INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The phosphoinositide pathway producing myo-inositol-1,4,5-trisphosphate (IP₃) is involved in the release of Ca²⁺ ions from the intracellular stores in several nonexcitable cell types. In hepatocytes, hormones using this pathway include -adrenergic hormones, vasopressin, angiotensin II, and adenosine nucleotides (ATP, ADP, AMP), among others, and the response evoked by these hormones is a periodical variation of cytosolic Ca²⁺ concentration called Ca²⁺ oscillations (Woods et al., 1986; Rooney et al., 1989), recently reviewed by Thomas et al. (1996). Above a certain concentration of agonist Ca²⁺ no longer oscillates, but remains chronically elevated.

The generation of Ca²⁺ oscillations has been intensely investigated at the theoretical level (Stucki and Somogyi, 1994; Sneyd et al., 1995). Based on the growing experimental data, the most recent mathematical models involve the participation of intracellular Ca²⁺ stores as a major source of Ca²⁺ for the oscillatory process. After binding of the agonist to its membrane receptor, the sequential activation of the phosphoinositide pathway components, i.e., G proteins, phospholipase C, and production of IP₃, results in the activation of the IP₃ receptor/Ca²⁺ channel located in the membrane of intracellular Ca²⁺ stores. Once activated, the IP₃ receptor/Ca²⁺ channel allows the rapid release of Ca²⁺ into the cytosol, which lasts until feedback mechanisms close the channel, allowing the cell to restore both cytosolic Ca²⁺ and intracellular store Ca²⁺ to their resting value. In the continuous presence of agonist, the Ca²⁺ spiking repeats itself in a periodical way at a stable frequency that is mainly determined by the agonist concentration.

Continuous G-protein activation (Osipchuk et al., 1990) or continuous IP₃ infusion (Petersen et al., 1991; Wakui et al., 1989) have been shown to be sufficient ingredients for induction of Ca²⁺ oscillations in pancreatic acinar cells in whole-cell patch-clamp experiments. A nonmetabolized analog of IP₃ was also shown to elicit Ca²⁺ oscillations in these cells (Wakui et al., 1989), which constitutes a strong argument against a possible role of pulsatile production or pulsatile metabolism of IP₃. In rat hepatocytes, however, the situation might not be identical, because Ca²⁺-induced Ca²⁺ release, if present, does not represent a major mechanism of Ca²⁺ release, as in pancreatic acinar cells (Nathanson et al., 1992).

A higher agonist concentration is likely to result in a higher rate of IP₃ release in the cytosol and, consequently, a more pronounced stimulation of the IP₃ receptor. The present study investigates at the experimental level whether the frequency modulation imposed by the level of agonist is eventually controlled by the level of cytosolic IP₃ in rat hepatocytes. To this end, the effects of rapid release of IP₃ were investigated in nonstimulated hepatocytes and during agonist-induced trains of Ca²⁺ oscillations.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Rat hepatocytes were isolated from male Wistar rats and attached to collagen-coated glass coverslips as described previously (Ubl et al., 1994).

Fluorescence microscopy

The fluorescence microscopy setup was composed of an image intensifier coupled to a video camera (Videoscope International, Washington, DC) attached to an inverted microscope (Nikon, Tokyo, Japan). Cells were observed with a 100×/1.3 N.A. objective lens (Zeiss, Germany). A 100-W xenon lamp (Nikon) and a fluorescence filter wheel (Sutter Instruments, Novato, CA) were used to excite the specimen. The filter cube contained a 510-nm dichroic mirror and a 520-nm long-pass emission filter. The video image was digitized with an 8-bit image processor (Leutron, Glattburg, Switzerland), and a custom-made computer program allowed the selection of regions of interest and intensity measurement inside single cells. Fluorescence intensity measurements were made every 170-200 ms for kinetics measurements, or every 5-8 s during long-lasting trains of oscillations to minimize the illumination time.

Microinjection and flash photolysis

For cytosolic Ca²⁺ measurements the fluorescent probe Fluo-3 was chosen because its visible excitation at 480 nm does not cause unwanted photolysis of photolabile IP₃ analogs. Single isolated hepatocytes attached to coverslips were microinjected using pipettes beveled to ~50 M (BV-10 Microelectrode Beveler, Sutter Instruments) and back-filled with an injection solution containing 2 mM Fluo-3 and 1.5 mM myo-inositol-1,4,5-trisphosphate P⁴⁽⁵⁾-1-(2-nitrophenyl)ethyl ester (caged IP₃) or 1.4 mM 1-(-glycerophosphoryl)D-myo-inositol-4,5-diphosphate P⁴⁽⁵⁾-1-(2-nitrophenyl) ethyl ester(caged GPIP₂). Pressure injection was performed with the Eppendorf 5246 microinjection system (Eppendorf, Hamburg, Ger-many). The cells were then allowed to recover for 20-30 min before the experiments were started. The injected cells on coverslips were then transferred to the fluorescence microscope. Experiments were performed at room temperature.

The intensity of Fluo-3 signals at resting cytosolic Ca²⁺ levels for different intensities of illumination and image intensifier gain were compared to signals obtained for known Fluo-3 concentrations in whole-cell patch-clamp experiments (Chatton et al., 1995). Based on the calculated dilution of Fluo-3, the final caged IP₃ concentration in the injected cells was estimated to be 20-25 µM.

A high-intensity xenon flashlamp system (XF-10; Hi-Tech Scientific, Salisbury, England) was used to deliver flashes of UV light with a duration of 1 ms (Rapp and Güth, 1988). A UG-11 filter was inserted in the light path to select a band of UV light (300-390 nm), and a liquid light guide with quartz focusing lens was used to bring the light to the microscope stage and illuminate the cells with a 2-3-mm² spot.

The flash energy was adjusted by changing both the charge voltage and the combination of capacitors, to produce a maximum calculated output energy of 331 J. The timing of flash-lamp triggering and image acquisition was controlled by computer. A cytosolic Ca²⁺ response could be routinely observed after up to a dozen of consecutive flashes with caged IP₃, indicating that only a small fraction of the pool of caged compound injected underwent photolysis during one flash.

Solutions and materials

Bath solution was Ham's F12 culture medium containing 10 mM HEPES and 1.8 mM CaCl₂ and set to pH 7.4, or a similar solution containing (in mM) 145 NaCl, 5.6 KCl, 0.8 MgSO₄, 1.8 CaCl₂, 10 HEPES, 10 glucose. The pipette solution for microinjection contained (in mM) 138 K-glutamate, 8 NaCl, 5 MgCl₂, 1 EGTA, 0.5 CaCl₂, 10 HEPES and was adjusted to pH 7.2.

Caged IP₃ and (caged GPIP₂) were from Calbiochem (La Jolla, CA). Fluo-3 (K⁺ salt) was from Molecular Probes (Eugene, OR). ADP, Ham's F12 medium, and bovine serum albumin were from Sigma (St. Louis, MO). The other chemicals were from Fluka (Buchs, Switzerland).

Expression of data

Data are presented as means ± SE. When appropriate, a nonpaired Student's t-test was performed to assess the statistical significance of differences.

Mathematical model of Ca²⁺ oscillations

In this study, we consider a minimal one-pool model that is an extended version of a previously published one (Somogyi and Stucki, 1991), in which the termination of a spike was caused by an emptying of the intracellular Ca²⁺ stores. Simultaneous measurements of Ca²⁺ in the cytosol and in the intracellular stores did not agree with this prediction (Chatton et al., 1995): the termination of a spike was not accompanied by an emptying of the stores, which pointed to an active participation of the IP₃ channel in the mechanism of spike termination. Therefore, we now consider the IP₃ channel to be in one of three states: closed, open, or inactivated (Fig. 1). The transition of the channel from the closed to the open state is brought about by increasing IP₃ and an autocatalytic effect of cytosolic Ca²⁺ on the opening probability of the channel. At higher concentrations of cytosolic Ca²⁺, the channel tends to become inactivated. These assumptions are supported by the bell-shaped activation-inactivation characteristics of the channel by cytosolic Ca²⁺ (Bezprozvanny et al., 1991). It is most important to assume that the transition from the inactivated to the closed channel at low basal cytosolic Ca²⁺ is a slow process. There must be a refractory period in this transition that allows cytosolic Ca²⁺ to return to its basal level before the system becomes excitable again and exhibits oscillations of a limit-cycle type. Conversely, increasing the agonist stimulation would lead the system to an overstimulated state (i.e., chronically elevated cytosolic Ca²⁺), the transition to which may show quasisinusoidal and damped oscillations (Somogyi and Stucki, 1991). In essence, the present model is a variant of the model formulated by Keizer's group (De Young and Keizer, 1992).

View larger version (24K): [in this window] [in a new window]   FIGURE 1   Model of Ca2+ oscillations in rat hepatocytes. The model is based on the three functional states of the IP3 receptor/Ca2+ channel: open (O), closed (C), inactivated (I). Plain arrows indicate fluxes, dotted arrows indicate stimulatory or inhibitory action. H, hormone; R, receptor; G, G-proteins; PLC, phospholipase C; PIP2, phosphatidyl-myo-inositol-4,5-bisphosphate; Ca2+cyt, cytosolic Ca2+; Ca2+IS, intracellular store Ca2+; , rate of Ca2+ entry; , rate of active Ca2+ extrusion; k, rate of active Ca2+ pumping into the intracellular stores; k', rate of Ca2+ leak from the stores.

The second modification of the previous model consisted of explicitly introducing the formation and degradation of IP₃. At constant hormone concentration, IP₃ reaches a steady-state concentration sufficiently elevated to trigger oscillations or overstimulation. Application of a pulse of UV light to uncage IP₃ can be simulated by transiently perturbing the IP₃ level. The same goes for caged GPIP₂, which is assumed to react with the same receptor site on the IP₃ receptor/Ca²⁺ channel, but with an affinity and degradation rate different from that of IP₃.

The pertinent equations used in the model are shown in Table 1. A comprehensive analysis of the dynamic properties of this model as well as a linear stability analysis of its steady states have been performed as previously published (Stucki and Somogyi, 1994; Stucki, 1978) and will be presented in detail elsewhere.

All programs were coded in either Fortran (Absoft, Rochester Hills, MI) or Mathematica (Wolfram Research, Champaign, IL), and calculations were performed on a Macintosh 9500 computer.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Characterization of caged IP₃- and caged GPIP₂-induced Ca²⁺ response

In a first step, the response of isolated rat hepatocytes to rapid photorelease of IP₃ or its slowly metabolized analog GPIP₂ was characterized. Cells were individually microinjected with photolabile IP₃ analogs and Fluo-3 before the experiments. The amount of caged IP₃ or caged GPIP₂ injected (20-25 µM) was sufficient to obtain a Ca²⁺ response from multiple UV flashes.

Fig. 2 shows the typical time course of an IP₃-induced and GPIP₂-induced Ca²⁺ transient. To compare the two types of transients, the characteristic kinetic parameters were analyzed for each peak. The time delay (lag time) between the flash and the start of Ca²⁺ increase was measured. To estimate the rate of Ca²⁺ increase, the time from the end of the lag period to the half-maximum Ca²⁺ concentration (t_50%) and to the peak maximum (t_100%) were both measured. As can be seen in the graphs and in Table 2, both the lag time after photolysis of the two caged analogs and the rate of Ca²⁺ concentration increase are similar. In the case of IP₃-induced transients, the recovery of Ca²⁺ concentration followed single exponential kinetics with correlation coefficients greater than 0.990. As can be seen in Fig. 2, the kinetics of recovery was markedly different after a GPIP₂-induced spike. The recovery generally consisted of two phases: a first rapid phase followed by a slow phase, where Ca²⁺ gradually returned to baseline. The curves were fitted with a double exponential decay equation with mean correlation coefficients of 0.991. The kinetic parameters of the Ca²⁺ transients observed after photorelease of IP₃ and GPIP₂ are summarized in Table 2. All kinetic parameters of IP₃- and GPIP₂-induced transients were the same statistically, except the second phase of the GPIP₂ decay.

View larger version (11K): [in this window] [in a new window]   FIGURE 2   Kinetics of Ca2+ response after photolysis of caged IP3 or caged GPIP2. Isolated rat hepatocytes were microinjected with the Ca2+-sensitive dye Fluo-3 and photolabile analogs of IP3. Typical responses are presented for two different cells microinjected with caged IP3 (left) or caged GPIP2 (right), respectively. One-millisecond flashes of UV light were applied at times indicated by the closed triangles in the graph. Ca2+ signals were fluorimetrically measured using Fluo-3. Kinetic parameters and statistics of the two types of responses are summarized in Table 2.

We next investigated the effects of flash intensity, i.e., of the amount of IP₃ delivered, on the Ca²⁺ response. A series of five flashes of increasing energy in the range 50-220 J were applied to single cells from four experiments. The kinetics parameters of Ca²⁺ transients revealed that the lag time, t_50% and t_100%, decreased with increasing flash intensity (data not shown). Increasing the energy of flash also increased the amplitude of spiking. Fig. 3 A shows the spike amplitude plotted against the energy of flash. Because of the variation in the amount of Ca²⁺ dye Fluo-3 injected, the ordinate has been normalized to 100% at 151 J. The graph shows that, in the range of energy considered, a quasilinear relationship exists between the amplitude of Ca²⁺ spiking and the energy delivered during a flash. It should be noted that below a certain threshold of energy, ~40 J, no Ca²⁺ transient was evoked, and there was a tendency toward plateauing at higher energy values (data not shown), which is consistent with either a saturation of IP₃ receptors or a maximum Ca²⁺ release from the intracellular stores. It should also be noted that each flash depletes the pool of available caged IP₃ by an amount proportional to the flash intensity, which could cause the steepness of the amplitude-energy relationship to be underestimated somewhat.

View larger version (11K): [in this window] [in a new window]   FIGURE 3   Relationship between Ca2+ spike amplitude and flash energy. (A) The energy of electrical discharge during flashes (Joules) was varied while measuring IP3-induced Ca2+ transients using Fluo-3. Because of the variability in the amount of fluorescent dye present in the cells, the amplitude of Ca2+ transients was normalized to 100% at 151 J. (B) The rate of decay (s1) normalized to the rate measured at 151 J is plotted against the Ca2+ spike amplitude. The decay data were obtained from the experiments described in A. Data are means ± SE from four separate preparations.

These experiments showed that, as was observed with guinea pig hepatocytes (Chiavaroli et al., 1994), in rat hepatocytes, the amplitude of Ca²⁺ spikes is dependent on the amount of IP₃ delivered, whereas increasing the dose of an IP₃-producing agonist like phenylephrine does not influence the amplitude of spiking, but only the frequency of oscillations.

In this series of experiments, the recovery phase of Ca²⁺ spikes was found to occur faster with increasing flash energy. As a direct consequence (Fig. 3 B), the rate of Ca²⁺ recovery was found to be proportional to the amplitude of the Ca²⁺ spikes. This observation is consistent with negative feedback imposed by Ca²⁺ on its own release from the stores by the IP₃ receptor (Bezprozvanny et al., 1991).

Photorelease of IP₃ and GPIP₂ during agonist-induced Ca²⁺ oscillations

As discussed above, increasing the dose of several hormones administered to oscillating cells leads to a frequency increase in Ca²⁺ oscillations. An increase in agonist concentration should induce a stronger activation of G proteins and phospholipase C, and ultimately result in a higher production of intracellular IP₃. The next series of experiments was designed to investigate the effects of a rapid perturbation of the IP₃ concentration during sustained hormone-induced Ca²⁺ oscillations.

Single hepatocytes microinjected with Fluo-3 and caged IP₃ were stimulated with the -adrenergic agonist phenylephrine at a concentration adequate to obtain cytosolic Ca²⁺ oscillations (Fig. 4 A). After a few regularly spaced oscillatory transients, a flash of UV light was applied while the Ca²⁺ concentration was at its resting level, i.e., during the interval between transients. As can be seen in the figure, the photorelease of IP₃ (arrowheads) evoked a fast Ca²⁺ spike, and when Ca²⁺ had recovered, normal oscillatory spiking resumed. Fig. 4 B depicts the frequency variations (min¹) of Ca²⁺ oscillations, calculated from the time intervals between Ca²⁺ maxima, along the experiment. The open symbols and arrows indicate the IP₃-triggered spikes. In the example shown, the second flash was triggered as soon as Ca²⁺ had reached the baseline after the oscillatory peak, whereas the first flash occurred somewhat later in the interval. In both cases, however, the Ca²⁺ peak that followed the IP₃-induced spike and the subsequent ones came after the same time interval as was found before the flashes. Thus, photorelease of IP₃ induced a Ca²⁺ spike and reset the phase of oscillations without altering the frequency.

View larger version (17K): [in this window] [in a new window]   FIGURE 4   Photorelease of IP3 during phenylephrine-induced cytosolic Ca2+ oscillations. (A) Hepatocytes injected with caged IP3 were stimulated with phenylephrine to elicit sustained Ca2+ oscillations, measured with Fluo-3. At times indicated by the set of closed triangles, flashes of UV light were applied and the released IP3 immediately induced Ca2+ mobilization. The following oscillatory spike came with a delay that corresponded to the period set by the agonist. (B) The frequency of oscillation (min1) calculated from the period between Ca2+ spikes in the experiment shown in A is plotted for each measured Ca2+ transient. Closed circles are agonist-induced oscillatory transients. Open circles and arrows indicate caged IP3-induced Ca2+ spikes. The oscillatory Ca2+ transient that followed the flash-induced spike arrived with the same frequency as the initial frequency. A representative experiment from a total of four is shown.

To examine whether this feature was specific to -adrenergic hormones, two other agonists able to elicit Ca²⁺ oscillations but acting on different classes of receptors were tested, namely adenosine diphosphate (ADP), an agonist of the P_2y purinoceptor (Dixon et al., 1990), and (Arg⁸) vasopressin (vasopressin), which activates V₁ receptors in hepatocytes (Dasso and Taylor, 1994). As can be seen in Figs. 5 and 6, the same phenomenon was observed when the agonist was ADP (n = 4 experiments) or vasopressin (n = 4 experiments). These experiments indicated that the observed phase shift of oscillations produced by IP₃ photorelease and its lack of effect on the frequency of Ca²⁺ oscillations were not restricted to Ca²⁺ oscillations induced by -adrenergic stimulation, but appeared to apply to all agonists that use the phosphoinositide signaling pathway in hepatocytes.

View larger version (16K): [in this window] [in a new window]   FIGURE 5   Photorelease of IP3 during ADP-induced Ca2+ oscillations. Experiments were performed using the same protocol as in Fig. 4, but with ADP as the agonist. The figure shows typical traces from four experiments.

View larger version (20K): [in this window] [in a new window]   FIGURE 6   Photorelease of IP3 during vasopressin-induced Ca2+ oscillations. Experiments were performed using the same protocol as in Fig. 4, but with vasopressin as the agonist. The figure is a representative experiment of four.

We then asked whether the extent of phase shift introduced by IP₃ photorelease depended on when it was applied in the interval between oscillatory transients. Fig. 7 shows that IP₃ photorelease triggered at different time points after an oscillatory transient produced a shift of ~100% of the basal frequency, regardless of whether it occurred just after an oscillatory transient or near the end of the normal period imposed by the agonist concentration.

View larger version (13K): [in this window] [in a new window]   FIGURE 7   Relationship between the extent of phase shift and the position of IP3 photorelease in the interval of Ca2+ oscillations. The amount of phase shift in percentage of the oscillatory frequency imposed by the hormone is plotted against the position of IP3 photorelease within the interval. The position of IP3-induced spike relative to the preceding oscillatory Ca2+ transient was scaled to the basal interval imposed by the agonist. Data are pooled from experiments using the agonists phenylephrine, ADP, or vasopressin.

The fact that photorelease of IP₃ did not produce any increase in the frequency of Ca²⁺ oscillations led us to consider the possibility that the degradation of the IP₃ molecules photoreleased was occurring rapidly enough to prevent any frequency effect. If this were really the case, one would expect to obtain different results by using nonmetabolized analogs of IP₃.

The same protocol as in Figs. 4-6 was therefore applied with hepatocytes microinjected with caged GPIP₂. Stimulation with phenylephrine evoked sustained Ca²⁺ oscillations, and when a flash of UV light was triggered in the same range of energy as for caged IP₃ experiments, an immediate overstimulated Ca²⁺ response was observed in five experiments (data not shown). Such a response is typically found when the agonist concentration exceeds a certain threshold, where Ca²⁺ oscillations turn into a chronic elevation of cytosolic Ca²⁺ (also called overstimulation). The flash intensity was therefore reduced and the same protocol repeated. Fig. 8 A shows that a response considerably different from that of the caged IP₃ experiments was obtained. First, the oscillatory Ca²⁺ transient that directly followed the GPIP₂-induced spike came with a shorter delay than the regular oscillatory period. Second, the frequency of the oscillatory peaks that followed GPIP₂ photorelease was higher compared with the basal frequency and gradually decreased to the original value imposed by the agonist concentration (Fig. 8 B).

View larger version (21K): [in this window] [in a new window]   FIGURE 8   Photorelease of GPIP2 during phenylephrine-induced cytosolic Ca2+ oscillations. (A) Hepatocytes injected with caged GPIP2 were stimulated with phenylephrine to elicit sustained Ca2+ oscillations, measured with Fluo-3. At times indicated by the set of closed triangles, flashes were applied and the released GPIP2 promptly induced Ca2+ mobilization. The following oscillatory spikes came with a shorter delay than the one set by the agonist. The figure depicts a representative experiment from a total of seven. (B) The frequency of oscillation (min1) calculated from the period between Ca2+ spikes in the experiment presented in A is plotted for each measured Ca2+ transient. Closed circles represent agonist-induced oscillatory transients and open circles and arrows indicate caged GPIP2-induced spikes. The oscillatory Ca2+ transients that follow the GPIP2-induced spikes have an increased frequency that progressively declined back to the initial frequency.

To quantify these effects, the mean frequency of Ca²⁺ oscillations induced by phenylephrine was calculated from the oscillatory period before photolysis of the caged compounds and compared with the period between the IP₃- or GPIP₂-induced Ca²⁺ spike and the next oscillatory transient. Whereas the photorelease of IP₃ did not produce any increase in frequency (103.3 ± 2.7%, n = 5 experiments, 9 flashes), GPIP₂ significantly increased the frequency to 134.3 ± 6.3% (n = 7 experiments, 15 flashes, p = 0.001).

We postulate that the phase shift produced by IP₃, its lack of effect on the frequency of oscillations, and the effects of GPIP₂ can be decomposed into two phenomena: 1) the transient elevation of Ca²⁺ induced by the IP₃ delivery puts the IP₃ receptor in an inactivated state (i.e., resets the phase of oscillations), and 2) the degradation of IP₃ occurs at the same rate as or faster than the recovery of Ca²⁺ to the baseline, meaning that the level of IP₃ has decayed to the level imposed by the agonist, while the IP₃ receptor was still in an inactivated state.

To test whether these hypotheses are sound, we used the mathematical model of Ca²⁺ oscillations described above and simulated the photorelease of IP₃ and GPIP₂, taking into account their rates of degradation. Fig. 9 shows the results of such simulations, where the rate of degradation of GPIP₂ was set to a 17-fold slower value than IP₃. Cytosolic Ca²⁺ oscillations started as soon as IP₃ was increased to its elevated steady state. Cytosolic oscillations were accompanied by periodic depletion of the intracellular store. Toward the end of the presented simulation, the Ca²⁺ pumping into the stores was stopped, which emptied the stores. As can be seen in the graph, the stores were only partially depleted during oscillations, as was observed experimentally (Chatton et al., 1995), indicating that spike termination is not caused by the emptying of the stores. These simulations yielded results strikingly similar to those of their experimental counterpart: the simulated IP₃ release inserted a phase shift without accelerating the oscillations, whereas a simulated GPIP₂ release resulted in a transient frequency increase in the oscillations.

View larger version (29K): [in this window] [in a new window]   FIGURE 9   Computer simulations of IP3 and GPIP2 photorelease during agonist-induced Ca2+ oscillations. Variations of intracellular store Ca2+ and cytosolic Ca2+ are shown in the upper and middle panels, respectively. The lower panel represents the time course of IP3 (dotted trace) and GPIP2 (bold trace) changes. Toward the end of the simulated experiment, inhibition of the intracellular store Ca2+ pump was simulated, causing an emptying of the stores, a concomitant Ca2+ release into the cytosol, and the immediate cessation of the oscillations. All data are in arbitrary units. The following parameters were used in the simulations: k' = 0.01, k = 3.0, kp = 0.45, kd = 0.12,  = 1.0,  = 3.0, a0 = 2.0, Ka = 15.0, n = 3, b0 = 5.0, Kb = 0.2, m = 2, c = 0.001, d = 0.01,  = 50 (see Fig. 1 and Table 1 for details on the model).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The purpose of this study was to investigate the role of IP₃ in setting the frequency of agonist-induced cytosolic Ca²⁺ oscillations in rat hepatocytes. This study has directly demonstrated, by using transient as well as more prolonged perturbation of IP₃ receptor stimulation, that the level of IP₃ indeed influences agonist-induced Ca²⁺ oscillations.

Kinetic properties of IP₃ and GPIP₂-induced Ca²⁺ mobilization

Photolysis of caged IP₃ produced Ca²⁺ spikes that had both faster activation kinetics and a faster return to baseline than agonist-induced Ca²⁺ oscillatory transients. This different behavior can be explained by the photochemical reaction that releases IP₃ with a half-time of 3 ms or less (Walker et al., 1987), which is orders of magnitude more rapid than the phospholipase C-mediated production of IP₃ (Chiavaroli et al., 1994) in hepatocytes. In addition, in a bolus release of IP₃, as opposed to a continuous IP₃ production by the phospholipase C, one can expect that IP₃ will disappear quickly. Metabolism of IP₃ has been shown to occur with a half-time of a few seconds (Hughes et al., 1988; Wang et al., 1995; Bird et al., 1992). The half-life of Ca²⁺ recovery measured in the present study (5.7 sec) is very close to the half-life of IP₃ metabolism of 7.6 s measured in rat parotid acinar cells (Hughes et al., 1988) and of 9 s obtained in neuroblastoma cells (Wang et al., 1995). Therefore, the recovery could be explained, in principle, solely by the degradation of the photoreleased IP₃.

A strikingly different Ca²⁺ response was observed after rapid photorelease of GPIP₂. GPIP₂ appears to possess an affinity for the IP₃ receptor about one order of magnitude lower than that of IP₃, but to be capable nevertheless of maximally activating Ca²⁺ signaling (Bird et al., 1992). In our experiments, photorelease of GPIP₂ resulted in Ca²⁺ signals with activation kinetics indistinguishable from that of IP₃, but with recovery kinetics consisting of more than one phase. The first phase of recovery had a half-time similar to that of the IP₃ recovery. A second phase occurred with ~25-fold slower kinetics. Very similar kinetics values were observed with guinea pig hepatocytes (Wootton et al., 1995) when another nonmetabolized IP₃ analog, 5-thio-IP₃, was used. A complex Ca²⁺ sensitivity of the IP₃ receptor has been demonstrated with isolated receptors (Bezprozvanny et al., 1991) and various intact cell preparations (Oancea and Meyer, 1996): an increasing Ca²⁺ concentration activated the IP₃ receptor up to a critical concentration, where Ca²⁺ turns into an inhibitory factor. We hypothesize that the initial phase of recovery from photorelease of GPIP₂ originated from a rapid but partial inactivation of the IP₃ receptor caused by the elevated Ca²⁺. The second phase would be due primarily to the slow metabolism of GPIP₂, which occurs over several minutes (Bird et al., 1992).

Because the first phase of GPIP₂ recovery occurred at the same rate as the recovery in an IP₃-induced Ca²⁺ transient, we propose that the Ca²⁺-induced inactivation of the IP₃ receptor and the metabolism of IP₃ have the same or very close kinetics and could not be distinguished in our measurements.

Perturbation of cytosolic IP₃ concentration during agonist-induced Ca²⁺ oscillations

It has been shown in many investigations involving hepatocytes as well as other types of cells that stimulation with certain agonists leads to cytosolic free Ca²⁺ oscillations, the frequency of which is determined by the concentration of the applied agonist (Thomas et al., 1996). This frequency modulation rather than amplitude modulation of signaling is thought to play a role in certain cellular processes. Increasing the dose of agonist, which results in frequency increase, should be accompanied by a stronger activation of G proteins, phospholipase C, and ultimately a higher level of cytosolic IP₃. Most of the recent models describing Ca²⁺ oscillations would predict that the parameter eventually responsible for setting the pace of oscillations is the level of IP₃ (Stucki and Somogyi, 1994), determined by the counteracting processes of generation and degradation of IP₃.

To investigate this issue experimentally, we took advantage again of the technique of flash photolysis, with microinjected photolabile precursors of IP₃ to perturb the cytosolic level of IP₃ during sustained Ca²⁺ oscillations. Photolysis of caged IP₃ triggered during the interval between oscillatory transients resulted in the generation of a rapid Ca²⁺ spike. This observation indicated that, in rat hepatocytes, if a refractory state exists between two oscillatory transients, it has to be considered partial and reversible. Unexpectedly, photorelease of IP₃ during Ca²⁺ oscillations had no influence on the frequency of oscillations, but had the effect of introducing a phase shift in the oscillations or, in other words, of resetting the clock. The influence of IP₃ photorelease on the oscillatory pattern can be due to a direct action of IP₃ on its intracellular receptor as well as to an indirect one through elevated cytosolic Ca²⁺ that followed the IP₃ release. We observed that rapid photorelease of IP₃ consistently introduced a shift of the same duration as the period set by the agonist concentration, regardless of when it was applied in the interval between two oscillatory transients (i.e., just after a transient, in the middle or near the end of the period). One can therefore conclude that its effect is to reset the IP₃ receptor to its state of maximum inactivation.

The fact that photorelease of IP₃ did not increase the frequency of Ca²⁺ oscillations speaks again in favor of a highly active metabolism of IP₃ (Hughes et al., 1988; Wang et al., 1995) that is capable of restoring the level of IP₃ to its "preflash" level, probably at the same time as or before Ca²⁺ has returned to baseline. The implication is that the parameter that resets the phase of oscillations is probably the cytosolic Ca²⁺ concentration itself. As seen in Fig. 3 B as well as in isolated IP₃ receptors (Bezprozvanny et al., 1991), an elevated cytosolic Ca²⁺ has the tendency to inactivate the release of Ca²⁺ from the stores. Another indication that this effect also occurs in rat hepatocytes came from recent experiments (Chatton et al., 1997), during which photorelease of IP₃ during maximum -agonist stimulationwhere cytosolic Ca²⁺ remains chronically elevatedinduced a Ca²⁺ spike above the elevated baseline. This spike was followed by an undershooting of the Ca²⁺ signal and brief oscillations around the elevated steady state. We believe that the observed undershooting is consistent with the Ca²⁺ spike causing a transient inactivation of the IP₃ receptor.

If the lack of effect of caged IP₃ on the oscillatory frequency was due to its fast metabolism, then releasing a nonmetabolized analog of IP₃ should affect the frequency of oscillations. Caged GPIP₂ was chosen for this purpose, even though it undergoes a slow degradation, ~200-fold slower that IP₃ in mouse lacrimal acinar cells (Bird et al., 1992), but probably somewhat faster in rat hepatocytes. For computer simulations, we used a GPIP₂ degradation rate 17-fold slower than that of IP₃, based on the kinetics of recovery presented in Table 2.

Photorelease of GPIP₂, indeed, resulted in a very different type of response during agonist-induced Ca²⁺ oscillations. First, an overstimulatory response was readily induced when the energy of photolysis was a little too high, which has never been observed with caged IP₃. Second, when overstimulation was avoided by reducing the amount of light delivered, an increase in frequency of the Ca²⁺ transients after the photorelease was observed. This acceleration was transient, however, and lasted up to ~6 min, consistent with the expected rate of GPIP₂ metabolism. Computer simulations based on a minimal model that takes into account the respective degradation rates of IP₃ and GPIP₂ effectively reproduced the observed cell responses to photorelease of IP₃ analogs.

Taken together, our experiments, supported by mathematical modeling, demonstrated that an enhanced activation of IP₃ receptors increases the frequency of oscillations and indicated that a refractory state is initiated by a Ca²⁺ spike. The duration of this state depends on the amount of IP₃ present. Thus the level of IP₃ directly influences the frequency of oscillations, the elevated Ca²⁺ being responsible mainly for putting the system into a state of reversible inactivation.