Calmodulin-sensitive adenylyl cyclases mediate AVP-dependent cAMP production and Cl⁻ secretion by human autosomal dominant polycystic kidney cells

Abstract

In autosomal dominant polycystic kidney disease (ADPKD), binding of AVP to the V2 receptor (V2R) increases cAMP and accelerates cyst growth by stimulating cell proliferation and Cl⁻-dependent fluid secretion. Basal cAMP is elevated in human ADPKD cells compared with normal human kidney (NHK) cells. V2R mRNA levels are elevated in ADPKD cells; however, AVP caused a greater increase in global cAMP in NHK cells, suggesting an intrinsic difference in cAMP regulation. Expression, regulatory properties, and receptor coupling of specific adenylyl cyclases (ACs) provide temporal and spatial regulation of the cAMP signal. ADPKD and NHK cells express mRNAs for all nine ACs. Ca²⁺-inhibited ACs 5 and 6 are increased in ADPKD cells, while Ca²⁺/CaM-stimulated ACs 1 and 3 are downregulated. ACs 1, 3, 5, and 6 were detected in cyst cells in situ, and codistribution with aquaporin-2 suggests that these cysts were derived from collecting ducts. To determine the contribution of CaM-sensitive ACs to AVP signaling, cells were treated with W-7, a CaM inhibitor. W-7 decreased AVP-induced cAMP production and Cl⁻ secretion by ADPKD cells. CaMKII inhibition increased AVP-induced cAMP, suggesting that cAMP synthesis is mediated by AC3. In contrast, CaM and CaMKII inhibition in NHK cells did not affect AVP-induced cAMP production. Restriction of intracellular Ca²⁺ switched the response in NHK cells, such that CaM inhibition decreased AVP-induced cAMP production. We suggest that a compensatory response to decreased Ca²⁺ in ADPKD cells switches V2R coupling from Ca²⁺-inhibited ACs 5/6 to Ca²⁺/CaM-stimulated AC3, to mitigate high cAMP levels in response to continuous AVP stimulation.

autosomal dominant polycystic kidney disease (ADPKD) is characterized by the presence of numerous fluid-filled cysts that disrupt the normal renal architecture, leading to vascular damage, extensive nephron loss, interstitial fibrosis, and progressive decline of renal function. ADPKD is caused by mutations in PKD1 (85% of the cases) or PKD2 (15%), genes that encode polycystin-1 (PC1) and polycystin-2 (PC2), respectively (35, 60). PC1 is a large transmembrane protein with extracellular domains involved in cell-cell and/or cell-matrix interactions. PC2, also called TRPP2, is an integral protein with six transmembrane domains that functions as a Ca²⁺-permeable cation channel (13). PC1 and PC2 interact to form a multifunctional signaling complex involved in intracellular Ca²⁺ signaling and epithelial cell development and repair (22, 58). Functional loss of the polycystins disrupts intracellular Ca²⁺ signaling and decreases steady-state Ca²⁺ levels, which transform tubule epithelial cells into poorly differentiated cells characterized by aberrant cell proliferation (33, 69).

The extraordinary appearance of ADPKD kidneys is due to the accumulation of fluid within hundreds or thousands of cysts caused by fluid secretion (20, 60). cAMP stimulates net fluid secretion driven by transepithelial Cl⁻ secretion involving the coordinated function of transporters and ion channels within the apical and basolateral membranes (49, 60). Chloride enters the cell through basolateral NKCC1, an electrically neutral Na⁺-K⁺-2Cl⁻ cotransporter that brings these ions into the cell using the transmembrane Na⁺ gradient. The basolateral Na⁺-K⁺-ATPase pumps Na⁺ out of the cells, and K⁺ channels provide an efflux mechanism for K⁺. The net effect is an increase in intracellular Cl⁻ above its electrochemical gradient, keeping Cl⁻ poised for rapid efflux across the luminal membrane with cAMP activation of CFTR Cl⁻ channels (5, 21, 61). The apical Cl⁻ conductance and basolateral K⁺ conductance create a lumen-negative transepithelial electrical potential that drives passive Na⁺ transport through the paracellular pathway. The net addition of Na⁺ and Cl⁻ into the luminal fluid drives the osmotic movement of water into the cyst cavity (5, 49, 61).

Intracellular cAMP is regulated by the activities of adenylyl cyclases (ACs), which catalyze the formation of cAMP from ATP, and phosphodiesterases (PDEs), which degrade cAMP to AMP. Cellular specificity and cellular compartmentalization are important features of cAMP signaling. Compartmentalization of the cAMP signal relies on localization of ACs at the plasma membrane and A kinase-anchoring proteins (AKAPs), which hold PKA to specific compartments in close proximity of the receptor, AC, phosphodiesterases, and effector molecules (10, 14, 47). Binding of AVP, an important antidiuretic hormone, to the V2 receptor (V2R) stimulates cAMP production by adenylyl cyclases (ACs) in cells of the collecting duct and distal nephron, predominant sites for renal cyst formation (53, 59).

In mammals, there are nine closely related membrane-associated ACs. Regulatory properties and tissue distribution of AC isoforms are important for specificity and compartmentalization of the cAMP signal (12, 27, 50). ACs 1, 3, and 8 are stimulated by Ca²⁺/calmodulin (CaM), whereas ACs 5 and 6 are inhibited by Ca²⁺ in a CaM-independent manner. ACs 2, 4, 7, and 9 are insensitive to Ca²⁺ (9, 36). The functional role of specific AC isoforms has been difficult to establish because of low AC expression and the lack of isoform-specific inhibitors (2, 4, 23, 26, 45, 48). mRNA and protein expression for almost all AC isoforms have been found in rat and mouse kidneys (2, 4, 26, 48). AC5 and AC6 are thought to be the primary ACs that mediate the effect of AVP on intracellular cAMP in collecting ducts (6, 25); however, recent studies have suggested that AVP-dependent AC activation and cAMP accumulation are regulated, in part, by CaM-stimulated ACs (26, 48, 51). AC3, a CaM-sensitive AC isoform, was shown to be expressed in rat and mouse medullary collecting ducts (26, 48), and knockdown of either AC3 or AC6 using small interfering RNA (siRNA) in primary cultured mouse inner medullary collecting duct (IMCD) cells significantly reduced AVP-stimulated cAMP accumulation (48). AC3 knockout mice have a significant decrease in glomerular filtration rate compared with control animals; however, water balance and urine osmolarity were normal (39, 42). AC6^−/− mice, on the other hand, have impaired ability to concentrate urine (7, 42, 43). These results with knockout mice raise questions on the physiological relevance of AC3 relative to AC6 on regulating the AVP-mediated regulation of intracellular cAMP and water permeability of the collecting duct in mice.

In the present study, we examined AC isoform expression in human ADPKD and normal human kidney (NHK) cells and tissues and determined the effects of intracellular Ca²⁺ and CaM inhibition on cAMP production and transepithelial Cl⁻ secretion in response to AVP. The results show that intracellular Ca²⁺ regulates specific AC isoforms associated with V2R and determines the magnitudes of AVP-dependent cAMP accumulation and Cl⁻ secretion by ADPKD cells.

MATERIALS AND METHODS

Primary Cultures of ADPKD and Normal Human Kidney Cells

ADPKD kidneys were obtained from hospitals participating in the Polycystic Kidney Research Retrieval Program with the assistance of the PKD Foundation (Kansas City, MO) and from the Biospecimen Shared Resource at the Kansas University Medical Center (KUMC). ADPKD primary cultures were prepared by the PKD Research Biomaterials and Cellular Models Core at KUMC as previously described (61, 71–73). Briefly, superficial cysts were cleaned of adherent fibrous tissue and fat, and domes of the cysts were cut free and digested in DMEM/F12 containing type IV collagenase (220 IU/ml, Worthington Biochemical, Lakewood, NJ) to loosen the epithelial cells from the stromal tissue. The resultant cell suspension was centrifuged, and the pellet was rinsed and plated in T75 flasks containing DMEM/F12 supplemented with 5% FBS (Hyclone, Logan, UT), penicillin/streptomycin (P/S), and 5 μg/ml insulin, 5 μg/ml transferrin, and 5 ng/ml sodium selenite (ITS; Collaborative Biomedical Products, Bedford, MA) until cells were ∼75% confluent.

NHKs unsuitable for transplantation were obtained from the Midwest Transplant Network (Kansas City, KS). The cortex was minced, and cells were released by digestion in collagenase and incubated in DMEM/F12 supplemented with P/S, ITS, and 5% FBS until they were ∼75% confluent. ADPKD and NHK cells were lifted from the plastic flasks and either frozen in culture medium containing 10% DMSO for storage in liquid N₂ or seeded directly for experiments. The protocol for the use of surgically discarded kidney tissues complies with Federal regulations and was approved by the Institutional Review Board at KUMC. Primary human ADPKD cells have provided an excellent model for the investigation of cellular mechanisms involved in the pathogenesis of PKD, particularly cAMP-induced cell proliferation and Cl-dependent fluid secretion (60). Previous studies have indicated that freshly isolated collecting ducts have a greater cAMP response to AVP and CaM inhibition compared with cultured collecting duct cells, indicating that cultured cells may underestimated the amplitude of the cAMP response; however, the relative contribution of specific ACs appears to be consistent between the two preparations (48). In the current study, cells were passaged a total of only three times to limit the loss of epithelial characteristics.

RNA Extraction and Quantitative Real-Time PCR

To compare AC isoforms, NHK and ADPKD cells were plated on plastic petri dishes (100-mm diameter) and grown in media containing 1% FBS for 3 days. The media was changed to 0.002% FBS for 24 h before RNA extraction. Surface cysts of ADKPD kidneys and cortical tissues from NHK kidneys were snap frozen and stored at −80°C. Total RNA was isolated and cDNA was synthesized as previously described (62). In brief, total RNA was isolated from 30 mg of frozen tissue or cells grown on a 100-mm petri dish using an RNA extraction buffer containing β-mercaptoethanol and precipitated in 70% EtOH. Total RNA was loaded onto the column of an RNeasy Mini Kit (Qiagen) and eluted with RNase-free water. RNA quality was verified using an Agilent Bioanalyzer (Santa Clara, CA). RNA from human whole brain (Clontech, Palo Alto, CA) was used as a positive control for AC isoform expression. cDNA was synthesized from total RNA using the Invitrogen SuperScript III first-strand synthesis system for RT-PCR. Primers for all AC isoforms (68) were purchased from IDT (Coralville, IA), and PCR was performed using a SmartCycler (Cepheid, Sunnyvale, CA) as previously described (68). In brief, PCR was performed for 40 cycles. Denaturation and extension temperatures/times were 95°C/30 s and 68°C/min, respectively. One-minute annealing times were used for all reactions at 52°C for isoforms 1, 2, and 8, at 50°C for isoform 3, at 55°C for isoforms 4, 5, and 7, and at 58°C for isoforms 6 and 9 (68). PCR products were separated on nondenaturing 2% polyacrylamide gels using a 100-bp DNA ladder (Promega) as a DNA size standard. Gels were stained with ethidium bromide and visualized using a Fluor-S Multi-Imager (Bio-Rad, Hercules, CA). PCR products were isolated using a gel extraction kit (Qiagen, Valencia, CA) according to the manufacturer's protocol. PCR products were sequenced using automated sequence analysis at the DNA sequencing core facility at KUMC and sequences confirmed by a BLAST search. Relative expression of ACs and V2R in ADPKD and NHK cells and native tissues (n =5) was determined by SYBR green real-time PCR using a SmartCycler (Cepheid). The threshold cycle (Ct), the cycle number at which amplification crossed a fixed threshold, was determined for all the ACs, V2R, and GAPDH, a housekeeping gene. Change in relative expression was calculated using a comparative CT method as previously described (62). AC primers were purchased from IDT (Coralvill, IA) and the PCR reactions were performed as described above. GAPDH and V2R primers were obtained from SuperArray (Frederick, MD), and PCR was performed according to the manufacturer's instructions. AC, V2R, and GAPDH primers have similar efficiency (E), ranging from 1.9 to 2.0, where E = [10(−1/slope)].

Western Blot Analysis

ADPKD and NHK cells (0.5× 10⁶) were seeded onto plastic petri dishes (100-mm diameter) containing DMEM/F12 medium with 1% FBS. The serum was reduced to 0.002% when the cells were ∼75% confluent, and then cells were allowed to grow for an additional 24 h. Cell lysates were prepared, and total protein content was determined using the dye-binding assay from Bio-Rad (71). Tissues (0.5 g) were homogenized with the aid of a mechanical homogenizer (Polytron Brinkman, Westbury NY) in ice-cold lysis buffer containing a protease inhibitor cocktail and phosphatase inhibitors. The homogenates were centrifuged for 10 min at 10,000 g to remove debris and insoluble material, and supernatants were assayed for protein content using a dye-binding assay from Bio-Rad.

AC expression was detected by Western blot analysis using validated commercially available isoform-specific antibodies. Antibodies to AC1 (V-20), AC3 (C-20), and AC5/6 (C-17) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). We tested AC8 (H-270; Santa Cruz Biotechnology) by Western blotting, but the antibody detected multiple bands and was not used for analysis. While these antibodies had previously been characterized and have been widely used to detect AC expression (17, 26, 40, 48), we validated them using appropriate positive control cell lysates for AC1 (IMR-32 cells, sc-2409), AC3 (SK-N-SH cells, sc-2410), and AC5/6 (HISM cells, sc-2229). Proteins were visualized by using an enhanced chemiluminescence system (ECL; Amersham Biosciences). Band intensity was detected and quantified by the Fluor-S MAX Multi-Imager system. Western blot analysis using 100 μg of cell or tissue lysate was performed as previously described (34, 73). Immunoblots were probed with antibodies for AC1, AC3, and AC5/6 at 1:200 dilutions. Blots were reprobed with an anti-GAPDH antibody (1:1,000; Abcam, Cambridge, MA) to verify equal protein loading. Secondary antibodies (anti-rabbit or anti-mouse IgG conjugated to horseradish peroxidase) were purchased from Santa Cruz Biotechnology. To enhance the signal intensity of ACs with low expression levels, we utilized a SuperSignal Western blot enhancer (Thermo Scientific, Rockford, IL). The bands for each AC were of the proper molecular mass as those reported in the literature and indicated on the product data sheet.

Immunohistochemistry

ADPKD and NHK tissues were fixed in 4% paraformaldehyde, embedded in paraffin wax, and sectioned for immunocytochemistry as previously described (62). Briefly, 5-μm sections were deparaffinized and rehydrated in Reveal (Biocare Medical, Concord, CA) with a Decloaker. Endogenous peroxidase activity was destroyed by incubating sections in 3% H₂O₂ in methanol for 10 min, and then the sections were incubated with isoform-specific AC antibodies (1:100 for ACs 1, 3, and 5/6) or with an aquaporin-2 antibody (ab15081; 1:2,000) ± excess blocking peptides overnight at 4°C. Sections were rinsed three times with PBS, and the antigen was detected using a Zymed SuperPicTure polymer detection kit (San Francisco, CA). Sections were then counterstained with a hematoxylin solution. Images of ADPKD and NHK sections were captured using a ×20 objective (bar = 100 μm).

Measurements of Intracellular cAMP

ADPKD and NHK cells (5 × 10⁴) were seeded into individual chambers (1.9 cm²) of 24-well plates in DMEM/F12 containing 5% FBS, ITS, and P/S (3). After 48 h, the medium was changed to 1% FBS to reduce the rate of basal growth, and the cells were allowed to grow for an additional 5 days. Serum was reduced to 0.002% for 24 h before the start of the experiments. Since it is not possible to measure cAMP and protein in the same monolayer, we plated two groups of monolayers and treated the monolayers identically. One group was used to determine cAMP content, and the other was used to determine protein content. ADPKD and NHK cells were treated with 10⁻⁸ M AVP or 1 μM forskolin (FSK) for 15 min. W-7 (50 μM) was added for 30 min alone or 15 min before the addition of AVP or FSK. Experiments were performed in either the presence or absence of IBMX (100 μM), a broad-spectrum PDE inhibitor.

To examine the effect of W-7 in mouse cells, M1 cells were seeded (7.3 × 10⁵ cells/0.332 cm²) on cell culture supports (6.5-mm diameter, Transwell-COL, CoStar, Cambridge, MA) in DMEM/F12 containing 5% FBS and P/S and grown as confluent monolayers (3). M1 cells were treated with 10⁻⁷ M AVP or 1 μM FSK for 15 min. W-7 (50 μM) was added for 30 min alone or 15 min before the addition of AVP.

To examine the effect of Ca²⁺ restriction on cAMP accumulation (see Fig. 8), NHK cells were treated with either 10 μM verapamil, an L-type Ca²⁺ channel blocker, for 24 h or 10⁻⁸ M BAPTA, a Ca²⁺ chelator, for 30 min. Stock solutions for AVP, FSK, and verapamil were prepared in water. IBMX, W-7, and BAPTA were dissolved in DMSO. Equivalent concentrations of DMSO were added to the media of control and treatment groups. Each treatment was performed in quadruplicate. After the treatment, the media were removed and an 80% methanol-20% water mixture was used to extract cAMP from the cells. cAMP content was determined with an enzyme immunoassay system as previously described (3) and according to the recommended manufacturer's protocol (Amersham Pharmacia Biotech, Little Chalfont, UK). cAMP levels were normalized to total cellular protein expressed as picomoles cAMP per milligram protein (see Fig. 5B) or expressed as picomoles cAMP per monolayer, yielding qualitatively similar results. Total protein content was similar between ADPKD and NHK cells. Treatments with DMSO and W-7 ± AVP did not alter total protein content, confirming the low toxicity of these drugs at the concentrations that were used.

Fig. 8.

Effect of intracellular Ca²⁺ restriction on CaM-sensitive AVP-mediated cAMP production in NHK cells. Pretreatment with 10 μM verapamil (VP) for 24 h or 10 nM BAPTA, a Ca²⁺ chelator, for 30 min did not significantly affect AVP-mediated cAMP accumulation in NHK cells (n = 4 monolayers/experiment). There was no significant effect of W-7 alone on AVP-induced cAMP production. By contrast, reduction in intracellular Ca²⁺ concentration by treatment with either BAPTA or verapamil switched the W-7 response, such that CaM inhibition decreased AVP-mediated cAMP production, similar to the effect of W-7 in ADPKD cells. Similar results of CaM inhibition were observed with 200 μM monodansylcadaverine (MDC) and 100 μM trifluoperazine (TFP), chemically unrelated CaM inhibitors (data not shown). #P < 0.001 compared with control treatment (Cont). *P < 0.001 compared with AVP alone, AVP+VP, or BAPTA.

Fig. 5.

Comparison of basal and AVP-induced cAMP levels in human ADPKD and NHK cells. ADPKD and NHK cells were plated into 6-well culture plates and treated with either control media or media containing 10 nM AVP. Intracellular cAMP was extracted from the cells using 80% methanol, and cAMP content was measured using the cAMP Biotrak Enzymeimmunoassay System (Amersham). A: cAMP content was corrected for total protein. Since is not possible to measure cAMP and protein in the same monolayer, we plated 2 plates of monolayers and treated the monolayers identically. One group was used to determine cAMP content, and the other was used to determine protein content. ADPKD and NHK cell monolayers had similar protein content (84 ± 1.7 and 117 ± 1.4 μg/well, respectively, n = 4, not statistically different). ADPKD cells had elevated cAMP and a greater cAMP response to AVP than NHK cells. B: similar results were obtained when cAMP levels were expressed as pmol/monolayer. Therefore, unless noted, we expressed cAMP as pmol/monolayer. #P < 0.001 compared with control. **P < 0.05 and $P < 0.001 compared with NHK cells with the same treatment.

Short-Circuit Current

ADPKD and NHK cells (2.5 × 10⁵) were plated on permeable Transwell supports (Snapwell, 12-mm diameter; CoStar). Snapwell supports containing confluent monolayers were inserted into modified Ussing chambers (Harvard Apparatus, Hollison, MA). The position of the monolayer and the supporting membrane separated the apical and the basolateral compartments. Both surfaces of the cell monolayer were bathed in a HCO₃⁻-Ringer solution maintained at 37°C and equilibrated in 5% CO₂ balanced with room air. Two dual-voltage clamp devices (Warner Instruments, Hamden, CT), each attached to four electrodes per chamber, were used to measure short-circuit current (I_SC) in four monolayers simultaneously (41). Benzamil (10 μM) was added to the apical medium for 15 min before the experiment to block the Na⁺ current. Vasopressin (10⁻⁸ M) or 1 μM FSK was added to the basolateral media until steady-state currents were achieved. W-7 (50 μM) was added to the basolateral media. I_SC measurements were carried out as previously described (61).

Chemicals

AVP, FSK, monodansylcadaverine (MDC), trifluoperazine (TFP), and verapamil were purchased from Sigma (St. Louis, MO). W-7, IBMX, and KN-93 were purchased from Calbiochem (San Diego, CA). BAPTA was purchased from Molecular Probes (Eugene, OR).

Statistics

Data are expressed as means ± SE. Statistical significance was determined by one-way analysis of variance and a Student-Newman-Keuls posttest for multiple comparisons or unpaired t-test for comparison between control and treated groups.

RESULTS

Expression of AC Isoforms in Human ADPKD and NHK Cells

To determine which AC isoforms are expressed in human ADPKD and NHK cells, we used RT-PCR to amplify mRNA of individual ACs using human isoform-specific primers, as described previously. The brain expresses all AC isoforms (68); therefore, mRNA from whole human brains (Clontech) was used as a positive control for AC mRNA expression. We found mRNA for all nine membranous ACs in ADPKD and NHK cells (Fig. 1A). Amplification of RNA from whole brains yielded products of the expected sizes (68). Primers for AC2, AC4, and AC8 resulted in amplification of more than one product, as shown previously (68). PCR bands of the expected sizes (68) were sequenced and confirmed by a BLAST search. Cortical tissue from NHK and individual cyst walls from ADPKD kidneys expressed all nine AC isoforms (data not shown), similar to the primary cells.

Fig. 1.

Expression of adenylyl cyclase (AC) isoforms in autosomal dominant polycystic kidney disease (ADPKD) and normal human kidney (NHK) cells. A: representative RT-PCR products (160–550 bp) using primers for human ACs 1–9 are shown for ADPKD and NHK cells. Human brain mRNA, which expresses all 9 isoforms, was used as a positive control for each primer set. Some primers amplified more than one product in the human brain (e.g., AC4) and in ADPKD and NHK cells (e.g., AC2 and AC8). Amplified products were confirmed by sequencing. The results were confirmed using mRNA from two additional ADPKD and NHK cell preparations. B: summary of quantitative real-time RT-PCR determination of AC expression in ADPKD cells relative to NHK cells (set to 1.0, dashed line). Total RNA was isolated from primary cultures of ADPKD and NHK cells. RNA was reverse transcribed into cDNA, and equal amounts of cDNA were amplified with a Cepheid Smart Cycler. AC expression was normalized to GAPDH, and relative fold-change in AC expression was calculated by the ΔΔC_T method, where C_T is the threshold cycle (62). Experiments were performed in three ADPKD and NHK cell preparations cultured from different kidneys.

To compare AC mRNA levels in cells derived from ADPKD and NHK, we used quantitative real-time RT-PCR. AC expression was normalized to GAPDH, and the relative fold-change in AC expression was calculated by the ΔΔCT method (62). Ca²⁺-inhibited ACs 5 and 6 were 2.5-fold higher in ADPKD cells compared with NHK cells, whereas Ca²⁺/CaM-stimulated ACs 1, 3, and 8 were decreased 40, 60, and 20%, respectively (Fig. 1B). There were no differences in expression levels of ACs 2 and 9. ACs 4 and 7 are isoforms that are not regulated by intracellular Ca²⁺. AC4 expression was increased twofold, while AC7 expression was decreased 60% in ADPKD cells compared with normal renal cells. Changes in AC 1, 3, and 5/6 expression in ADPKD cells compared with NHK cells were confirmed by Western blot analysis in both cell and tissue lysates (Figs. 2, A and B). To analyze the protein expression of Ca²⁺-inhibited ACs 5 and 6 and Ca²⁺/CaM-activated ACs 1 and 3, we used antibodies that have previously been used to characterize AC expression (17, 26, 40, 48). To further validate these antibodies, we used control cell lysates known to express ACs 1, 3, and 5/6 (data not shown). AC5 and AC6 have high homology, and a single antibody recognizes both isoforms (AC5/6). Even though we found a decrease in AC1 mRNA levels in ADPKD cells, there were no differences in AC1 protein (98 kDa) expression in either cells or tissues. AC3 (160 kDa) expression was slightly decreased in cells (0.8 ± 0.05 vs. 1.0 ± 0. 09, n = 4) and tissues (0.8 ± 0.10 vs. 1.0 ± 0.04, n = 2) from ADPKD kidneys compared with NHK kidneys, respectively (Fig. 2, A and B). By contrast, protein levels for AC5/6 (132 kDa) were significantly increased in ADPKD cells (1.3 ± 0.03 vs. 1.0 ± 0.02, n = 4, P < 0.05) and tissues (2.4 ± 0.44 vs. 1.0 ± 0.03, n = 3, P < 0.001) compared with NHK tissue.

Fig. 2.

Protein expression of ACs 1, 3, and 5/6 in ADPKD and NHK cells and tissues. Lysates (100 μg/well) of ADPKD and NHK cells (A) and tissues (B) were loaded into 7.5% polyacrylamide gels for Western blot analysis. Proteins were separated by electrophoresis and then transferred to a nitrocellulose membrane. Membranes were blotted with 5% milk and probed with well-characterized AC isoform-specific antibodies (1:200 dilution). Bands were visualized using enhanced chemiluminescence (ECL), and the protein size was estimated from the relative position of molecular weight markers using a Fluor-S MAX imager (Bio-Rad). ACs 5 and 6 share high amino acid sequence homology; therefore, a single antibody detected both isoforms. Antibodies to ACs 1, 3, and 5/6 detected bands of the predicted molecular masses as reported in the literature and in product data sheets (Santa Cruz Biotechnology). Bar graphs represent AC expression, normalized to GAPDH (AC/GAPDH), for ADPKD (black bars) and NHK (white bars). For cells (A), n = 4 experiments from 4 different NHK and ADPKD kidneys and for tissues (B), n = 3 experiments from 2 different NHK and ADPKD kidneys were performed. Representative blots are shown below the corresponding bar graphs.*P < 0.05 compared with NHK.

Next, we examined AC isoform expression and localization in ADPKD tissue sections by immunostaining (Fig. 3). ACs 1, 3, and 5/6 were detected in cyst-lining epithelial cells; however, the staining was variable among different cysts even within a single kidney and even more so between kidneys from different patients. In larger cysts, the cyst-lining epithelial cells stained multiple AC isoforms as well as aquaporin-2 (AQP2), suggesting that these cysts were derived from collecting ducts (15, 24). As previously reported in normal kidneys (6, 26), ACs 1, 3, and AC5/6 appeared to have similar distribution within multiple tubular segments, including the collecting ducts as demonstrated by parallel staining with AQP2 (Fig. 4, A–H). Immunostaining for ACs was eliminated by preincubation of the antibody with the appropriate blocking peptide. Similar patterns of staining were detected in two additional ADPKD and NHK sections. These data demonstrate that CaM-sensitive ACs 1 and 3, as well as Ca²⁺-inhibited AC5/6, are expressed in human ADPKD cysts and tubules of normal kidneys.

Fig. 3.

ACs 1, 3, and 5/6 immunostaining of sections of ADPKD kidneys. ADPKD tissue sections were incubated with isoform-specific AC antibodies (1:100) for ACs 1, 3, and 5/6 or with an aquaporin-2 antibody (AQP2; 1:2,000) ± excess blocking peptides (BP) overnight at 4°C. Representative staining showed that AC1 (A), AC3 (C), and AC5/6 (E) were detected in cyst-lining epithelial cells, indicated by arrows. AQP2 (G) colocalized with ACs 1, 3, and 5/6 in cyst-lining epithelial cells, suggesting that collecting duct-derived cysts express multiple AC isoforms. Staining was variable among different cysts even within a single kidney and even more so between kidneys from different patients. Immunostaining for ACs 1, 3, and 5/6 and AQP2 was eliminated by preincubation of the antibody with excess BP (B, D, F, and H). A similar pattern of staining was detected in 2 additional ADPKD kidneys. Images were captured using a ×20 objective (scale bar = 100 μm).

Fig. 4.

ACs 1, 3, and 5/6 immunostaining of sections of NHK. Tissue sections of NHK were incubated with antibodies (1:100) to ACs 1, 3, and 5/6 or AQP2 (1:2,000) ± excess BP overnight at 4°C. ACs 1, 3, and AC5/6 (A, C, and E, respectively) had similar distribution within multiple tubule segments, including the collecting ducts, as indicated by AQP2 staining (G). Immunostaining for ACs and AQP2 was eliminated by preincubation with the BP (B, D, F, and H). Similar patterns of staining were detected in 2 additional NHK. Images were captured using a ×20 objective (scale bar = 100 μm). Glomeruli are denoted by asterisks (*). Arrows indicate collecting ducts.

Effect of CaM Inhibition on AVP-Stimulated cAMP Accumulation in ADPKD and NHK Cells

Renal cAMP is elevated in PKD animals, including pcy mice, jck mice, PCK rats, and Pkd2^WS25/− mice (18, 46, 57, 70). It has been proposed that increased intracellular cAMP in PKD kidneys is caused by decreased intracellular Ca²⁺ concentration, secondary to mutations in the PKD genes, resulting in higher activity of Ca²⁺-inhibited AC isoforms (ACs 5 and 6) (54). We found that human ADPKD cells have a significantly higher basal cAMP level compared with NHK cells (9.1 ± 0.7 vs. 4.7 ± 1.5 pmol/mg of protein, n = 2, P < 0.05) (Fig. 5A) or 1.1 ± 0.1 vs. 0.7 ± 0.1 pmol/monolayer, n = 9, P < 0.01 (Fig. 5B). These results suggest that there are intrinsic differences in cAMP regulation between cystic and normal cells. AVP significantly increased intracellular cAMP in ADPKD cells from 4.7 ± 1.5 to 127.9 ± 7.6 pmol/mg protein (P < 0.001) (Fig. 5A). Surprisingly, AVP consistently had a greater effect on global cAMP in NHK cells compared with ADPKD cells (198.2 ± 9.6 and 127.9 ± 7.6 pmol/mg protein P < 0.001, respectively) (Fig. 5A). ADPKD and NHK cells had similar protein content (84.0 ± 1.7 and 117.0 ± 1.4 μg/well, n = 4 wells). Consequently, cAMP levels normalized for protein content (Fig. 5A) or expressed as picomoles/monolayer (Fig. 5B) were qualitatively the same. In the remaining experiments, cAMP levels were expressed as picomoles cAMP per monolayer (see Figs. 6–8 and Table 1).

Fig. 6.

Effect of calmodulin (CaM) inhibition on basal, AVP-, and forskolin (FSK)-mediated cAMP production in ADPKD and NHK cells. ADPKD and NHK cells were plated into 6-well culture plates and treated with either control media or media containing 10 nM AVP or 1 μM FSK ± 50 μM W-7. Intracellular cAMP was extracted from the cells using 80% methanol, and cAMP content was measured using the cAMP Biotrak Enzymeimmunoassay System (Amersham). A: pretreatment with 50 μM W-7 for 15 min significantly reduced both AVP- and FSK-mediated cAMP accumulation in ADPKD cells (*P < 0.001 vs. AVP and †P < 0.001 vs. FSK) but did not have a significant effect in NHK cells. The numbers of experiments are indicated in parentheses, ranging from 4 to 9 experiments (4 monolayers/experiment). B: IBMX, a global phosphodiesterase (PDE) inhibitor, was added during the experiment to eliminate the possible role of CaM-sensitive PDE regulation of cAMP levels. Basal cAMP was elevated and both the AVP- and FSK-induced cAMP increased to a greater extent in the presence of IBMX. The effect of W-7 treatment was similar in the presence of IBMX, indicating that the difference in the effect of CaM inhibition between normal and cystic cells was due to cAMP synthesis. Overall, cAMP levels were higher in the presence of IBMX in both ADPKD and NHK cells. Statistical comparisons between groups in the presence and absence of IBMX are indicated in Table 1. #P < 0.001 compared with control treatment (Cont). *P < 0.001 compared with AVP alone. †P < 0.001 compared with FSK alone.

Table 1.

Effect of CaM inhibition on basal, AVP-, and FSK-stimulated cAMP levels in ADPKD and NHK cells

	ADPKD		NHK
	No IBMX	IBMX	No IBMX	IBMX
Control	1.1 ± 0.08** (9)	7.4 ± 0.8@ (4)	0.6 ± 0.06 (9)	7.3 ± 0.8@ (4)
W-7	0.6 ± 0.05## (6)	3.5 ± 0.7## (4)	0.74 ± 0.07 (8)	6.8 ± 0.8& (4)
AVP	11.8 ± 1.04# (9)	18.7 ± 2.8# @ (4)	21.28 ± 1.9# $ (9)	26.6 ± 2.3# $$ (4)
W-7 + AVP	5.4 ± 0.55* (6)	8.02 ± 1.6* (4)	23.5 ± 2.5 (4)	27.9 ± 1.5 (4)
FSK	21.7 ± 1.1# (5)	27.9 ± 1.5# & (4)	23.6 ± 2.4# (6)	33.2 ± 1.6# & (4)
W-7 + FSK	7.4 ± 0.45† (4)	10.3 ± 0.7† (4)	18.4 ± 2.1 (4)	28.3 ± 2.7& (4)

Values are means ± SE of cAMP levels (pmol/monolayer) of autosomal dominant polycystic kidney disease (ADPKD) and normal human kidney (NHK) cells under different experimental conditions. FSK, forskolin. The number of experiments is indicated in parentheses, ranging from 4 to 9 (4 monolayers/experiment). Statistical significance was determined by 1-way analysis of variance and Student-Newman-Keuls posttest for multiple comparisons or unpaired t-test for comparison between control and treated groups in the presence or absence of IBMX.

^**P < 0.01, compared with control-treated NHK cells (No IBMX).

^*P < 0.001, AVP+W-7 compared with AVP alone. $P < 0.001, compared with the effect of AVP between ADPKD cells and NHK (No IBMX). $$P < 0.01, compared with the effect of AVP between ADPKD and NHK cells in the presence of IBMX. #P < 0.001 and ##P < 0.05, compared with corresponding control group.

^†P < 0.001, FSK+W-7 vs. FSK alone. @P < 0.001 and &P < 0.05, compared with the same treatment in the presence and absence of IBMX.

To determine whether the difference in the amplitude of the cAMP response to AVP between ADPKD and NHK cells was due to differences in PDE activity, we compared the effects of AVP in the presence and absence of IBMX. AVP-mediated cAMP levels were greater in NHK cells compared with ADPKD in either the absence (21.2 ± 1.7 vs. 11.8 ± 1.04 pmol/monolayer, P < 0.0001) or presence of IBMX (26.6 ± 2.3 vs. 18.7 ± 2.8 pmol/monolayer, P < 0.001) (Fig. 6, A and B). Despite a greater effect of AVP on global intracellular cAMP in NHK cells compared with ADPKD cells, we found that V2R mRNA levels were 1.4-fold greater in ADPKD cells than normal cells. Similarly, there was a twofold greater V2R mRNA expression in ADPKD tissues compared with NHK tissues.

To determine the relative contribution of CaM-sensitive ACs 1, 3, and 8 and CaM-insensitive ACs 5 and 6 on basal and AVP-induced cAMP production, we treated ADPKD and NHK cells with W-7, a CaM inhibitor (Fig. 6) (48, 51). W-7 (50 μM) was added for 15 min before the addition of 10⁻⁸ M AVP; then, cAMP was isolated after an additional 15 min. Treatment of ADPKD cells with W-7 reduced AVP-mediated cAMP levels by 55%, consistent with the role of CaM-stimulated ACs in AVP signaling (Fig. 6A). In addition, W-7 alone decreased basal cAMP levels in ADPKD cells (0.6 ± 0.5 pmol/monolayer vs. 1.1 ± 0.1, P < 0.05). By contrast, W-7 had no effect of AVP-induced cAMP production in NHK cells. The results are summarized in Table 1. CaM protein levels were not significantly different between NHK and ADPKD cells (1.0 ± 0.15 vs. 1.1 ± 0.12, n = 4); therefore, the difference in the W-7 effect was not due to differences in CaM expression. In addition, treatment with W-7 did not alter total protein content in ADPKD and NHK cells, confirming that the inhibitory effect of W-7 on AVP- and FSK-mediated cAMP levels in ADPKD cells (Fig. 6) was not due to cellular toxicity of W-7.

FSK directly activates all AC isoforms except AC9, with an EC₅₀ between 5 and 10 μM (44). Previously, W-7 was shown to significantly decrease cAMP production by a submaximal concentration of FSK (1 μM) in a rat IMCD suspension, acutely isolated mouse IMCDs, and primary mouse IMCD cells, suggesting that the majority of membrane-bound AC activity was CaM stimulated and that CaM inhibition was acting at the level of ACs instead of having an effect at the receptor or G protein level (26, 48). In the present study, 1 μM FSK increased intracellular cAMP from 1.1 ± 0.1 to 21.7 ± 1.2 pmol/monolayer (P < 0.001) in ADPKD cells, and incubation with W-7 caused a significant decrease in cAMP production (Fig. 6A and Table 1). The effect of W-7 on cAMP production by 1 μM FSK could indicate that CaM-activated ACs are more sensitive to this submaximal concentration of FSK than CaM-insensitive ACs (38). Thus inhibition of CaM would decrease the activity of this AC population. While we cannot rule out possible nonspecific effects of W-7 on AC activity, we found that chemically unrelated CaM inhibitors, MDC and TFP, also inhibited cAMP accumulation by 1 μM FSK. Maximal stimulation of AC activity with 10 μM FSK increased cAMP from 1.1 ± 0.1 to 40.7 ± 1.7, an effect that was insensitive to CaM inhibition (37.7 ± 1.7 pmol/monolayer; FSK + W-7). By contrast, CaM inhibition by W-7 did not cause a significant effect on basal, AVP-, or FSK-mediated cAMP accumulation in NHK cells. An economical interpretation of these findings is that there may be functional differences in the role of CaM-sensitive ACs in the AVP-mediated signaling between normal and cystic cells.

In contrast to the lack of an effect of W-7 on NHK cells, previous studies have shown that W-7 inhibits cAMP production by AVP and low concentrations of FSK in mouse and rat collecting duct cells (26, 48, 51), suggesting a species difference in the role of CaM-sensitive ACs. To examine the effect of CaM inhibition on a different species, we treated M1 cells, a mouse cortical collecting duct cell line, with AVP and FSK in the presence or absence of W-7. As expected, AVP and FSK increased cAMP levels in M1 cells (Fig. 7). CaM inhibition with W-7 significantly decreased AVP- and FSK-induced cAMP accumulation. These results support the idea that the relative contribution of CaM-stimulated ACs to AVP-induced cAMP production may be different between human and rodent.

Fig. 7.

Effect of calmodulin (CaM) inhibition on basal, AVP-, and FSK-mediated cAMP production in M1 cells. Cells were seeded on cell culture supports and grown as confluent monolayers as previously described (6). Cells were grown in 5% FBS for 10 days, serum was removed for 24 h, and then cells were treated with 100 nM AVP or 1 μM FSK in the presence or absence of 50 μM W-7. cAMP content was determined using an enzyme immunoassay system. Bar graphs represent means ± SE; n = 8 monolayers from 2 independent experiments. AVP increased cAMP levels in M1 cells from 1.0 ± 0.3 to 6.1 ± 0.9 pmol/monolayer (#P < 0.001). CaM inhibition with W-7 decreased AVP-mediated cAMP from 6.1 ± 0.9 to 1.0 ± 0.3 pmol/monolayer (*P < 0.001). W-7 also decreased FSK-mediated cAMP levels (†P < 0.05).

Since changes in intracellular cAMP by W-7 could be due to inhibition of Ca²⁺/CaM-activated PDEs (19, 64), we tested the effect of W-7 in the presence of 100 μM IBMX. In both ADPKD and NHK cells, treatment with IBMX significantly increased basal cAMP levels and amplified AVP- and FSK-mediated cAMP levels (Table 1). Overall, the effect of W-7 on basal, AVP-, and FSK-induced cAMP accumulation was similar in the presence and absence of IBMX (Fig. 6, A and B, and Table 1), consistent with the hypothesis that cAMP synthesis is largely mediated by CaM-dependent AC isoforms in ADPKD cells, but not NHK cells.

Steady-state intracellular Ca²⁺ concentrations were found to be lower in ADPKD cells compared with NHK cells (69), consistent with the predicted function of the polycystins (13). In normal renal cells, Ca²⁺ restriction was shown to cause a phenotypic switch in the cAMP mitogenic response such that cAMP stimulates ERK and cell proliferation, mimicking the response observed in ADPKD cells (69, 73). It is well established that changes in cellular levels of cAMP and Ca²⁺ are important for deciphering various stimuli. Moreover, changes in intracellular Ca²⁺ can modulate AC coupling to GPCR, such as V2R, and downstream components of the MEK/ERK pathway through interactions with various A-kinase-anchoring proteins, including AKAP79 (14, 28, 52, 66). Here, we determined whether a reduction in intracellular Ca²⁺ alters the response of NHK cells to CaM inhibition by W-7. NHK cells were treated for 30 min with 10 nM BAPTA-AM, a membrane-permeable Ca²⁺ chelator, or overnight with 10 μM verapamil, a Ca²⁺ channel blocker; then, W-7 was added for 15 min and AVP was added for an additional 15 min. Verapamil decreased AVP-induced cAMP production from 21.3 ± 1.9 to 15.3 ± 1.4 pmol/monolayer (n = 4 monolayers, P < 0.001). We also found that intracellular Ca²⁺ chelation with BAPTA caused a small reduction in the cAMP response to AVP. Restriction of intracellular Ca²⁺ with either BAPTA or verapamil caused a switch in the W-7 response such that CaM inhibition decreased AVP-mediated cAMP production in NHK cells, similar to the W-7 response in ADPKD cells (Fig. 8). It is possible that Ca²⁺ reduction increases the functional coupling of CaM-sensitive ACs, such as AC3, to V2R.

AC3 is the only AC known to be negatively regulated by Ca²⁺/CaM-dependent kinase II (CaMKII) (9, 12, 63). To determine whether AC3 mediates cAMP production, ADPKD cells were treated with 10 μM KN-93, a CaMKII inhibitor. KN-93 alone caused a slight, but statistically significant, increase in intracellular cAMP (1.4 ± 0.4, KN-93 vs. 1.1 ± 0.1 pmol/monolayer, control, P < 0.05) and amplified the AVP-mediated cAMP accumulation (19.5 ± 2.8 vs. 11.8 ± 1.0 pmol/monolayer, AVP alone P < 0.001). By contrast, KN-93 had no significant effect on FSK-mediated cAMP production (24.6 ± 2.9; FSK+KN-93 vs. 21.6 ± 1.2 pmol/monolayer; FSK alone). In NHK cells, KN-93 did not affect basal, AVP-, or FSK-mediated cAMP accumulation, similar to the lack of an effect of CaM inhibition by W-7. These results indicate that the AVP response in ADPKD cells is mediated in part by a CaM-activated, CaMKII -inhibited AC, namely AC3.

Effect of CaM Inhibition on AVP- and FSK-Stimulated Cl⁻ Secretion by ADPKD and NHK Cell Monolayers

cAMP-dependent Cl⁻ and fluid secretion is responsible for fluid accumulation within cysts, leading to massive enlargement of ADPKD kidneys (29, 49, 61). To examine the relative contributions of CaM-sensitive ACs on AVP-mediated Cl⁻ secretion, we tested the effect of W-7 on AVP- and FSK-stimulated I_SC in polarized ADPKD and NHK cell monolayers (Fig. 9). Cells were treated with 10 μM benzamil, an amiloride derivative and high-affinity ENaC inhibitor, at the beginning of the experiment to block the Na⁺ current. Basolateral addition of 10 nM AVP (Fig. 9, A and C) or 1 μM FSK (Fig. 9, B and D) increased I_SC across ADPKD (Fig. 9, A and B) and NHK (Fig. 9, C and D) cell monolayers. The effect of AVP on anion secretion was slightly greater in NHK cells compared with ADPKD cells, consistent with the greater cAMP increase in the normal cells (Fig. 5 and Table 1). By contrast, the effect of 1 μM FSK on I_SC was similar between NHK and ADPKD cell monolayers. In ADPKD cells, CaM inhibition with W-7 significantly decreased AVP-mediated and FSK-mediated I_SC to baseline levels (Fig. 10), whereas W-7 had almost no effect on AVP- and FSK-induced I_SC by NHK cells, similar to the lack of an effect on cAMP production. These results support the hypothesis that CaM inhibition reduces the amplitude of cAMP production by AVP and FSK in ADPKD cells and that this has functional consequences regarding Cl⁻ secretion.

Fig. 9.

Effect of CaM inhibition on AVP- and FSK-induced anion secretion by ADPKD and NHK cell monolayers. ADPKD cells (A and B) and NHK (C and D) cells were grown for 7–10 days on permeable Transwell supports and mounted in modified Ussing chambers for measurement of short-circuit current (I_SC). Benzamil (10 μM) was added to the apical media to block the Na⁺ current. In ADPKD cell monolayers, basolateral AVP (A) and FSK (B) increased I_SC. W-7 caused a rapid and complete inhibition of AVP- and FSK-induced anion secretion in ADPKD monolayers (A and B). In NHK cell monolayers, AVP (C) and FSK (D) stimulated anion secretion, consistent with the effect of cAMP production. However, inhibition of CaM with W-7 had very little or no effect on anion secretion in NHK monolayers (C and D).

Fig. 10.

Composite of the effect of CaM inhibition on AVP- and FSK-induced anion secretion by ADPKD and NHK cell monolayers. In the presence of benzamil (Cont), positive I_SC is consistent with anion secretion, specifically Cl⁻ (41). Basolateral AVP (10 nM) increased anion secretion in NHK cells to a greater extent than ADPKD cells, similar to the difference in the AVP effect on cAMP production (Fig. 5), whereas FSK (1 μM) increased anion secretion to a similar level between ADPKD and NHK cell monolayers. After a new steady state was achieved, 50 μM W-7 was added to inhibit CaM. W-7 alone had no effect on basal current (data not shown). In ADPKD cells, CaM inhibition blocked AVP- and FSK-induced anion secretion, but had very little or no effect on I_SC in NHK cell monolayers. Data represent means ± SE. The numbers of monolayers per treatment group are indicated in parentheses. #P < 0.01 compared with control. *P < 0.001 vs. AVP. †P < 0.05 vs.. FSK alone.

DISCUSSION

V2R is the major regulator of AC activity and cAMP production in principal cells of the collecting duct and cyst epithelial cells and is a target for ADPKD treatment (16, 56, 60, 65). In this study, we examined the functional role of CaM-sensitive AC isoforms in AVP-mediated cAMP synthesis in human ADPKD and NHK cells. The key observations are 1) human ADPKD cyst epithelial cells have a higher baseline cAMP level compared with NHK cells; 2) AVP caused a greater cAMP response in normal renal cells despite greater V2R mRNA expression in ADPKD cells; 3) ADPKD and NHK cells and tissues express mRNA for all nine membrane AC isoforms, but the relative contributions of the Ca²⁺/CaM AC isoforms in AVP-mediated cAMP production differ between cystic and normal cells; 4) inhibition of CaM decreased AVP-induced cAMP production and Cl⁻ secretion by ADPKD cells, but had no effect in normal cells; 5) decreasing intracellular Ca²⁺ in NHK cells appeared to switch V2R signaling to a CaM-sensitive AC; and 6) CaMKII inhibition increased basal and AVP-induced cAMP production in ADPKD cells, implicating AC3, a CaM-activated and CaMKII-inhibited AC, in V2R signaling.

Elevated Renal cAMP in PKD

Renal cAMP levels have been shown to be increased in several animal models of PKD (16, 18, 56, 60, 64, 70). The elevation in intracellular cAMP in cystic epithelial cells could be due to overexpression and hyperactivation of V2R, differences in the rate of cAMP synthesis by ACs, or degradation by PDEs (18, 31, 32, 57, 64). Moreover, increased plasma levels of AVP in response to a decline in urine concentrating capacity of cystic kidneys may also contribute to elevated production of cAMP. It has long been established that there is a concentration defect in kidneys of ADPKD patients, including children, even before measurable renal insufficiency (11, 30, 55). It has also been hypothesized that elevated renal cAMP in cystic epithelial cells is a direct consequence of decreased intracellular Ca²⁺ levels caused by mutations in PC-1 or PC-2, leading to upregulation of Ca²⁺-inhibited AC6 and downregulation of PDE1 (1, 57, 64). In the present study, we found that basal cAMP levels were significantly elevated in cultured human ADPKD cells compared with normal renal cells (Figs. 5 and 6), suggesting that there is an intrinsic difference in cAMP regulation. This difference in basal cAMP levels appears to be dependent on PDE activity since cAMP levels are the same in the presence of IBMX, a global PDE inhibitor (Fig. 6B and Table 1). Thus, in the absence of AVP, elevated basal cAMP in ADPKD cells may be caused by a decrease in the rate of cAMP degradation by PDE1, a Ca²⁺/CaM-activated PDE. FSK-mediated cAMP levels were the same between NHK and ADPKD cells in either the presence or absence of IBMX (Table 1). FSK directly stimulates all ACs (except AC9) and significantly increases the V_max for cAMP synthesis, exceeding the rate of catalysis (44). FSK-activated cAMP levels are less sensitive to changes in PDE activity. Even in the absence of IBMX, FSK-mediated cAMP levels would be the same between ADPKD and NHK cells if global expression of FSK-sensitive AC isoforms is relatively the same.

Renal V2R expression has been shown to be elevated in animals with hereditary cystic disease (18, 31, 32, 57). In the current study, V2R mRNA levels was found to be 1.4-fold higher in human ADPKD cells compared with NHK cells and 2-fold elevated in ADPKD tissues compared with NHK tissues. Despite higher V2R mRNA expression in ADPKD cells, AVP caused a greater increase in intracellular cAMP in NHK cells compared with ADPKD cells (Table 1). These results are consistent with the previous observation that primary cultures of ADPKD cystic cells have AC hyporesponsiveness to agonist stimulation (67). Treatment with IBMX significantly enhanced the AVP response in ADPKD cells and caused a small, but insignificant, increase in NHK cells. However, the amplitude of the AVP response remained greater in NHK cells (Fig. 6 and Table 1). Considering that the FSK response was similar between ADPKD and NHK cells, the disparity in the AVP response is most likely due to changes in the relative contribution of AC isoforms to AVP-mediated cAMP production and/or coupling of V2R to specific AC isoforms and not due to differences in total AC abundance.

AC Isoform Expression in Human ADPKD Cyst Epithelial Cells and NHK Cells

Previous studies have investigated the expression pattern of AC isoforms and their specific roles in AVP-mediated cAMP production in mouse and rat isolated collecting ducts and cells (26, 48). To our knowledge, the expression and functional role of specific AC isoforms in AVP-mediated cAMP synthesis in human ADPKD cells and normal collecting duct cells have not been previously characterized. We found that all nine membrane-bound AC isoforms were expressed in ADPKD and NHK cells (Fig. 1) and tissue (data not shown). mRNA levels of ACs 1, 3, and 8 were decreased in ADPKD cells compared with NHK cells (Fig. 1B), whereas mRNA for Ca²⁺-inhibited ACs 5 and 6 was increased. Correspondingly, AC5/6 protein was significantly upregulated and AC3 was slightly downregulated in ADPKD cells and tissues compared with NHK (Fig. 2). There was no difference in AC1 protein levels.

ACs 1, 3, and 5/6 localized to collecting ducts and other tubule segments of normal kidneys and cystic epithelia of ADPKD kidneys (Figs. 3 and 4). Thus ADPKD cells harbored no surprises in relation to AC expression profiles in other tissues where the expression of multiple AC isoforms has been reported (12, 23). Coexpression of several ACs may be necessary to fine tune the regulation of the cAMP signal in response to specific agonists and cellular conditions (8, 12). In mice and rat kidneys, Ca²⁺-inhibited and Ca²⁺/CaM-activated AC isoforms were found to be expressed in IMCD cells (26, 48), suggesting that multiple AC isoforms may be responsible for mediating the AVP-dependent cAMP response. While collecting duct cells express multiple ACs, the relative contribution of specific isoforms to cAMP synthesis in response to receptor activation depends on intracellular Ca²⁺ levels and other regulators of AC activity. Coupling of V2R to specific AC isoforms may be an important mechanism to adjust the gain of the response to a cAMP agonist.

AVP-Induced cAMP Synthesis in ADPKD Cells Appears to be Mediated by AC3

Recent studies designed to characterize the role of specific ACs in AVP-mediated cAMP signaling and water permeability in collecting ducts have challenged a well-accepted view that ACs 5 and 6, Ca²⁺-inhibited ACs, are the only isoforms involved in AVP-stimulated cAMP synthesis (26, 48). AC3, a Ca²⁺/CaM-stimulated AC, was also found to be present in rat and mouse collecting duct cells and to contribute to the synthesis of cAMP in response to AVP (26, 48, 51). In addition, knockdown of either AC3 or AC6 using small interfering RNA (siRNA) in primary cultured mouse IMCD significantly reduced AVP-stimulated cAMP accumulation, consistent with a role of both AC isoforms in AVP-mediated cAMP synthesis. AC6 knockout mice were found to have a concentrating defect (7, 43), whereas AC3 knockout mice were normal (39), raising the question of the physiological relevance of AC3, relative to AC6, in regulation of water permeability of the collecting duct (39, 42).

We found stark differences in response to CaM inhibition of AVP-mediated cAMP production between ADPKD and normal cells. W-7 completely blocked AVP-induced cAMP production (Fig. 6) and anion secretion in ADPKD cells (Figs. 9 and 10), but had almost no effect in NHK cells. W-7 also inhibited the effect of a submaximal concentration of FSK (1 μM) on cAMP production and anion secretion in ADPKD cells, consistent with the effect of CaM inhibition being at the level of ACs (26). The extent to which W-7 decreased cAMP production also suggests that the majority of AVP-mediated AC activity was CaM sensitive, even though all isoforms appeared to be expressed. As predicted, inhibition of AVP-induced cAMP production by W-7 was overridden by maximal activation of ACs by 10 μM FSK. Furthermore, PDE inhibition did not prevent the divergent effects of W-7 in normal and cystic cells (Fig. 6B).

In light of these findings, it seems reasonable to conclude that AVP signaling is primarily coupled to CaM-sensitive ACs 1, 3, and/or 8 in ADPKD cells. The fact that the expression of these ACs is downregulated in ADPKD cells and tissues (Figs. 1 and 2) suggests a possible compensatory mechanism to restore normal cAMP signaling in the presence of elevated AVP. Moreover, changes in AC expression may not reflect changes in the subcellular localization of AC isoforms within microdomains that control the specificity and compartmentalization of the cAMP signal. Previously, AC3 was shown to be uniquely inhibited by CaMKII (12, 48). Indeed, we found that CaMKII inhibition increased AVP-mediated cAMP production by ADPKD cells, consistent with functional coupling of AC3 to V2R- and AVP-mediated cAMP production.

In NHK cells, CaM inhibition had no significant effect on cAMP levels and CaMKII inhibition only slightly increased cAMP, consistent with a lack of a role for AC3. These findings are in contrast to previous studies that showed that AC3 was involved in AVP-mediated cAMP signaling in rat IMCD suspension and freshly isolated mouse IMCD (26, 48, 51). This discrepancy may be due to species difference or differences between freshly isolated collecting ducts and primary collecting duct cells (74). Previously, Yasuda et al. (74) found that the amplitude of the cAMP response to G protein-coupled receptor ligands differed between primary IMCD cells and isolated collecting ducts. Strait et al. (48) reported that even though the responsiveness to CaM inhibition in cultured cells was less than that in freshly isolated IMCDs, the overall pattern of the W-7 response was the same between the two preparations. Therefore, it could be concluded that primary cultures are a reasonable model for the investigation of cAMP metabolism and the role of CaM-sensitive ACs in AVP-mediated signaling; however, there needs to be caution in extrapolation of the results from cultured cells to tissue. Primary cells may not completely differentiate, and therefore the relative contribution of components to cAMP metabolism may not fully represent the condition in situ.

To further address this issue, we analyzed the effect of W-7 on AVP-induced cAMP accumulation in mouse cortical collecting duct cells. Consistent with their previous work in mouse and rat IMCDs, we found that CaM inhibition with W-7 significantly decreased AVP- and FSK-mediated cAMP accumulation in M1 cells (Fig. 7). These results suggest that the relative contribution of CaM-stimulated ACs to AVP-induced cAMP production may be different between humans and rodents.

Reduction in Intracellular Ca²⁺ Changes the Response of CaM Inhibition

In NHK cells, Ca²⁺ restriction with Ca²⁺ channel blockers or Ca²⁺ chelators causes a phenotypic switch in the mitogenic response to cAMP, such that cAMP stimulates the B-Raf/MEK/ERK pathway and accelerates cell proliferation (73). This mechanism appears to involve the derepression of B-Raf, possibly by relieving the inhibitory effect of Ca²⁺-dependent kinases such as Akt (69, 71, 73). We have noted that the Ca²⁺ switch requires several hours, suggesting that additional Ca²⁺-dependent mechanisms may be necessary (73). In the current study, we found that treatment of NHK cells with the Ca²⁺ channel blocker verapamil for 24 h or BAPTA for 1 h changed the CaM sensitivity of the cAMP response to AVP, such that W-7 inhibited AVP-induced cAMP production (Fig. 8). These data support the hypothesis that persistently reduced intracellular Ca²⁺ levels in ADPKD cells switch AC regulation from Ca²⁺ inhibited (AC 5/6) to Ca²⁺/CaM stimulated (AC 3), thereby minimizing the overall amplitude of the AVP-mediated increase in cAMP production. Hyporesponsiveness to AVP stimulation in ADPKD cells has previously been reported (67). While this may diminish the magnitude of fluid secretion linked to Cl⁻ transport into cysts, the underlying process appears to uncover a mitogenic response of cyst mural cells to cAMP that promotes accelerated cyst enlargement (69, 71–73).

Summary

According to our results, cAMP production by ACs differs importantly in human ADPKD and normal renal cells, and the relative contribution of different AC isoforms to the cAMP response may differ between different species. A Ca²⁺/CaM-stimulated AC isoform, most likely AC3, mediates AVP-induced cAMP synthesis and anion secretion by human ADPKD cells, but not NHK cells. However, normal renal cells made to behave like ADPKD cells by reducing intracellular Ca²⁺ had increased activity of CaM-sensitive AC3 to synthesize cAMP. It is possible that during low-Ca²⁺ conditions, human collecting duct cells recruit Ca²⁺/CaM-dependent ACs to V2R. Coupling of Ca²⁺/CaM-stimulated ACs to V2R instead of Ca²⁺-inhibited ACs is likely to be a compensatory mechanism to limit the amplitude of the AVP effect and mitigate excessively high cAMP levels in cells with decreased intracellular Ca²⁺. Currently, chemical development of potent isoform-selective small-molecule inhibitors is still in an early stage (45); however, the identification a functional role for specific ACs in ADPKD progression will be beneficial as ACs become a viable drug target (37). Studies using AC3 and AC6 mouse knockouts may be valuable tools for investigating the relative role Ca²⁺/CaM-regulated AC isoforms in the pathogenesis of PKD in vivo.

GRANTS

This work was supported by grants from the National Institutes of Diabetes and Digestive and Kidney Diseases (R01DK081579 to D. P. Wallace) and The Polycystic Kidney Disease (PKD) Foundation (to C. S. Pinto). Generation of primary cell cultures by the PKD Research Biomaterials and Cellular Models Core was funded by the PKD Foundation and the KUMC Kidney Institute.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: C.S.P. and D.P.W. provided conception and design of research; C.S.P., G.A.R., E.N., and C.W. performed experiments; C.S.P., G.A.R., and D.P.W. analyzed data; C.S.P., G.A.R., and D.P.W. interpreted results of experiments; C.S.P. and D.P.W. prepared figures; C.S.P. and D.P.W. drafted manuscript; C.S.P. and D.P.W. edited and revised manuscript; C.S.P., G.A.R., E.N., C.W., and D.P.W. approved final version of manuscript.