Mechanisms of vasopressin-induced intracellular Ca²⁺ oscillations in rat inner medullary collecting duct

Abstract

Arginine vasopressin (AVP) causes increase in intracellular Ca²⁺ concentration with an oscillatory pattern. Ca²⁺ mobilization is required for AVP-stimulated apical exocytosis in inner medullary collecting duct (IMCD). The mechanistic basis of these Ca²⁺ oscillations was investigated by confocal fluorescence microscopy and flash photolysis of caged molecules in perfused IMCD. Photorelease of caged cAMP and direct activation of ryanodine receptors (RyRs) by photorelease of caged cyclic ADP-ribose (cADPR) both mimicked the AVP-induced Ca²⁺ oscillations. Preincubation of IMCD with 100 μM 8-bromo-cADPR (a competitive inhibitor of cADPR) delayed the onset and attenuated the magnitude of AVP-induced Ca²⁺ oscillations. These observations indicate that the cADPR/RyR pathway is capable of supporting Ca²⁺ oscillations and endogenous cADPR plays a major role in the AVP-induced Ca²⁺ oscillations in IMCD. In contrast, photorelease of caged inositol 1,4,5-trisphosphate (IP₃) induced Ca²⁺ release but did not maintain sustained Ca²⁺ oscillations. Removal of extracellular Ca²⁺ halted ongoing AVP-mediated Ca²⁺ oscillation, suggesting that it requires extracellular Ca²⁺ entry. AVP-induced Ca²⁺ oscillation was unaffected by nifedipine. Intracellular Ca²⁺ store depletion induced by 20 μM thapsigargin in Ca²⁺-free medium triggered store-operated Ca²⁺ entry (SOCE) in IMCD, which was attenuated by 1 μM GdCl₃ and 50 μM SKF-96365. After incubation of IMCD with 1 nM AVP in Ca²⁺-free medium, application of extracellular Ca²⁺ also triggered Ca²⁺ influx, which was sensitive to GdCl₃ and SKF-96365. In summary, our observations are consistent with the notion that AVP-induced Ca²⁺ oscillations in IMCD are mediated by the interplay of Ca²⁺ release from RyRs and a Ca²⁺ influx mechanism involving nonselective cation channels that resembles SOCE.

oscillation of intracellular Ca²⁺ concentration ([Ca²⁺]_i) serves as the signal transduction mechanism for many physiological stimuli in both excitable and nonexcitable cells, with information being encoded in both the frequency and the amplitude of the Ca²⁺ signal (2). There is emerging evidence that Ca²⁺ oscillation in renal epithelium is an integral part of the signaling transduction process for regulation of water reabsorption (8, 45). A physiological dose of AVP has been shown to trigger intracellular Ca²⁺ mobilization and Ca²⁺ oscillations in inner medullary collecting duct (IMCD) (32, 45), and the Ca²⁺ mobilization is necessary for AVP-stimulated apical exocytosis and osmotic water permeability (8, 45). AVP also induced Ca²⁺ oscillations in mouse thick ascending limb, which was associated with AVP-stimulated secretion of nucleotides (30). Ca²⁺ oscillations were also reported recently in tubular epithelial cells near the macula densa (23). They are modulated by luminal NaCl concentration and luminal flow, and are possibly related to juxtaglomerular signaling. Furthermore, angiotensin II-triggered Ca²⁺ oscillations have been recorded in descending vasa recta pericytes (11, 51). It has been suggested that Ca²⁺ oscillations in pericytes are due to repetitive cycles of ryanodine-sensitive sarcoplasmic reticulum (SR) Ca²⁺ release and SKF-96365-sensitive refilling of SR Ca²⁺ stores.

Our previous studies (8, 45) have provided the initial characterization of AVP-induced Ca²⁺ oscillation in IMCD. With the use of confocal fluorescence microscopy to monitor [Ca²⁺]_i and apical exocytosis in individual cells, AVP was found to trigger a rapid increase in [Ca²⁺]_i followed by sustained repetitive Ca²⁺ oscillations. These Ca²⁺ responses were mediated through a cAMP-dependent mechanism and mimicked by an agonist of exchange protein directly activated by cAMP (Epac) (46). Removal of extracellular Ca²⁺ did not prevent the initial rise of [Ca²⁺]_i induced by AVP but abolished the sustained Ca²⁺ oscillation (45). In the absence of extracellular Ca²⁺, ryanodine completely obliterated AVP-induced Ca²⁺ mobilization and AVP-stimulated increase of osmotic water permeability (8). These results suggested that both extracellular Ca²⁺ influx and intracellular Ca²⁺ release contribute to the AVP-induced Ca²⁺ response.

Ca²⁺ oscillation can be driven by influx of extracellular Ca²⁺, Ca²⁺-induced Ca²⁺ release (CICR) from intracellular Ca²⁺ stores, or both (33). Ryanodine receptors (RyRs) and inositol 1,4,5-trisphosphate (IP₃) receptors (IP₃Rs) are expressed in endoplasmic reticulum (ER) of IMCD (8, 19, 44). Both receptors are capable of supporting CICR and Ca²⁺ oscillations (13, 14). Ca²⁺ release through RyRs or IP₃Rs may deplete intracellular Ca²⁺ stores and trigger store-operated Ca²⁺ entry (SOCE), which can serve as a mechanism to replenish the depleted Ca²⁺ stores and sustain Ca²⁺ oscillations (33, 34). To capture the dynamics of Ca²⁺ release from RyRs and IP₃Rs from the ER of IMCD, flash photolysis of caged cyclic ADP-ribose (cADPR) and caged IP₃ was used in the present study to directly activate RyRs and IP₃Rs, respectively. The specific contributions of RyRs, IP₃Rs, and SOCE in the AVP-induced Ca²⁺ oscillations were evaluated. Our results indicate that AVP-induced Ca²⁺ oscillation in IMCD is mediated in part by the endogenous production of cADPR and is maintained through the interplay of Ca²⁺ release from RyRs and Ca²⁺ influx through SOCE.

MATERIALS AND METHODS

Isolation and perfusion of single IMCD segment.

Experiments were carried out in accordance with guidelines for the care and use of research animals. All experiments were performed under protocols approved by the Institutional Animal Care and Use Committee at the University of South Florida, in accordance with Public Health Service Policy on Humane Care and Use of Laboratory Animals. Experiments were conducted in IMCD isolated from male Sprague-Dawley rats (60–100 g body wt; Harlan). Rats were treated with furosemide (5 mg) intraperitoneally for 30 min and then killed with an overdose of 5% halothane through a Fluotec Mark-3 vaporizer (Ohmeda). Furosemide treatment was used to wash out the medullary osmolarity gradient, in order to minimize osmotic shock when the medulla is placed in isotonic dissection solution (5). The kidneys were rapidly removed through a midline abdominal incision and placed in an ice-cold dissecting solution. The dissection solution consisted of (in mM) 120 NaCl, 25 NaHCO₃, 2 K₂HPO₄, 1.2 MgSO₄, 2 CaCl₂, 5.5 glucose, and 5 sodium acetate. Terminal IMCD segments were dissected from the inner half of the inner medulla (5). The isolated piece of IMCD was then transferred to a temperature-controlled perfusion chamber (Vestavia) mounted on a Leica DMI6000 inverted microscope, which was coupled to a Leica TCS SP5 confocal scanning unit. The IMCD was then cannulated and perfused with glass concentric pipettes by the method developed by Burg (3). Luminal and bath perfusates were identical to the dissecting solution. All solutions were gassed with 95% O₂ and 5% CO₂ before use, and pH was adjusted to 7.4. For Ca²⁺-free perfusate, CaCl₂ was replaced by 2 mM EGTA. For the sulfate- and phosphate-free solution, K₂HPO₄ and MgSO₄ were replaced by equimolar KCl and MgCl₂, respectively.

Measurement of intracellular calcium.

The changes of [Ca²⁺]_i in IMCD cells induced by AVP and other pharmacological agents were determined from confocal fluorescence images of the perfused tubules as described previously (8, 45). In brief, 5 μM fluo-4 AM (Molecular Probes) was loaded into the IMCD from the peritubular solution at room temperature for 30 min. The tubule was then washed and incubated at 37°C for another 30 min for deesterification before measurements were begun. Confocal fluorescence images were acquired from the lower surface of the perfused IMCD with the 488-nm laser line. All images were collected with a Leica ×63 plan-apochromat objective (numerical aperture 1.2, water immersion). Emission was collected with a spectral window of 495–540 nm at 0.5 or 1 Hz and stored digitally. The spatial and temporal variations of [Ca²⁺]_i in individual IMCD cells were measured from the stored images with Leica Application Suite Advanced Fluorescence software.

Flash photolysis of caged compounds in IMCD.

Cell-permeant caged Ca²⁺ (DMNP-EDTA-AM, 20 μM), caged cAMP (DMNB-caged cAMP, 200 μM), and caged IP₃ [d-2,3-O-isopropylidene-6-O-(2-nitro-4,5-dimethoxy)benzl-myo-inositol 1,4,5-trisphosphate-hexakis(propionoxymethyl)ester (IP₃/PM), 1 μM] were dissolved in DMSO and loaded into IMCD together with fluo-4 AM for 30 min. Caged cADPR (NPE-caged cADPR, 40 μM) is cell impermeant. It was loaded into IMCD with a reversible permeabilization procedure (12, 40). In brief, the cannulated IMCD was washed with regular IMCD perfusate at 37°C for 15 min, and then the peritubular perfusate was switched to a Ca²⁺-free permeabilization solution containing 40 μM caged cADPR for 2 min. The permeabilization solution was then removed, and IMCD was incubated at 37°C with regular IMCD perfusate for 60 min for membrane recovery. Fluo-4 AM was then loaded to IMCD after the recovery period. Permeabilization solution contained (in mM) 137 NaCl, 3 KCl, 5 glucose, and 20 piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES), with 1 mg/ml bovine serum albumin and 0.4 IU/ml streptolysin-O. This permeabilization procedure did not affect AVP-induced Ca²⁺ oscillation in IMCD cells (see Fig. 5B), indicating that it did not compromise this signaling process.

Fig. 5.

Ca²⁺ oscillations in IMCD induced by 1 nM AVP. In the presence of 10 μM nifedipine (27 cells/2 tubules; A), after the reversible permeabilization procedure using streptolysin-O (34 cells/3 tubules; B), and in the presence of 40 μM xestospongin C (39 cells /4 tubules; C), AVP was added to the bath at t = 0. Filled square indicates that mean value is significantly different from the baseline before AVP exposure (P < 0.05). Inset: time courses collected from 4 IMCD cells of the same tubule (red, blue, green, and black). D: corresponding mean power spectra of oscillations in A, B, and C. Dashed lines are SE.

Flash photolysis of the caged compound was induced by an UV nitrogen-pulse laser (Laser Science, 337 nm, 4 ns/pulse, 300 μJ/pulse) coupled to a quartz optical fiber of 200-μm diameter, which was mounted on a micromanipulator. To visualize and to align the laser spot on the IMCD before UV flash photolysis, the UV laser pulses were directed through a tunable-dye control module that converted UV laser pulses into visible laser pulses centered at 640 nm (48). The firing frequency (30 Hz) and number of laser pulses in each burst were controlled with an analog-to-digital board (Data Translation DT-2801A) driven by a computer. The duration of photorelease was controlled by the number of laser pulses as described previously (48). The caging moiety in DMNB-caged cAMP and caged IP₃ has the maximum of reabsorption at 370 nm, while that of NPE-caged cADPR is at 270 nm. UV laser at 337 nm is more efficient to uncage cAMP and IP₃ than cADPR. Thirty and sixty pulses of UV laser were used to uncage cAMP and cADPR, respectively.

Flash photolysis of caged signaling molecules allowed rapid manipulation of the intracellular environment with minimal disturbance in confocal image acquisition. The use of consecutive short UV laser pulses (300 μJ/4 ns) to induce flash photolysis conferred more control to minimize the potential tissue damage due to UV radiation than the use of a conventional flash lamp. A single flash from a xenon flash lamp has energy on the order of 50–100 mJ/pulse and lasts for a few microseconds.

Chemicals.

AVP was purchased from Bachem. SKF-96365 and xestospongin C were purchased from Calbiochem. DMNP-EDTA-AM, DMNB-caged cAMP, NPE-caged cADPR, and fluo-4 AM were purchased from Molecular Probes. IP₃/PM was purchased from SiChem. Thapsigargin and streptolysin-O were from Sigma.

Data analysis.

Time series of fluo-4 emission variations in individual IMCD cells were sampled at 0.25 or 0.5 Hz for spectral analysis. Each time series was normalized with respect to the baseline before AVP stimulation or photolysis and was subjected to linear trend removal. Two hundred fifty-six data points were used to calculate the power spectrum with an algorithm based on fast Fourier transform (47). Results are reported as means ± SE. Statistical significance was calculated by using Student's t-tests for paired or unpaired data. Only one IMCD was studied in each rat.

RESULTS

Flash photolysis of caged Ca²⁺ and cAMP induced Ca²⁺ mobilization in IMCD.

The efficacy of the UV pulsed laser system and the sensitivity of the detection system were first evaluated by using a photosensitive calcium chelator in IMCD. In IMCD loaded with DMNP-EDTA and fluo-4, a single UV laser pulse (4 ns) was sufficient to induce a detectable increase in fluo-4 emission (Fig. 1). A graded accumulative increase of fluo-4 emission was observed when the number of laser pulses was increased in each successive laser burst. The mean normalized fluo-4 emission after one, two, and three UV laser pulses was 1.12 ± 0.05, 1.32 ± 0.08, and 1.57 ± 0.06 (29 cells/2 tubules), respectively. These observations indicated that a single UV laser pulse provides sufficient energy to initiate photolysis of in perfused IMCD, and that the extent of photolysis can be controlled by the number of laser pulses used.

Fig. 1.

Flash photolysis of Ca²⁺ chelator in inner medullary collecting duct (IMCD). A: mean normalized time course of changes in fluo-4 emission from intact IMCD cells induced by photolysis of DMNP-EDTA. Photorelease of Ca²⁺ was triggered by UV laser pulse. Each arrow indicates a burst of laser pulses; 1, 2, and 3 pulses of UV laser were applied sequentially. Dashed lines are SE (29 cells/2 tubules). B: 3 consecutive images (sampling at 1 Hz) showing the changes of fluo-4 emission induced by photolysis of DMNP-EDTA with 2 UV laser pulses before (a), during (b), and after (c) photolysis.

We have shown previously (8, 45) that AVP-induced Ca²⁺ oscillations are mediated by cAMP. To examine whether cAMP-mediated Ca²⁺ oscillation is intact in IMCD after exposure to the UV laser illumination, DMNP-caged cAMP was introduced into perfused IMCD. A burst of 30 UV laser pluses delivered over 1 s triggered a rapid increase in cytosolic Ca²⁺, which was followed by repetitive [Ca²⁺]_i oscillations (Fig. 2A). The time delay between the uncaging of cAMP and the onset of rise in cytosolic [Ca²⁺]_i ranged from 4 to 10 s. The mean normalized fluo-4 emission of the initial spike was 1.49 ± 0.06 (49 cells/5 tubules). The corresponding mean power spectrum had multiple peaks (see Fig. 4B). This means that the oscillatory frequency is time varying. Most of the power of oscillation was found in the frequencies below 0.1 Hz. A log scale was used to enhance the resolution of oscillatory frequencies in the lower frequency range. Each of these frequencies was associated with a different power level. A similar pattern was found in oscillations triggered by AVP (45). These observations confirmed that cAMP alone is sufficient to induce Ca²⁺ oscillation in IMCD cells, and UV flash photolysis did not compromise the ability of IMCD cells in generating Ca²⁺ response.

Fig. 2.

Flash photolysis of caged cAMP and caged cyclic ADP-ribose (cADPR) in IMCD. Mean normalized time courses of changes in fluo-4 emission from intact IMCD cells induced by photolysis of caged cAMP (49 cells/5 tubules; A) and caged cADPR (26 cells/3 tubules; B) are shown. A burst of 30 and 60 UV laser pulses (at 30 Hz) were used for uncaging cAMP and cADPR, respectively. Insets: time courses collected from 4 IMCD cells of the same tubule (red, blue, green, and black). Arrow indicates the laser burst. Filled square indicates that mean value is significantly different from the prephotolysis baseline (P < 0.05). Dashed lines are SE.

Fig. 4.

Flash photolysis of caged inositol 1,4,5-trisphosphate (IP₃) in IMCD. A: mean normalized time course of changes in fluo-4 emission from intact IMCD cells induced by uncaging of IP₃ (37 cells/3 tubules). Photolysis of caged IP₃ was induced by 3, 10, and 30 UV laser pulses. Arrow indicates the laser burst. Filled symbols indicate that mean values are significantly different from the prephotolysis baseline (P < 0.05). Inset: preincubated IMCDs with 40 μM xestospongin C (green line, 38 cells/3 tubules) abolished Ca²⁺ mobilization triggered by photolysis of caged IP₃ (red line, 26 cells/3 tubules). Each arrow indicates a burst of laser pulses. B: mean power spectra of oscillations in fluo-4 emission induced by photolysis of caged IP₃ in A and by caged cAMP and caged cADPR from Fig. 2. Dashed lines are SE.

Effects of photoreleased cyclic ADP-ribose and IP₃ in IMCD.

We have shown previously (8, 45) that AVP-induced Ca²⁺ mobilization is dependent on ryanodine-sensitive Ca²⁺ stores. To further investigate the RyR-dependent Ca²⁺ oscillations, IMCDs were loaded with NPE-caged cADPR, an endogenous agonist of RyR, by a reversible permeabilization protocol using streptolysin-O (50). Photorelease of cADPR with 60 UV laser pulses delivered over 2 s triggered a rapid increase in cytosolic Ca²⁺ followed by [Ca²⁺]_i oscillations. The mean normalized time course for fluo-4 emission is shown in Fig. 2B. The time delay between the uncaging of cADPR and the rise in cytosolic [Ca²⁺]_i was ∼10–20 s. The mean normalized fluo-4 emission at the peak of oscillation was 1.18 ± 0.04 (26 cells/3 tubules). The corresponding mean power spectrum is shown in Fig. 4B. The distribution of power spectral density was similar to that triggered by uncaging cAMP, but with less power in all oscillatory frequencies.

To test whether cADPR is involved in AVP-induced Ca²⁺ oscillations, cell-permeant 8-bromo-cADPR (100 μM) was used as a competitive inhibitor to suppress the endogenous activity of cADPR. Preincubation of IMCD with 8-bromo-cADPR for 25 min suppressed the amplitude and delayed the onset of the AVP-induced Ca²⁺ oscillations by ∼200 s (Fig. 3B). The mean normalized fluo-4 emission at the peaks of Ca²⁺ oscillations was reduced from 1.61 ± 0.17 (45 cells/3 tubules) to 1.25 ± 0.05 (42 cells/4 tubules) in the presence of 8-bromo-cADPR. The mean power spectra of AVP-induced oscillations in the absence and presence of 8-bromo-cADPR are shown in Fig. 3C. Both had multiple peaks and a broad distribution of spectral density below 0.1 Hz. The continuous presence of AVP seemed to partially overcome the inhibition of 8-bromo-cADPR. Ca²⁺ oscillations developed gradually after the initial 200 s. These observations suggest that endogenous cADPR may play a crucial role in the initiation of AVP-induced Ca²⁺ oscillations.

Fig. 3.

Inhibition of AVP-induced Ca²⁺ oscillations by 8-bromo-cADP-ribose in IMCD. A and B: AVP-induced Ca²⁺ oscillations (45 cells/3 tubules; A) and AVP induced Ca²⁺ oscillations after 25-min incubation with 100 μM 8-bromo-cADP-ribose (42 cells/4 tubules; B). C: corresponding mean power spectra of oscillations in A and B. Filled square indicates that mean value is significantly different from the baseline before AVP exposure (P < 0.05). Dashed lines are SE.

To investigate the involvement of IP₃-sensitive Ca²⁺ stores in mediating Ca²⁺ oscillations, the effects of photolytic release of IP₃ in fluo-4 emission were examined in IMCD. The mean normalized time courses of changes in fluo-4 emission in responding to flash photolysis of caged IP₃ are shown in Fig. 4A. Three UV pulses induced a small and slow transient increase of cytosolic Ca²⁺ after a time delay. Ten UV pulses triggered an initial jump in cytosolic Ca²⁺, followed by slow decay. The time delay between the laser pulses and initial increase of cytosolic Ca²⁺ was reduced compared with photolysis induced by three UV pulses. There was an upward trend in the cytosolic Ca²⁺ after the initial decay. Thirty UV pulses triggered a larger and more synchronized initial Ca²⁺ release. Cytosolic Ca²⁺ then remained steady above the baseline after a short decay. No consistent periodic spikes of cytosolic Ca²⁺ similar to those induced by cADPR were observed in all three levels of IP₃ uncaging. The mean power spectra of fluo-4 emission induced by 10 and 30 laser pulses are shown in Fig. 4B. No prominent peak in the power spectra was found. By increasing the loading concentration of caged IP₃/AM from 1 μM to 4 μM, one single UV pulse was sufficient to trigger a detectable Ca²⁺ response. Preincubation of IMCD with xestospongin C (40 μM; a common antagonist of IP₃Rs) completely obliterated the effects of photolytic release of caged IP₃ in Ca²⁺ mobilization (Fig. 4A, inset), confirming that IP₃ mobilized intracellular Ca²⁺ via IP₃Rs in IMCD.

The mean normalized time course of changes in fluo-4 emission induced by AVP in the presence of xestospongin C is shown in Fig. 5C. There were multiple peaks in the corresponding mean power spectrum (Fig. 5D), indicating that xestospongin C did not prevent AVP-induced Ca²⁺ mobilization and oscillations. However, the amplitude of the initial Ca²⁺ spike triggered by AVP was attenuated and delayed by xestospongin C. The mean normalized fluo-4 emission of the initial spike was reduced from 1.61 ± 0.17 (45 cells/3 tubules; Fig. 3A) to 1.21 ± 0.05 (39 cells/4 tubules, P < 0.05), which suggested that IP₃ may contribute in part to the initial spike of the AVP-induced intracellular Ca²⁺ release or xestospongin C may have other effects in the intracellular Ca²⁺ stores in addition to blocking IP₃Rs (9).

Store-operated Ca²⁺ entry in IMCD.

Our previous study (45) showed that removal of extracellular Ca²⁺ did not prevent the initial rise of [Ca²⁺]_i but inhibited the sustained oscillations induced by AVP in IMCD. This suggested that entry of extracellular Ca²⁺ was required to maintain the AVP-induced Ca²⁺ oscillations. Ca²⁺ entry was not mediated by L-type voltage-gated Ca²⁺ channel, as nifedipine (10 μM) did not inhibit AVP-induced Ca²⁺ oscillations (Fig. 5, A and D). Treatment of IMCD with thapsigargin (20 μM) to inhibit the ER Ca²⁺-ATPase in Ca²⁺-free medium caused an increase in intracellular Ca²⁺ due to Ca²⁺ leak from intracellular stores (data not shown). Readdition of 2 mM Ca²⁺ to the peritubular perfusate induced a rapid influx of Ca²⁺ in IMCD (Fig. 6), which was absent without the use of thapsigargin to deplete the intracellular Ca²⁺ stores. The latter observations are consistent with the report that changes of peritubular [Ca²⁺] alone induced only slight or no change in [Ca²⁺] in rat IMCD (4). The mean normalized fluo-4 fluorescence at the peaks of Ca²⁺ entry was significantly reduced from 2.64 ± 0.08 (126 cells/7 tubules) to 1.41 ± 0.14 (33 cells/3 tubules, P < 0.05) in the presence of 50 μM SKF-96365 and to 1.61 ± 0.06 (88 cells/5 tubules, P < 0.05) in the presence of 1 μM GdCl₃. SKF-96365 and a low concentration of Gd³⁺ are two commonly used antagonists for SOCE and nonselective cation channels. These results indicated that depletion of intracellular Ca²⁺ stores triggered SOCE in IMCD.

Fig. 6.

Activation of store-operated Ca²⁺ entry (SOCE) by thapsigargin in IMCD. IMCDs were incubated with 20 μM thapsigargin in the absence of extracellular Ca²⁺ for 25 min. Subsequent addition of 2 mM Ca²⁺ to the peritubular perfusate resulted in a rapid extracellular Ca²⁺ entry (126 cells/7 tubules), which was inhibited by GdCl₃ (88 cells/5 tubules) and 50 μM SKF-96365 (33 cells/3 tubules). The Ca²⁺ entry triggered by readdition of bath Ca²⁺ was absent without thapsigargin incubation (62 cells/5 tubules). Filled symbols indicate that mean value is significantly different from the equivalent data point in the control (P < 0.05). Dashed lines are SE.

To determine whether AVP could trigger extracellular Ca²⁺ influx similar to that induced by thapsigargin, IMCD was incubated with 1 nM AVP in Ca²⁺-free medium for 30 min. Readdition of 2 mM Ca²⁺ to the peritubular perfusate produced a rapid Ca²⁺ entry (Fig. 7). The mean normalized fluo-4 fluorescence at the peaks of Ca²⁺ entry was significantly reduced from 2.72 ± 0.09 (135 cells/8 tubules) to 1.32 ± 0.11 (35 cells/3 tubules, P < 0.05) in the presence of 50 μM SKF-96365 and to 1.89 ± 0.07 (54 cells/4 tubules, P < 0.05) in the presence of 1 μM GdCl₃.

Fig. 7.

Activation of SOCE by AVP. IMCDs were incubated with 1 nM AVP in the absence of extracellular Ca²⁺ for 25 min. Subsequent addition of 2 mM Ca²⁺ to the peritubular perfusate resulted in a rapid extracellular Ca²⁺ entry (135 cells/8 tubules), which was sensitive to GdCl₃ and (54 cells/4 tubules) and 50 μM SKF-96365 (35 cells/3 tubules). Filled symbols indicate that mean value is significantly different from the equivalent data point in the control (P < 0.05). Dashed lines are SE.

To determine whether AVP-induced Ca²⁺ oscillation required calcium entry, Ca²⁺ oscillation in IMCD was first induced by 1 nM AVP. Removal of extracellular Ca²⁺ immediately halted the Ca²⁺ oscillation. Readdition of 2 mM Ca²⁺ to the peritubular perfusate triggered a large and rapid Ca²⁺ entry, which was followed by Ca²⁺ oscillation (Fig. 8). These results together with the aforementioned data are consistent with the notion that extracellular Ca²⁺ entry activated subsequent to Ca²⁺ release from ryanodine-sensitive Ca²⁺ stores participates in the AVP-induced Ca²⁺ oscillations in rat IMCD.

Fig. 8.

Dependence of AVP-induced Ca²⁺ oscillations on extracellular Ca²⁺. Ca²⁺ oscillations were triggered by 1 nM AVP. Removal of extracellular Ca²⁺ halted the Ca²⁺ oscillations. Dashed lines are SE (11 cells from the same IMCD).

DISCUSSION

Previous studies have shown that AVP at a physiological concentration elicits robust Ca²⁺ mobilization and oscillations through a cAMP-dependent pathway and the Ca²⁺ mobilization is essential for AVP-induced apical incorporation of aquaporin-2 (AQP2) and water permeability changes in IMCD (8, 45, 46). Here we used UV laser-induced photorelease of cAMP to demonstrate that cAMP alone is sufficient to mimic AVP in initiating Ca²⁺ oscillation. We furthermore provide novel evidence suggesting that AVP-induced Ca²⁺ response is activated by the endogenous RyR agonist cADPR and maintained by Ca²⁺ influx via SOCE. However, AVP-induced Ca²⁺ oscillation is independent of Ca²⁺ release stimulated by IP₃ and Ca²⁺ influx through voltage-gated Ca²⁺ channels. Hence this cAMP-cADPR-RyR-SOCE pathway may represent an integrated Ca²⁺ signaling mechanism for the regulation of water permeability in IMCD.

AVP enhances renal water reabsorption through the activation of heteromeric G_s protein-coupled V₂ receptor, causing a rise in cAMP concentration, PKA-dependent phosphorylation of AQP2, and translocation of AQP2 to plasma membrane of IMCD cells (29). Increasing evidence suggests that Ca²⁺ signaling plays a pivotal role in this regulatory process. AVP has been shown to both increase cAMP and induce Ca²⁺ mobilization in IMCD (4, 8, 10, 22, 27, 38, 45). Chelation of intracellular Ca²⁺ with BAPTA completely abolished the AVP-induced apical exocytosis and increase in osmotic water permeability without affecting cAMP production (8, 45). AVP-induced Ca²⁺ mobilization appears to be a downstream effect following activation of adenylyl cyclase. It is mimicked by the membrane-permeant cAMP analog 8-(4-chlorophenylthio)cAMP (8) and by photorelease of caged cAMP as observed in the present study. The time delay of 4–10 s in the onset of caged cAMP-induced Ca²⁺ response is congruent with the delay in the increase of membrane capacitance observed after photorelease of cAMP in cultured IMCD cells (26), suggesting that Ca²⁺ response could be a rate-limiting step for cAMP-induced exocytosis.

AVP-induced Ca²⁺ mobilization is dependent on RyR-gated Ca²⁺ stores. Inhibition of RyR completely obliterates AVP-induced Ca²⁺ response and water permeability change in IMCD (8, 45). Immunofluorescence, immunoblotting, and RT-PCR experiments showed that RyR1 is predominantly expressed and colocalized with AQP2 in IMCD cells, supporting a cooperative association between the two types of channels (6, 8). The mechanism by which AVP activates RyR is unclear. Ca²⁺ influx through voltage-gated or non-voltage-gated Ca²⁺ channels is known to activate RyRs through CICR. However, AVP has been shown to trigger ryanodine-sensitive Ca²⁺ release in IMCD cells in the absence of extracellular Ca²⁺ (45), suggesting that the process is independent of Ca²⁺ influx-mediated CICR. Studies in other cell systems have shown that agonists including angiotensin II, endothelin-1, and acetylcholine can activate RyRs through stimulation of ADP-ribosyl cyclase to generate the endogenous RyR agonist cADPR (1, 18, 49), which increases the open probability of RyRs perhaps through interactions with FK506-binding protein (39, 41). Our observations strongly suggest that the AVP-induced Ca²⁺ oscillations in IMCD involve endogenous cADPR, because the Ca²⁺ response was delayed and suppressed by cADPR antagonist 8-bromo-cADPR and could be mimicked by photorelease of caged cADPR.

AVP may stimulate cADPR production through the cAMP signaling pathway. It has been reported in cardiac myocytes and chromaffin cells that application of cAMP analogs or activation of adenylate cyclase enhanced cADPR synthesis (20, 28, 43), an effect that could be blocked by the PKA inhibitor H-89 or Rp-adenosine 3′,5′-cyclic monophosphorothioate (Rp-cAMP[S]), an antagonist of cAMP for PKA binding. However, the AVP-induced Ca²⁺ mobilization does not seem to involve PKA, since it was not prevented by the PKA inhibitors H-89 and KT-5720 (46). It is most likely mediated by another cAMP effector, Epac, because the Epac-selective cAMP agonist 8-pCPT-2′-O-Me-cAMP mimicked AVP in triggering Ca²⁺ oscillations, which was blocked by ryanodine but not by Rp-cAMP[S] (46). Although a functional link between Epac and activation of ADP-ribosyl cyclase or cADPR synthesis has not been established, it warrants future investigations.

Multiple IP₃R subtypes have been identified in IMCD cells (19, 44), but their role in the AVP-induced Ca²⁺ mobilization is not well defined. Photorelease of IP₃ in IMCD activated a robust increase of [Ca²⁺]_i, which was blocked by xestospongin C. However, xestospongin C did not abolish AVP-induced Ca²⁺ mobilizations and oscillations. These observations are consistent with the previous report that AVP at physiological concentration does not activate the phosphoinositide signaling pathway in IMCD (7). A closer examination of the data revealed that the initial spike of AVP-induced Ca²⁺ release was attenuated and delayed by xestospongin C, suggesting that IP₃ might still contribute in part to the initial Ca²⁺ release. However, xestospongin C is also known to inhibit ER Ca²⁺ pump in addition to IP₃Rs (9). If this occurs in IMCD cells, xestospongin C might reduce the calcium load in the ER and alter the dynamic of intracellular Ca²⁺ release.

In addition to RyR-gated Ca²⁺ release, AVP-induced Ca²⁺ oscillation is maintained by Ca²⁺ influx in IMCD cells. Removal of extracellular Ca²⁺ abbreviates the sustained repetitive AVP-induced Ca²⁺ oscillation to a transient Ca²⁺ release (45). We found that AVP activates a nifedipine-insensitive Ca²⁺ entry pathway, which is similar to the Ca²⁺ entry activated by thapsigargin in IMCD. It is blocked by the nonselective cation channel blocker SKF-96365 and is sensitive to a low concentration of Gd³⁺, bearing resemblance to SOCE. However, 1 μM Gd³⁺ could only reduce Ca²⁺ influx induced by thapsigargin and AVP, while the same concentration of Gd³⁺ abolishes SOCE in other cell types (17, 36). These results suggest that the SOCE in IMCD could be different from the classical SOCE, which is mediated by Ca²⁺ release-activated Ca²⁺ (CRAC) channels (36).

Previous studies showed that depletion of intracellular Ca²⁺ stores activates capacitative Ca²⁺ entry to replenish the Ca²⁺ content of ER/SR. Both RyR- and IP₃R-gated Ca²⁺ stores have been implicated in SOCE. Recent studies demonstrated that the stromal interaction molecule 1 (STIM1) operates as the endoplasmic Ca²⁺ sensor (24, 35, 37). Upon Ca²⁺ depletion of ER, STIM1 proteins aggregate, translocate, and couple to Orai1 as well as other associated proteins including transient receptor potential (TRP) channels in the plasma membrane to activate Ca²⁺ entry (21, 25, 31, 42). TRPC3 and TRPC6 are expressed in rat IMCD (15). TRPC3 is found in apical membrane and is colocalized with AQP2 in vesicles (16). However, SOCE and its molecular counterparts have not been characterized in IMCD. Nevertheless, the prominent Ca²⁺ entry signal elicited by thapsigargin and AVP with similar sensitivity to SKF-96365 and Gd³⁺ suggests that a Ca²⁺ entry mechanism resembling SOCE is an important Ca²⁺ pathway in IMCD cells. Our observations, however, did not exclude the possibility that other nonselective cation channels may also be involved in the AVP-induced Ca²⁺ entry. In conclusion, our present observations in conjunction with those in previous studies have unraveled a cAMP-cADPR-RyR-Ca²⁺ influx signaling system that may play a central role in the regulation of water permeability in IMCD.

GRANTS

This study was supported by National Institutes of Health Grants R01-DK-60501 to K.-P. Yip and HL-071835 to J. S. K. Sham.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).