INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE AQUAPORINS CONSTITUTE a family of membrane proteins that function as selectively permeable water channels to facilitate water movement across membranes in all eukaryotic organisms (1). The subtype aquaporin-2 (AQP-2) is the only known water channel to be rapidly regulated by a hormone. The antidiuretic hormone arginine-8 vasopressin (AVP) causes an increase in intracellular cAMP concentration in principal cells of the inner medullary collecting duct (IMCD) by activation of the G_s/adenylyl cyclase system and, as a consequence, activation of protein kinase A (PKA). PKA, in turn, phosphorylates AQP-2, localized in intracellular vesicles (9, 14), which subsequently fuse with the cell membrane. The integration of AQP-2 into the plasma membrane increases the water permeability of the bilayer and allows efficient water movement down an osmotic gradient through the principal cell. Withdrawal of AVP leads to the endocytosis of AQP-2 and cessation of water resorption. This mechanism is known as the shuttle hypothesis (31). Disturbances in the AVP V₂ receptor (V₂R)-AQP-2 signaling cascade cause severe dysregulation in water homeostasis through a massive loss of water via the kidney, i.e., diabetes insipidus (3).

In addition to the AVP-induced exocytotic insertion of AQP-2-bearing vesicles into the cell membrane (short-term regulation), an increase in AQP-2 synthesis (long-term regulation; see Refs. 5 and 16) is observed in Brattleboro rats after prolonged treatment with AVP for several days. On a cellular level, a major problem in the investigation of AQP-2 regulation has been the lack of an authentic cell line expressing the AQP-2 protein at levels adequate for experiment. Usually, primary cultures halt, or at least rapidly downregulate, AQP-2 expression (8). To overcome this problem, tissue preparations of freshly isolated collecting ducts or cell lines transfected with the AQP-2 cDNA (4, 13, 30) have been used to study the cellular action of AVP. Our aim was to define conditions favoring the long-term upregulation of AQP-2 in authentic primary cultures in vitro.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture. IMCD primary cultures were prepared as described (22). Briefly, rats were killed by decapitation, and kidney inner medulla, including papilla, was removed and cut into small pieces. Tissue was digested in phosphate-buffered saline (PBS; in mM: 137 NaCl, 2.7 KCl, 1 KH₂PO₄, and 10 Na₂HPO₄, pH 7.4) containing 0.2% hyaluronidase (Boehringer, Mannheim, Germany) and 0.2% collagenase type CLS-II (Biochrom, Berlin, Germany) at 37°C for 90 min. Thereafter, cells were centrifuged, washed three times, and seeded at a density of ~10⁵ cells/cm² on petri dishes or glass coverslips coated with either rat-tail collagen (Serva, Heidelberg, Germany) or type IV collagen (5 µg/cm²) (TEBU, Frankfurt am Main, Germany; or Becton-Dickinson, Heidelberg, Germany). This seeding density yields 10-14 cm² of confluent monolayer per animal used, resulting in approximately ten 100-mm culture dishes from 40 animals. Preparations of this size yield ~5,000 µg of total membrane protein. Because IMCD cells are accustomed to high osmotic challenges within the kidney medulla, Dulbecco's modified Eagle's medium (DMEM), adjusted to 600 mosmol/l by the addition of 100 mM NaCl and 100 mM urea, was used to establish growth conditions with preferential selectivity for IMCD principal and intercalated cells (22). Penicillin (0.5 IU/ml), streptomycin (0.5 µg/ml), glutamine (2 mM), nonessential amino acids (1%), and fetal calf serum (5%) were routinely added. Dibutyryl cAMP (DBcAMP; 500 µM) was added as indicated. The cell culture medium was exchanged three times per week. For Western blot analysis, cells were used from days 1-6 after seeding. For all other experimental procedures, cells from day 6 were used.

Antisera. Polyclonal antisera against AQP-2 were raised against synthetic peptides corresponding to the 15 COOH-terminal amino acids of AQP-2 with an additional NH₂-terminal tyrosine residue. One peptide (Y-VELHSPQSLPKA) corresponded to the human (19), and the other (Y-VELHSPQSLPKA) corresponded to the rat sequence. The peptides were coupled to keyhole limpet hemocyanin and the conjugates were used for immunization. Both antisera revealed similar properties. The antiserum raised against the human sequence was used for Western blotting experiments, whereas the antiserum against the rat sequence, producing less background, was used in immunofluorescence experiments. The anti-Rab3 monoclonal antibody Cl.42.1 (7), kindly provided by Reinhard Jahn (Goettingen, Germany), recognizes the Rab3 subtypes a, b, and c.

Preparation of cell membranes. Crude membranes were prepared as described previously (18) with modifications. All procedures were performed at 4°C. One 60-mm culture dish (2 × 10⁶ cells) was washed twice with PBS, and cells were scraped from the dish and homogenized with 10 strokes at 750 rpm in a glass-Teflon homogenizer. The homogenate was centrifuged at 230 g for 2 min. The supernatant was then centrifuged at 200,000 g for 1 h. The pellet was resuspended in 100 µl PBS and stored at 80°C.

Preparation of fractions enriched in cell plasma membranes or intracellular vesicles. Membrane fractions enriched in vesicles or cell plasma membranes were prepared as described (20). All procedures were performed at 4°C. Cells grown on a 60-mm dish were washed twice with PBS, scraped, and homogenized with 5 strokes at 1,250 rpm in a glass-Teflon homogenizer in 1.5 ml dissecting buffer (0.3 M sucrose, 25 mM imidazole, 1 mM EDTA, 8.5 µM leupeptin, and 1 mM phenylmethylsulfonyl fluoride, pH 7.2). The homogenate was centrifuged at 4,000 g for 15 min, and the supernatant was collected. The remaining pellet was resuspended in dissecting buffer and rehomogenized with 3 strokes at 1,250 rpm, and the centrifugation was repeated as above. The two supernatants were combined, and low-speed (LS) and high-speed (HS) pellets were prepared by consecutive centrifugations at 17,000 g and 200,000 g for 1 h, respectively. The pellets represent fractions enriched for either plasma membranes (LS) or intracellular vesicles (HS). The pellets were resuspended in 150 µl dissecting buffer and stored at 80°C.

Immunoblotting. Immunoblotting was performed as described (30). Proteins were separated by SDS-PAGE (30 µg/lane, 12% polyacrylamide) and transferred onto nitrocellulose filters (Schleicher & Schuell, Dassel, Germany). The filters were blocked in blocking buffer (5% nonfat dried milk, 150 mM NaCl, 1% Triton X-100, and 20 mM Tris-HCl, pH 7.4) for 1 h at room temperature and incubated with antiserum against AQP-2 or monoclonal antibody against Rab3 proteins for 2 h or overnight, respectively. After washing, we used either alkaline phosphatase-conjugated goat anti-rabbit IgG or goat anti-mouse IgG (Dianova, Hamburg, Germany) as the second antibody. The filters were washed and stained with 0.56 mM 5-bromo-4-chloro-3-indolyl phosphate (BCIP) and 0.48 mM nitro blue tetrazolium (NBT) (Roth, Karlsruhe, Germany) in 10 mM Tris-HCl (pH 9.5).

Isolation of RNA from IMCD cells and Northern blots. Northern blots were essentially performed as described previously (25). We isolated total RNA from IMCD cells using the RNA Clean kit (AGS, Heidelberg, Germany). Size separation of RNA (10 µg) on 1.1% agarose gels containing 2.2 M formaldehyde was followed by vacuum blotting for 1 h onto nylon membranes (Qiagen, Hilden, Germany) and ultraviolet cross-linking (2 × 90 s at 312 nm). We performed prehybridization and hybridization with [-³²P]dCTP-labeled rat AQP-2 cDNA (2.5 × 10⁶ cpm/ml) (Prime It II Kit; Stratagene, Heidelberg, Germany) in QuikHyb solution (Stratagene) at 60°C for 1 h. We washed membranes at increasing stringency, with a final wash in 0.1× SSC (1× SSC = 150 mM NaCl, 15 mM sodium citrate)/0.1% SDS at 60°C for 15 min. The membranes were exposed on X-Omat AR films (Kodak, Rochester, NY) for 6 h to 5 days.

Adenylyl cyclase assay. We performed the preparation of nuclei-free crude membrane fractions from IMCD cells and the adenylyl cyclase assay essentially as described (27), except that the reaction mixture contained 1.5 mM MgCl₂ and 1 mM EDTA. [³²P]cAMP was isolated by the two-column method (26).

Immunofluorescence studies. IMCD cells (8 × 10⁵) were seeded in 35-mm petri dishes containing coverslips coated with type IV collagen and grown for 6 days. Cells were washed twice with Krebs-Ringer-Henseleit buffer (in mM: 130 NaCl, 4.7 KCl, 2.5 CaCl₂, 1.2 KH₂PO₄, 1.2 MgSO₄, 10 HEPES, and 5.5 glucose, pH 7.25) and fixed with 2.5% paraformaldehyde in sodium cacodylate buffer (100 mM sodium cacodylate and 100 mM sucrose, pH 7.4) for 30 min. After three washes with PBS, cells were permeabilized in PBS containing 0.1% Triton X-100 for 5 min and rewashed three times. The coverslips were then transferred into a humidifying chamber and incubated for 45 min at 37°C with 30 µl AQP-2 antiserum (1:400 in PBS) and washed three times (10 min each wash) with cold PBS. Cells were reincubated in a humidifying chamber for 45 min at 37°C with 30 µl CY3-conjugated anti-rabbit IgG (Dianova) (1:400 final dilution in PBS). Coverslips were then mounted with a mixture of glycerol (70%) and PBS (30%). 1,4-Diazabicyclo[2.2.2]octane (DABCO) (Sigma, Deisenhofen, Germany) (100 mg/ml) was added to reduce photobleaching (24). Samples were visualized by confocal microscopy on a Zeiss Axiovert 135 Micro Systems laser scanning microscope.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Administration of AVP to Wistar or Brattleboro rats leads to a stimulation of AQP-2 expression (5, 16). We therefore tested whether an elevation of cAMP, the cellular response to AVP, caused an increase in AQP-2 expression in vitro. To avoid blunting of the cAMP response, due to internalization of the V₂ receptor under cell culture conditions (2), we used the membrane-permeable, weakly hydrolyzable cAMP analog DBcAMP. In consideration of the epithelial nature of IMCD cells, type IV collagen, a major component of basement membranes (15), was used side by side with rat-tail collagen for coating the petri dishes.

Figure 1 shows a Western blot analysis, performed with an AQP-2 antiserum, of membrane preparations from IMCD cells grown under various cell culture conditions [with or without DBcAMP added to the culture medium, and with petri dishes coated either with rat-tail collagen (rt) or type IV collagen (IV)]. For freshly prepared rat kidney inner medulla homogenate (IM), a narrow band with a molecular mass of 28 kDa and a broad band of 35-40 kDa were observed, representing the unglycosylated and glycosylated forms of AQP-2, respectively (19). These bands disappeared after day 2 when cells were grown under standard culture conditions, i.e., on rat-tail collagen without the addition of DBcAMP. A prolonged expression of AQP-2 was observed when the petri dishes were coated with type IV collagen instead of rat-tail collagen. Addition of DBcAMP to the culture medium overcame the downregulation of AQP-2 synthesis in primary cultured IMCD cells grown on either collagen type. With this modified cell culture protocol (i.e., DBcAMP supplementation of the culture medium and type IV collagen coating), 46 of 48 primary cultures expressing AQP-2 protein at levels adequate for experiment were obtained. We compared AQP-2 levels by densitometry of Western blots from six randomly selected IMCD cultures grown independently on type IV collagen with DBcAMP-supplemented culture medium. The levels were (in arbitrary units) 95,768 ± 8,170 (mean ± SD). Several attempts to subcultivate the primary cultured IMCD cells were not successful, indicating a high degree of differentiation. In contrast to immortalized cell lines, primary cultured IMCD cells could be kept over long periods under cell culture conditions without subcultivation. Western blot analysis of IMCD cells grown for 3 wk still revealed AQP-2 expression comparable to cells grown for 6 days (data not shown).

View larger version (64K): [in this window] [in a new window]   Fig. 1.   Influence of cell culture conditions on aquaporin-2 (AQP-2) expression in primary cultured inner medullary collecting duct (IMCD) cells. Proteins of IMCD membranes were analyzed by SDS-PAGE/immunoblotting with AQP-2 antiserum. Filter-bound antibodies were visualized with alkaline phosphatase-conjugated goat anti-rabbit IgG and 5-bromo-4-chloro-3-indolyl phosphate (BCIP)/nitro blue tetrazolium (NBT). Equal amounts of protein (30 µg/lane) were used. Rat inner medulla homogenate (IM) was used as control. Membranes from IMCD cells were prepared after cell culture for 1, 2, 4, or 6 days (1, 2, 4, and 6) and grown on rat-tail collagen (rt) or on type IV collagen (IV). Effect of addition of DBcAMP (500 µM) to culture medium is shown for IMCD cells grown on rat-tail collagen (rt+) or on type IV collagen (IV+). Arrows indicate nonglycosylated (ng) and glycosylated (g) forms of AQP-2.

Fractions enriched for AQP-2-bearing vesicles contain proteins involved in regulated exocytosis, including Rab3 proteins (11, 19). Figure 2 shows a Western blot of primary cultured IMCD cells stained with a monoclonal antibody against Rab3 proteins (Cl.42.1) (7). As in the case of AQP-2, Rab3 (27 kDa) expression was only detectable at day 6 when cells were grown in the presence of DBcAMP, indicating a cAMP-dependent regulation of Rab3 protein expression.

View larger version (39K): [in this window] [in a new window]   Fig. 2.   Expression of Rab3 proteins in primary cultured IMCD cells. Proteins were analyzed by SDS-PAGE/immunoblotting with monoclonal antibody specific for Rab3 proteins (Cl.42.1). Filter-bound antibodies were visualized with alkaline phosphatase-conjugated goat anti-mouse IgG and BCIP/NBT. Equal amounts of protein (30 µg/lane) were used. Rat brain homogenate was used as control. Membranes from IMCD cells grown without DBcAMP on type IV collagen (IV) or rat-tail collagen (rt) and from cells grown with DBcAMP supplementation (500 µM) on type IV collagen (IV+) or rat-tail collagen (rt+) were used.

Northern blot analysis (Fig. 3) confirmed the dependence of AQP-2 expression on culture conditions, although the mRNA levels did not correlate linearly with the protein expression levels. For cells grown on rat-tail collagen for 6 days, no signal was detectable, whereas those cells grown on type IV collagen showed a weak mRNA signal for AQP-2 at 1.8 kb. The addition of DBcAMP, at each medium exchange, strongly increased the signal irrespective of the collagen type used. Differences in mRNA levels were observed after 6 days of growth in the continuous presence of DBcAMP, followed by a 24-h deprivation of DBcAMP. Cells cultivated on type IV collagen exhibited diminished downregulation of AQP-2 mRNA compared with those grown on rat-tail collagen.

View larger version (44K): [in this window] [in a new window]   Fig. 3.   Detection of AQP-2 mRNA in primary cultured IMCD cells (Northern blot analysis). IMCD cells were grown for 6 days on type IV collagen (IV) or rat-tail collagen (rt) without () or with (+) DBcAMP supplementation (500 µM). In addition, cells were grown for 6 days with DBcAMP supplementation (500 µM) and for a further day without DBcAMP supplementation (IV+, rt+). Total RNA was isolated, and 10 µg were applied per lane. Blots were hybridized with radiolabeled AQP-2 cDNA probe. Signal runs at size of 1.8 kb. Control lane shows total RNA from freshly prepared IMCD cells (FP).

Aside from the retention of AQP-2 expression, we observed other advantageous effects of type IV collagen on cell growth. Cells cultivated on this coating reached confluence at day 2 after seeding, whereas cells grown on rat-tail collagen coating usually needed 1-2 additional days of growth until confluence. Therefore, all subsequently presented experiments were performed with IMCD cells grown on type IV collagen coating with DBcAMP supplementation of the cell culture media.

We performed adenylyl cyclase assays of membrane preparations from IMCD cells to demonstrate an intact coupling of the V₂R to adenylyl cyclase (Fig 4). Eighteen hours before the assay, DBcAMP was withdrawn from the culture medium. The addition of 1 µM AVP produced a 13-fold increase in adenylyl cyclase activity. Application of forskolin (100 µM) to IMCD cell membranes stimulated the adenylyl cyclase only slightly more (a 15-fold increase) than AVP, confirming both the robust expression of V₂R and the presence of a functionally intact signaling cascade from V₂R to the adenylyl cyclase.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Adenylyl cyclase activity in membranes derived from IMCD cells. IMCD cells were grown on type IV collagen coating for 6 days in continuous presence of DBcAMP (500 µM), followed by DBcAMP-free period of 18 h. Membranes were prepared as described in MATERIALS AND METHODS. Membrane preparations were left untreated or stimulated for 20 min with arginine-8 vasopressin (AVP) (1 µM) or forskolin (100 µM). Values are means ± SD derived from 3 independent experiments.

A characteristic feature of primary cultures is their relative inhomogeneity compared with monoclonal cell lines. Immunofluorescence laser confocal microscopy analysis with anti-AQP-2 antibodies on permeabilized IMCD cells revealed a clustered distribution of AQP-2-expressing cells in monolayers cultivated for 6 days on glass coverslips. About 50-70% of the cells expressed AQP-2. On a subcellular level, a punctate intracellular staining was observed 18 h after the withdrawal of DBcAMP, without subsequent stimulation (Fig. 5A). After stimulation with forskolin (Fig. 5B), an intense plasma membrane staining was observed, indicating the redistribution of AQP-2-bearing vesicles to the cell membrane. To confirm the involvement of the V₂R, AVP (1 µM; Fig. 5C) and a combination of AVP (1 µM) and SR-121463A (100 µM; Fig. 5D), a selective V₂R antagonist (28), were added 18 h after the withdrawal of DBcAMP. AVP stimulation elicited a similar result to that of forskolin stimulation. The V₂R antagonist abolished the AVP-induced redistribution of AQP-2 to the cell membrane. The membrane staining pattern in Fig. 5, B and C, suggested that the AQP-2 redistribution was directed to the lateral membrane compartment of the cells. Z sections (Fig. 5E), however, revealed that AQP-2 is delivered to both the lateral and apical membrane compartments of IMCD cells. We performed immunofluorescence experiments with 17 IMCD cultures (out of a total of 46 successfully grown IMCD primary cultures) and consistently found that 50-70% of the IMCD cells expressed AQP-2. The morphologies of AQP-2-expressing and -nonexpressing cells are very different when analyzed by phase-contrast microscopy and fluorescence microscopy with immunostained tight junctions to visualize cell borders (not shown). We found, generally, two morphologically distinct epithelial cell types. AQP-2-expressing epithelial cells showed long, straight cell borders, a nucleus embedded into the cytoplasm with one to three nucleoli, and several prominent granules within the cytoplasm. Epithelial cells that did not express AQP-2 showed very curved cell borders with indentations and invaginations to their neighboring cells and a protruding nucleus with one nucleolus, as indicated by a strong phase-contrast signal around the nucleus. These two morphologically distinct cell types occurred regularly in a clustered distribution as mentioned, whereby small islands of non-AQP-2-expressing cells were present in the large layer of AQP-2-expressing cells (i.e., principal cells). The relative areas covered by the two cell types were fairly constant, consistent with the small interexperimental variations in AQP-2 levels. Very few single fibroblasts were observed. Because of the different morphologies of AQP-2-expressing and -nonexpressing epithelial cells, we assume that the nonexpressing cells are intercalated cells.

View larger version (108K): [in this window] [in a new window]   Fig. 5.   Laser confocal microscopy analysis of IMCD cells. Cells were cultivated for 6 days in continuous presence of 500 µM DBcAMP on type IV collagen-coated glass coverslips. After DBcAMP-free period of 18 h, cells were incubated for 15 min with buffer (A), 100 µM forskolin (B), 1 µM AVP (C), or 1 µM AVP and 100 µM SR-121463A (a specific vasopressin V2 receptor antagonist) (D). Cells were labeled with protein G Sepharose-purified AQP-2 antiserum. Cells treated with unfractionated preimmune serum did not stain (data not shown). Scale bar, 10 µm. E: Z sections after forskolin application (100 µM) demonstrate the presence of AQP-2 in both basolateral and apical cell membranes.

We confirmed the redistribution of AQP-2 by using a biochemical approach (20). Eighteen hours after withdrawal of DBcAMP, IMCD cells were stimulated for 20 min with AVP (1 µM) or left untreated (buffer). After the preparation of LS, enriched for cell plasma membranes, and HS, enriched for intracellular vesicles, Western blot analysis was performed (Fig. 6). Freshly prepared IM served as a control. Comparison of the AQP-2 signal in LS and HS, from stimulated versus unstimulated IMCD cells, demonstrated a shift of AQP-2 from the intracellular vesicles to the plasma membrane and thus confirmed the data obtained by immunofluorescence microscopy.

View larger version (54K): [in this window] [in a new window]   Fig. 6.   Western blot analysis of IMCD cell fractions enriched for cell plasma membranes (LS, low-speed pellet) or intracellular vesicles (HS, high-speed pellet). Cells were grown for 6 days in continuous presence of DBcAMP (500 µM) on type IV collagen. After DBcAMP-free period of 18 h, cells were stimulated for 20 min with AVP (1 µM) or left untreated. Rat inner medulla homogenate (IM) was used as control. Left to right: lanes 2 and 3 show LS and HS derived from AVP-stimulated IMCD cells, respectively; lanes 4 and 5 show LS and HS derived from untreated IMCD cells, respectively. Proteins were analyzed by SDS-PAGE/immunoblotting with AQP-2 antiserum. Filter-bound antibodies were visualized with alkaline phosphatase-conjugated goat anti-rabbit IgG and BCIP/NBT. Equal amounts of protein (30 µg/lane) were used. Arrows indicate nonglycosylated (ng) and glycosylated (g) forms of AQP-2. In membrane preparations from cultured IMCD cells, an additional 17-kDa protein was detected by AQP-2 antiserum. This protein was recently identified as histone H2A1 (12).

Taken together, the data show that the primary cultured cells have preserved the key functions of principal cells in situ, i.e., expression of AQP-2 and redistribution of AQP-2 to the cell membrane in response to AVP or forskolin.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

AVP stimulates the long-term regulation of AQP-2 expression in vivo (5, 16). This finding suggests the involvement of cAMP in the regulation of AQP-2 expression. Consistent with this assumption, a cAMP-responsive element was identified in the promoter region of the human (29) and murine (21) AQP-2 genes. We illustrate here that cAMP is sufficient for the retention of rat AQP-2 expression in vitro.

An explanation of the observed biological effects of type IV collagen coating, namely, faster cell growth to confluence and the prolonged expression of AQP-2, cannot be offered as yet. Signaling pathways from the extracellular matrix (ECM) into the cell via integrins are well documented (10); in this case, however, the identity of the cellular components that interact with the ECM and lead to retardation of the downregulation of AQP-2 expression remains unclear.

The immunofluorescence studies indicate a redistribution of AQP-2 after stimulation to the lateral compartment, as well as to the apical membrane compartment, of IMCD cells. In LLC-PK₁ cells stably transfected with AQP-2 (13), and in AQP-2-overexpressing mouse IMCD cell monolayers (17), an almost exclusive sorting to the basolateral membrane has been reported. In contrast, stably transfected rabbit cortical collecting duct cells (30) and stably transfected MDCK cells (strain I) (4) exhibit apical sorting of the overexpressed AQP-2 protein. Several authors (6, 20, 23) have presented photomicrographs of cryosections of rat kidney inner medulla, which show the localization of AQP-2 in both apical and basolateral membrane compartments of principal cells. Thus the subcellular distribution of AQP-2 in primary cultured IMCD cells (apical and lateral plasma membranes) is reminiscent of that in principal cells in situ, i.e., in the kidney inner medulla.

The culture conditions used afford radical advantages for the in vitro study of AQP-2 regulation, as its natively harbored promoter appeared sufficiently strong to override the known downregulation of the protein under cell culture conditions in the absence of a transfected foreign promoter. Thus the cell culture model presented here, in contrast to those previously used (which afford information regarding trafficking only) (4, 13, 30), also facilitates the study of the regulation of AQP-2 at the gene level. The cellular model described here may also prove useful in the development of new therapeutic approaches to diuretic states.