced. Second, we observed that InsP₃Rs are more concentrated in the apical region of cholangiocytes, and InsP -mediated ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals may begin in regions where the InsP₃R is clustered.³⁹ Finally, ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals in cholangiocytes begin in the region where the type III InsP₃R is concentrated. This is consistent with the idea that the type III isoform of the InsP₃R acts as a molecular trigger for ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals, because Ca²⁺ released from this isoform stimulates further Ca²⁺ release in a positive feedback fashion.¹⁷ This isoform is localized to the apex of a number of epithelia in addition to cholangiocytes, including pancreatic and salivary acinar cells^40-42 and the nonpigmented epithelium of the ciliary bilayer of the eye,²⁵ and the apical region serves as the initiation site for ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves in each of these cell types. Furthermore, hormone-induced ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves spread from the apical to the basal pole in each of these cell types, as it does in cholangiocytes. However, there are at least 2 different mechanisms by which polarized ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves may spread. In acinar cells, all 3 InsP₃R isoforms are apical,^40-42 whereas the RyR is basolateral.¹⁵ The InsP₃R is required for apical initiation of the ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signal, whereas the RyR is needed for this signal to spread into the basolateral region.^30,37,43,44 In contrast, nonpigmented epithelium cells express only the type III InsP₃R apically and express the type I InsP₃R basolaterally.²⁵ Thus, the type I InsP₃R rather than the RyR is responsible for the spread of ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves into the basolateral region in that cell type. Because cholangiocytes express type I and II InsP₃Rs in the baso-lateral region and do not express the RyR, the molecular basis for ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves in cholangiocytes seems similar to what occurs in nonpigmented epithelium cells.

How does the current study increase our understanding of bile duct cell function? Evidence from other polarized epithelia suggests that apical ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals and apical-to-basal ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves regulate secretion. For example, apical increases in ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ in pancreatic acinar cells lead to targeting of vesicles to the apical membrane and thus induce exocytosis.³ Bile ductular secretion depends on apical insertion of water channels^45,46 and other membrane fusion events,^47,48 which establishes a potential role for apical ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signals in cholangiocytes. In addition, apical-to-basal ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves direct luminal Cl⁻ secretion in acinar cells.^3,49 Because ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ mediates apical bicarbonate secretion in cholangiocytes^33,50 and cholangiocytes express apical, Ca²⁺-activated Cl⁻ channels,^51,52 polarized ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ waves may similarly direct fluid and electrolyte secretion in bile ducts. The small size and intrahepatic location of cholangiocytes limits the ability to relate ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ signaling to secretion directly in these cells. However, NRC cells are known to form polarized monolayers that exhibit well-defined apical and basolateral domains²¹ and have been used to monitor secretion induced by ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ agonists.³⁴ Here we furthermore show that the predominant InsP₃R in NRC cells is the type III isoform, which is more concentrated apically, and that biliary solutes such as ATP⁵³ can induce ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ oscillations in these cells. Thus, the asymmetric distribution of InsP₃R isoforms may enable ${\mathrm{Ca}}_{\mathrm{i}}^{2+}$ to regulate secretion by cholangiocytes, and NRC cells seem to be an appropriate model in which to examine this.

Abbreviations

$C{a}_i^{2+}$

cytosolic Ca²⁺

InsP₃

inositol 1,4,5-trisphosphate

IP₃R

inositol 1,4,5-trisphosphate receptor

RyR

ryanodine receptor

NRC

normal rat cholangiocyte

ACh

acetylcholine

ATP

adenosine triphosphate

PCR

polymerase chain reaction

RT

reverse transcription

DiD

1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine perchlorate

mRNA

messenger RNA