Substantial rapid passive movement of water across biliary epithelia in response to osmotic gradients established by actively transported ions and solutes suggests that water channels (i.e., aquaporins [AQPs]) likely are involved in mechanisms of water transport by cholangiocytes, the epithelial cells that line the intrahepatic biliary tree.1–8 Moreover, we recently reported that siRNA-mediated suppression of AQP1 gene expression in rat cholangiocytes results in a marked decrease in transcellular water flow and transported volume of water in isolated intrahepatic bile ducts units.9 In addition to AQP1, rat and mouse cholangiocytes express at least five other AQPs (i.e., AQP0, AQP4, AQP5, AQP8, and AQP9), which also could be important in water transport by intrahepatic bile ducts.7, 8, 10 However, clarifying the role of AQPs in mediating water transport by biliary epithelia has been limited in part because of the absence of physiologically relevant experimental models. In this study, we combined the advantages of normal mouse cholangiocytes (NMCs), which lack endogenous AQP4 when grown in culture, forming a monolayer and a retroviral gene delivery system that allows integration of the gene of interest into the host cell chromosome, that is, in our model, stable transfection of cholangiocytes with exogenous functional AQP4. AQP4 has the highest single-channel water conductance among mammalian AQPs and, in rat, is expressed constitutively on the cholangiocyte basolateral plasma membrane domain.11 Given the physical proximity of the cholangiocyte basolateral plasma membrane to the peribiliary vascular plexus from which water destined for secretion into the biliary system is derived, AQP4 may be one of the AQPs involved in water transport by cholangiocytes, thus playing an important role in ductal bile formation. In this study, we established an AQP4-transfected polarized mouse cholangiocyte cell line suitable for functional studies of transepithelial water transport, and, using this model, we defined the importance of AQP4 in water transport across biliary epithelia.
Rodent cholangiocytes express 6 of the 11 known channel proteins called aquaporins (AQPs) that are involved in transcellular water transport in mammals. However, clarifying the role of AQPs in mediating water transport in biliary epithelia has been limited in part because of the absence of physiologically relevant experimental models. In this study, we established a novel AQP4-transfected polarized mouse cholangiocyte cell line suitable for functional studies of transepithelial water transport, and, using this model, we define the importance of this AQP in water transport across biliary epithelia. Polarized normal mouse cholangiocytes (NMCs) lacking endogenous AQP4 were transfected stably with functional AQP4 or cotransfected with functional AQP4 and a transport-deficient AQP4 dominant negative mutant using a retroviral delivery system. In transfected NMCs, AQP4 is expressed on both the mRNA and protein levels and is localized at both the apical and basolateral membranes. In nontransfected NMCs, the transcellular water flow, Pf, value was relatively high (i.e., 16.4 ± 3.2 μm/sec) and likely was a reflection of endogenous expression of AQP1 and AQP8. In NMCs transfected with AQP4, Pf increased to 75.7 ± 1.4 μm/sec, that is, by 4.6-fold, indicating the contribution of AQP4 in channel-mediated water transport across MNCs monolayer. In cotransfected NMCs, AQP4 dominant negative reduced Pf twofold; no changes in Pf were observed in NMCs transfected with the empty vector. In conclusion, we developed a novel polarized mouse cholangiocyte monolayer model, allowing direct study of AQP4-mediated water transport by biliary epithelia and generated data providing additional support for the importance of AQP4 in cholangiocyte water transport. (HEPATOLOGY 2004;39:109–116.)
Materials and Methods
All chemicals were of highest commercially available purity and were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise indicated.
Normal mouse cholangiocytes (a gift from Dr. Yoshi Ueno, Tohoko University, Japan) were grown in Eagle's minimum essential medium supplemented with 10% fetal bovine serum (FBS), 1,000 units/mL of penicillin and 100 μg/mL streptomycin. GPE86 cells were grown in Dulbecco's minimum essential medium supplemented with 10% FBS, 1,000 units/mL of penicillin and 100 μg/mL streptomycin. NMCs and GPE86 cells routinely were passaged once and twice per week, respectively, using trypsin.
Cloning of AQP4 and AQP4 (LY246/247PP) Into pLNCX Vectors.
The retroviral shuttle vector (pLNCX) was purchased from Clontech (Palo Alto, CA). The plasmid containing the rat AQP4 cDNA (pAQP4) was purchased from ATCC (Rockville, MD). Briefly, the AQP4 cDNA was excised with EcoR I from pAQP4 and filled in using Klenow DNA polymerase. The pLNCX vector was digested and linearized with Hpa I, creating blunt ends. Subsequently, a blunt ligation was performed that yielded the pLNCX (AQP4) plasmid. The AQP4 dominant negative protein construct was created using site-directed PCR mutagenesis. The sense PCR primer (5′ATTGAGCTCAAAGCTTGCCATGTACCCATACGACGTCCCAGACTACTGGGGT-3′) contained a Sac I and Hind III site and a hemagglutinin epitope tag. The antisense PCR primer (5′-ACGTTTGAGCTCCACGTCAGGACAGAAGACATACTCGTAAGGTGGACCTGCCAGCACAGCGCCTATGATTGGTCCAACCC-3′) contained two point mutations (LY246/247PP) and a Sac I site. The PCR reaction was carried out with 200 ng of pAQP4 template using Pwo I polymerase. The 730-bp amplicon was isolated from a 1.0% (wt/vol) agarose gel and subsequently digested with Sac I. pAQP4 was digested with Sac I and gel purified. The 730-bp SacI fragment was ligated into the Sac I site of pAQP4, thereby creating the pAQP4 (LY246/247PP) construct. pAQP4 (LY246/247PP) plasmid was digested with Hind III, the 1300-bp Hind III mutated AQP4 fragment was cloned into the Hind III site of pLNCX, creating the pLNCX (LY246/247PP) construct (hereafter referred to as LY246/247PP).
Transfection of the GPE86 Retroviral Packaging Cell Line.
The GPE86 cells were transfected using Lipofectamine Plus Reagent (Gibco BRL, Grand Island, NY) according to the manufacturer's directions. Briefly, the GPE86 cells were grown to 50% confluency the day of transfection and the cells were washed twice with 1X phosphate-buffered saline. Next, 0.5 μg of plasmid DNA was mixed with the Lipofectamine Plus Reagent and added to the cells. The cells were incubated for 4 h after that the plasmid/Lipofectamine Plus Reagent mix was replaced with standard GPE86 culture medium. After 48 h, the medium was replaced with medium containing 800 μg/mL of G418 to select for stable clones.
Retroviral Transfection of NMCs With the Plasmid Constructs.
Twenty-four hours before retroviral transfection, the G418 was removed from the stably transfected GPE86 cells. NMCs were grown to 50% confluency the day of retroviral transfection. The day of retroviral transfection, the medium, containing the viral lysate, was collected from the stably transfected GPE86 cells and was filter-sterilized using a low protein-binding 0.45 μm cellulose acetate syringe filter. Next, polybrene (8 μg/mL; Sigma) was added to the filtered viral lysate medium, and the viral lysate was added directly to the NMCs. The NMCs were cultured overnight, and the medium was replaced the next day. 48 h after transfection, NMCs medium containing 4,000 μg/mL of G418 was used to select for stably transfected cells. The LY246/247PP viral lysate was used to cotransfect the stably transfected pLNCX (AQP4) NMCs using the aforementioned procedure.
Transfection of NMC With AQP4.
To generate a cholangiocyte cell line that constitutively expressed AQP4, the rat AQP4 cDNA was subcloned into a retroviral expression vector to form pLNXC (AQP4). The expression vector was transfected into the GPE86 packaging cell, and stable clones were harvested 10 to 14 days after selection with G418. The tissue culture supernate, containing the recombinant retrovirus with the AQP4 coding sequence from the stably transfected GPE86 cells, was used to transfect NMCs. The NMCs were selected with G418 to generate AQP4 stably transfected NMCs. In a similar manner, the tissue culture supernate was harvested from pLNCX (LY246/247PP) stably transfected GPE86 cells and was used to cotransfect the AQP4-NMCs. The cotransfection generated a cell line in which the mouse cholangiocytes express AQP4 and the dominant negative protein of AQP4.
Confirmation of AQPs Message in NMCs.
RNA was isolated from the NMCs using TRI Reagent (Sigma) according to the manufacturer's protocol. Briefly, the cells were homogenized in TRI reagent and the RNA was extracted with bromochloropropane (Sigma). Next, the RNA was precipitated with isopropanol, washed with 75% ethanol, air dried, and resuspended in RNAsecure Resuspension Solution (Ambion, Austin, TX). The first strand complimentary DNA (cDNA) was synthesized using the SuperScript preamplification system (GIBCO BRL, Grand Island, NY). Specific oligonucleotides were synthesized based on the published sequence for AQP1, AQP4, and AQP8: specifically for AQP1, 5′ GTTTATTTCGGTCCTCCAGTCTC 3′ (forward) and 5′ CGTGTCAGAAGAGAGACGTAAC 3′ (reverse); for AQP4, 5′ CCAAACGGACTGATGTTACT 3′ (forward) and 5′ AGGATCAAGTCTTCTGTCTCC 3′ (reverse); for LY246/246PP anchored polymerase chain reaction (PCR) primers, 5′ CAGACTACTGGGGTATGAGTG 3′ (forward) and 5′ GCAATGCTGAGTCCAAAG 3′ (reverse); for AQP8, 5′ TTCTGACAATGCTGCTGTTGGTATTG 3′ (forward) and 5′ CTCTTTTGGGCGGATTAAGATTT 3′ (reverse). Polymerase chain reaction was performed using GeneAmp PCR reagent kit and Ampli Taq DNA polymerase (Perkin Elmer, Norwalk, CT) according to the instructions from the vendor. The PCR products were sequenced by Mayo Molecular Core Facility (Rochester, MN) to confirm sequence identity.
Protein Confirmation in Transfected NMCs.
NMCs were grown on collagen-coated coverslips and fixed by 0.1 M PIPES, pH 6.95, 1 mM ethylene glycol-bis(β-amino-ethylether-N,N,N′,N′-tetra acetic acid (EGTA), and 2% paraformaldehyde in 1X phosphate-buffered saline for 20 min at room temperature. The cells then were permeabilized in 0.2% Triton for 2 min and incubated for 1 h with AQP4 affinity-purified polyclonal (1:1000; Alpha Diagnostics International, San Antonio, TX) or hemagglutinin antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). The cells subsequently were washed, incubated with biotinylated secondary antibody, and developed with DAB (Sigma). A control without primary antibody was performed to confirm the absence of labeling.
Preparation of Apical and Basolateral NMCs Plasma Membranes.
Apical and basolateral plasma membranes were prepared from NMCs grown to confluency on rat tail collagen type I as previously described by us.12 Briefly, NMCs were sonicated in 0.3 M of sucrose containing protease inhibitors, and the plasma membrane fraction was obtained after centrifugation at 200,000g for 60 min on 1.3 M sucrose gradient. Next, the plasma membrane proteins were subfractionated by high-speed centrifugation through a discontinuous sucrose gradient to obtain highly purified apical and basolateral plasma membrane domains.
Subcellular Localization of AQP4 by Immunoblotting.
Solubilized plasma membrane fractions isolated from nontransfected and AQP4-transfected NMCs were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nylon membranes. After blocking, blots were incubated overnight at 4°C with affinity-purified AQP4 polyclonal antibody (2 μg/mL; Alpha Diagnostics International, San Antonio, TX). The blots were washed and incubated for 1 h at room temperature with horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (1:2,000 dilution), and bands were detected using the enhanced chemiluminescent plus detection system (ECL PLUS; Amersham, Arlington Heights, IL). Autoradiographs were obtained by exposing the nitrocellulose to Kodak XAR film (Rochester, NY).
Water Transport in Nontransfected and AQP4-Transfected NMCs.
The approach, which previously had been applied successfully for functional evaluation of transcellular water flow in Madin Darby Canine Kidney (MDCK) cells transfected with AQP1,13 was used in this study to measure transcellular water flow in nontransfected NMCs and NMCs transfected with functional AQP4. Briefly, NMCs were grown to confluence in a polarized manner in 0.33 cm2 collagen-coated Transwell inserts. Then, 4 days after seeding, the apical and basolateral compartments were washed twice with isotonic (290 mosM) Hepes-buffered saline (1X HBS: 135 mM NaCl, 5 mM KCl, 0.8 mM Na2HPO4, 25 mM Na Hepes, pH 7.4) and filled with 200 μL of hypertonic (i.e., 340, 390, or 490 mosM) HBS containing 30 μg/mL of phenol red and 800 μL of 1X HBS, respectively. After incubation for 2 h at 37°C, the content of the apical compartment was mixed with pipette and two aliquots of 50 μL were put into Eppendorf tubes (for duplicated measurements), diluted to 800 μL with 1X HBS, and analyzed for absorbency at 479 nm (A479), the isosbestic point for phenol-red. Any A479 from the basolateral compartment with a value >0.01 were considered leaky and were not included in further calculations. Transcellular water flow (Pf in μm/s) was calculated using the following equation:
where Jv is the transported volume of water (in microliters per second), iVola is the initial volume in the apical compartment (in microliters), iA is the initial absorbency, eA is the end absorbency, t is time (in seconds), Vm is the partial molecular volume of water (18 cm3/mol), A is area (in square centimeters), and δC is the osmotic gradient (in moles per cubic centimeter).
All values are expressed as mean ± SE. Statistical analysis was performed by the Student's t test, and results were considered statistically different at P < .05.
Expression of AQPs in Wild-type and Transfected NMCs.
Polarized NMCs lacking endogenous AQP4 were transfected stably either with the cDNA corresponding to the open reading frame of rat AQP4 to restore AQP4-mediated water transport across mouse cholangiocytes monolayer or cotransfected with both functional AQP4 and transport deficient AQP4 (LY246/247PP) dominant negative mutant to abolish AQP4-mediated water transport in transfected cholangiocytes. Data in Fig. 1A show that wild-type NMCs express mRNAs for two of three tested AQPs: AQP1 and AQP8. AQP4 mRNA was absent in wild-type NMCs but was abundant in NMCs transfected with AQP4 alone or cotransfected with AQP4 and the AQP4 dominant negative mutant construct (Fig. 1B). To verify expression of the AQP4 dominant negative mutant in cotransfected NMCs, anchored reverse transcription-polymerase chain reaction (RT-PCR) was performed using the sense primer anneal within the hemagglutinin epitope tag, which is unique to AQP4 dominant negative construct and therefore does not detect the AQP4 construct. Figure 1C shows that the AQP4 dominant negative mutant was amplified in mouse cholangiocytes, thereby confirming successful cotransfection of the NMCs. In contrast, the AQP4 message was not detected by RT-PCR in nontransfected NMCs or in NMCs transfected with a retroviral shuttle vector (pLNCX) alone (Fig. 1).
Using the affinity-purified rabbit antibody against AQP4 and the monoclonal antibody raised against the hemagglutinin epitope tag, we observed strong staining for AQP4 and AQP4 dominant negative mutant in AQP4-transfected and cotransfected NMCs, respectively, confirming expression of transfected and cotransfected proteins in the mouse cholangiocytes (Fig. 2). No AQP4 or hemagglutinin were detected in nontransfected NMCs (Fig. 2).
Subcellular Localization of AQP4 in Transfected NMCs.
Immunoblot analysis of highly purified apical and basolateral plasma membrane domains from nontransfected NMCs, AQP4-transfected NMCs, and AQP4 dominant negative cotransfected NMCs showed the expression of AQP4 on both the basolateral and apical plasma membrane domains in AQP4-transfected cholangiocytes, whereas no protein was detected in nontransfected cells (Fig. 3A). In AQP4 dominant negative cotransfected NMCs, mutated AQP4 also was expressed at the same location, that is, on both basolateral and apical membrane (Fig. 3B).
Water Transport in AQP4-Transfected NMCs.
The transported volume of water, Jv, and transcellular water flow, Pf, in response to osmotic gradients of 50 to 200 mosM established by the addition of sucrose or NaHCO3 to the HBS were measured in nontransfected NMCs, NMCs transfected with AQP4, or with both AQP4 and AQP4 dominant negative mutant. In nontransfected cholangiocytes, Jv was dependent and Pf was independent of the established osmotic gradients (Fig. 4), an observation consistent with universal characteristics of osmotic water transport. The Pf value was relatively high when compared with values known for tissues that do not express AQPs and likely is a reflection of endogenous expression of AQP1 and AQP8 in wild-type NMCs. In NMCs transfected with AQP4, Jv and Pf increased to −42.1 ± 1.9 nL/sec and 75.7 ± 1.4 μm/sec, respectively, that is, by 4.6-fold compared with nontransfected NMCs (i.e., −9.4 ± 1.8 nL/sec and 16.4 ± 3.2 μm/sec, respectively), indicating the contribution of AQP4 in channel-mediated transcellular water transport across MNCs monolayer (Fig. 5). In cotransfected NMCs, AQP4 dominant negative reduced Jv and Pf twofold; no changes in Jv and Pf were observed in NMCs transfected with the empty vector (i.e., −10.5 ± 2.9 nL/sec and 17.4 ± 5.5 μm/sec, respectively; Fig. 5).
The major findings reported here relate to development of a novel model suitable to direct study of AQP4-mediated water transport in biliary epithelia. We: (i) used a retroviral delivery system to transfect polarized NMCs with transport capable and transport-deficient AQP4 constructs and (ii) showed a relationship between cholangiocyte water permeability and expression of functional AQP4. The data are consistent with a potential role for AQP4 in water transport by biliary epithelia.
Although many observations exist for retroviral gene delivery into hepatocytes and other cells,14, 15 to our knowledge, this study represents the first example of retroviral gene delivery into cholangiocytes. Taking the advantage of a retrovirus delivery system, which allows integration of the gene of interest into the host cell chromosome, we successfully transfected the AQP4 coding sequence into NMCs, which subsequently was detected on both the mRNA and protein levels. Moreover, we successfully cotransfected the AQP4-NMCs with a retroviral plasmid containing the LY246/247PP dominant negative mutant construct of AQP4, which allowed us specifically to inhibit AQP4 expression and function in AQP4-transfected NMCs. Our data also indicate that a retroviral gene delivery approach, which we used in this study, does not alter the NMCs monolayer organization and, thus, suggest that AQP4-transfected and cotransfected NMCs are suitable for studying AQP4-mediated water transport in biliary epithelia.
AQP4 is a widely distributed mercury-insensitive water channel that is constitutively expressed in fluid transporting epithelia such as renal collecting duct principal cells,16–18 gastric parietal cells,19–22 excretory ducts in salivary and lacrimal glands,19, 23 small intestine,24 colon,25 and airway epithelia.26, 27 We previously reported that in the rat liver, AQP4 is expressed on the cholangiocyte basolateral membrane domain, and we hypothesized that such functional expression of AQP4 may be critically important in ductal bile formation.11 The current study was performed to define further the functional contribution of AQP4 to cholangiocyte water transport. Normal mouse cholangiocytes were chosen because: (i) when grown to confluence on collagen-coated inserts, they form a polarized monolayer with well-developed tight junctions and, thus, are suitable for transport studies; and (ii) polarized NMCs do not express AQP4, and so the significance of AQP4 in water transport by cholangiocytes could be addressed directly using transfection.
Transfected NMCs express AQP4 on both the mRNA and protein levels; however, the subcellular distribution of transfected AQP4 is different than for endogenous AQP4 in freshly isolated rat cholangiocytes. Transfected NMCs express AQP4 not only at the basolateral membrane, but also at the apical membrane. There are two possible explanations for this finding. First, many factors contribute to the variable polarity of AQPs in epithelial cells. For example, it has been shown that AQP2, which is primarily targeted to the apical plasma membrane in the collecting duct principal cells in kidney in response to vasopressin, in MDCK cells adapted to hypertonic culture medium is targeted to the basolateral membrane in response to forskolin.28 Furthermore, kidney slices taken from vasopressin-deficient Brattleboro rats show apical insertion of AQP2 in the inner medulla when exposed to vasopressin. In contrast, slices from normal rats show a marked basolateral insertion of AQP2 when challenged acutely with vasopressin in vitro. Finally, inner medullary principal cells in kidney slices from Brattleboro rats pretreated with vasopressin for 11 days in vivo show basolateral instead of apical AQP2 insertion when exposed to acute vasopressin stimulation in vitro.29 Thus, we cannot exclude that expression of AQP4 on both apical and basolateral membranes in transfected NMCs is a reflection of the functional state of mouse cholangiocytes in culture. Second, retroviral transfection of NMCs with functional or nonfunctional AQP4 can affect the cholangiocyte sorting machinery. Others have noted that overexpression of a transgene can result in undesirable missorting of the encoded protein,30, 31 a phenomenon that may have accounted for our findings. However, work necessary to address this question fully is beyond the scope of the study described.
Pf measured in nontransfected NMCs was 16.4±3.2 μm/sec, a value consistent with AQP-mediated transcellular water movement, and was independent of the size of osmotic gradient. This finding is consistent with the expression of AQP1 and AQP8 in mouse cholangiocytes, which could mediate transcellular water transport. Upon transfection of the NMCs with AQP4, Pf increased fivefold. Cotransfection of NMCs with a dominant negative AQP4 (LY246/247PP) mutant significantly reduced Pf when compared with the AQP4-transfected NMCs, suggesting that the increase in Pf in AQP4-transfected NMCs is a result of expression of this particular aquaporin; the mechanism for this effect is unclear. According to Shi and Verkman,32 mutant monomers of AQP4 may interact with AQP4 tetramer and functionally may alter the tetramer via a dominant negative effect when expressed in Xenopus laevis. The other possibility is that AQP4(LY246/247PP) dominant negative monomers may form heterotetramers with AQP4 monomers causing formation of a defective AQP4 heterotetramer. Thus, expression of AQP4 in transfected NMCs resulted in an increase in transepithelial water transport, suggesting that this AQP is responsible for a large amount of water transported by cholangiocytes.
In summary, this work provides additional support of the importance of AQPs in cholangiocyte water transport and potentially in ductal bile formation.