Mechanisms underlying InsP₃-evoked global Ca²⁺ signals in mouse pancreatic acinar cells

Abstract

1. In secretory epithelial cells, complex patterns of Ca²⁺ signals regulate physiological processes. How these patterns are generated is still not fully understood. In particular, the basis of global Ca²⁺ waves is not clear.

2. We have studied regional differences in InsP₃-evoked Ca²⁺ release in single mouse pancreatic acinar cells, using high-speed (∼90 frames s⁻¹), high-sensitivity Ca²⁺ imaging combined with rapid (10 ms) spot photolysis (2 μm diameter) of caged InsP₃. Within a single region we measured Ca²⁺ response latency and rate of rise to construct an InsP₃ dose-response relationship.

3. Spot InsP₃ liberation in the secretory pole region consistently elicited a dose-dependent, rapid release of Ca²⁺.

4. Spot InsP₃ liberation in the basal pole region of ∼50 % of cells elicited a similar dose-response relationship but with a lower apparent InsP₃ affinity than in the secretory pole. In the other cells, basal pole InsP₃ liberation did not elicit active Ca²⁺ release, even at the highest stimulus intensities we employed, although these same cells did respond when the stimulus spot was moved to different regions.

5. We conclude that in the basal pole active sites of rapid Ca²⁺ release have a lower functional affinity for InsP₃ than those in the secretory pole and are spread out in discrete sites across the basal pole. These properties explain the propagation of Ca²⁺ waves across the basal pole that are only observed at higher stimulus levels.

In secretory epithelial cells, agonists evoke complex Ca²⁺ responses (Kasai & Augustine, 1990; Toescu et al. 1992; Thorn et al. 1993) thought to be important in cell physiology (Petersen, 1992; Thorn, 1996). These responses include local Ca²⁺ release within the secretory (apical) pole region and, at higher stimulus intensities, global Ca²⁺ waves and oscillations that are initiated in the secretory pole and then spread as a wave across the cell to the basal pole (Kasai & Augustine, 1990; Toescu et al. 1992; Nathanson et al. 1992; Kasai et al. 1993; Thorn et al. 1993; Lee et al. 1997). The local Ca²⁺ signal in the secretory pole has now been shown to be due to the co-ordinated activity of a cluster of smaller elementary Ca²⁺ spikes (Thorn et al. 1996; Kidd et al. 1999; Fogarty et al. 2000). The basis of these local spikes appears to be the regional location of InsP₃ receptors, which have been immunolocalized to the secretory pole (Lee et al. 1997; Yule et al. 1997). The question that now arises is what is the basis of the global Ca²⁺ signal observed in these cells. We still have no direct evidence of the mechanisms of Ca²⁺ release that are present in the basal pole that might be recruited during a global Ca²⁺ wave. The aim of this paper is to provide an insight into the basal pole Ca²⁺ signal.

Functional evidence indicates that the basal pole region is capable of an active Ca²⁺ response (Kasai et al. 1993; Thorn et al. 1993), suggesting that Ca²⁺ release sites are present. However, immunological staining for InsP₃ receptors suggests a low density of receptors in this region of the cell (Lee et al. 1997; Yule et al. 1997). Furthermore, studies aimed at characterizing the basal pole response (Toescu et al. 1994; van de Put & Elliott, 1996; Ito et al. 1999) have been hampered by the fact that the global signal is always initiated in the secretory pole. This means that it has been impossible to study the basal pole signal in isolation, as it is always contaminated by the Ca²⁺ released from the secretory pole. To overcome this problem, and to study the basal pole signal in isolation, we have developed methods of localized, regional application of InsP₃.

Our results indicate that the basal pole contains discrete regions which are capable of active InsP₃-evoked Ca²⁺ release. This Ca²⁺ response is comparable to that in the secretory pole but has a lower functional affinity for InsP₃. However, other regions of the basal pole show no active InsP₃-evoked Ca²⁺ response. The low affinity and spatial separation of active sites of Ca²⁺ release are proposed to explain why the global Ca²⁺ responses always originate in the secretory pole and only invade the basal pole at higher stimulus intensities.

METHODS

Imaging

Outbred male albino mice were killed by cervical dislocation. Single, and small clusters of, mouse pancreatic acinar cells were prepared by collagenase digestion as previously described (Thorn & Petersen, 1992). Ca²⁺ imaging experiments were performed by microinjection of individual acinar cells with 70 000 MW dextran-coupled Calcium Green and caged InsP₃ (Molecular Probes) calculated to give a final, cellular concentration of 4-20 and 10-50 μm, respectively (these calculated concentrations matched estimates from comparison of the fluorescence of microinjected cells with fluorescence measurements in whole-cell patch clamp experiments where the Calcium Green concentration was known). The high-molecular weight Ca²⁺ indicator spread throughout all regions of the cell. The low concentration and slow diffusion rate of 70 000 MW dextran-coupled Calcium Green minimized artifacts due to diffusional redistribution of the Ca²⁺ signal by the probe. Imaging was carried out using a custom-built, high-speed, high-sensitivity, wide-field digital-imaging microscope (Fig. 1A) as previously described (Kidd et al. 1999). Briefly, single cells were illuminated with a visible laser (Coherent Innova 70, Coherent, Santa Clara, CA, USA) at 488 nm and imaged through a Nikon × 40 UV, 1.3 NA, oil-immersion objective through a 510 nm long-pass filter. Full-frame images (128 pixels × 128 pixels) were captured on a cooled, frame-transfer CCD camera (70 % quantum efficiency, 5 electrons readout noise at 500 frames s⁻¹; Massachusetts Institute of Technology (MIT), Lincoln Laboratory) with a pixel size of 200 nm at the specimen. All images were acquired at 90 frames s⁻¹ (11 ms between images) and with 5 ms exposures. Data were recorded to computer, then analysed with customized software following appropriate bleach correction and smoothing. We assumed a monoexponential decline for the bleaching. Errors in the correction led to minor discrepancies in the ratio signal (for example in Fig. 2 the initial ratio is slightly less than 0). Images were displayed as ΔF/F_o images (100(F–F_o)/F_o), where F is the recorded fluorescence and F_o is the mean fluorescence of the first 40 sequential image frames prior to flash photolysis.

Figure 1. The process of spot photolysis of InsP₃ was modelled in order to estimate the relative [InsP₃] that evoked regional Ca²⁺ release.

A, a simplified diagram of the high-speed microscope. An argon-krypton visible laser (VL; λ= 488 nm, shown in purple), was used to excite Calcium Green, while an Argon-Ion ultraviolet laser (UVL; λ= 351 and 361 nm, shown in pink), was used for photolysis of InsP₃. The UV beam was passed through a 3 mm aperture laser shutter (S), a telescope beam expander optic (BE), a converging lens (L) and a pinhole (P). The visible and UV beams were combined by a dichroic beam combiner (DM, 400DCLP, Chroma), and focused onto the cell through a Nikon × 40 UV, 1.3 NA, oil-immersion objective lens (O). The Calcium Green fluorescence, shown in green, was imaged onto the 128 pixel × 128 pixel image frame (IF) of the high-speed, frame-transfer CCD camera. At the end of an exposure, this image was quickly (∼50 μs) shifted into the storage frame (SF) for transfer to the PC. (Abbreviations: M, mirror; PZT, piezomotor; E, emission filter.) B, the photolysis spot produced within the cells was imaged using the Calcium Green fluorescence excited by the UV beam, and used to determine the spatial profile of the spot. The image shown is the average of five, 1 W, spots from five different cells. Scale bars here and in E and G are 1 μm. Intensity profiles through the average spot, both horizontally (○) and vertically (□) are shown to the left, and were well fitted with a gaussian profile (σ= 0.818 μm) having a full-width at half-maximal intensity (FWHM) of ∼2 μm. C, the actual total (integrated) spot photolysis power seen at the cell was measured and shown to vary linearly with the range of laser power settings used in the experiments. D, the image shows the 2-D gaussian profile of photon density (J cm⁻²) calculated from the total number of UV photons delivered by a 1 W laser setting at the spot in a 10 ms period. E, the fraction of InsP₃ that was uncaged was calculated from the photon density images from five laser power settings using the cross-section and quantum yield of caged InsP₃. For laser settings above 0.05 W, there is a region at the centre of the spot where the photon density is sufficient to uncage all the InsP₃ present. The size of this region increases with the laser power. F, the relationship between laser power and the average, relative [InsP₃] within a 4 μm diameter area, corresponding to the ΔF/F₀ measurement region, is not linear, due to the contribution from the expanding region of photolysis saturation. G, the relationship between laser power and the average, relative uncaged [InsP₃] at different radii from the centre of the spot is shown for five annuli of equal area (as indicated in the inset by the dotted lines at radii of 1, 1.41, 1.73, 2 and 2.25 μm).

Figure 2. The global Ca²⁺ response to global photolytic liberation of InsP₃.

Cells were microinjected with Calcium Green and caged InsP₃. The bright-field image shows the outline of the cell (estimated by eye from the fluorescence image) as a white line. This outline was used as the mask for the ΔF/F_o images. The secretory pole was identified by the clustering of zymogen granules. Cells responded, with a short latency and a secretory to basal pole Ca²⁺ wave, to global photolytic release of InsP₃ (time point indicated by arrow on graphs). Indicated on the graphs are the responses recorded in 2 μm diameter regions in the secretory pole and basal pole. The ΔF/F_o images of Calcium Green fluorescence are proportional to the changes in intracellular [Ca²⁺] and are shown as a pseudocolour representation according to the colour scale on the right. The asterisk indicates the orientation of the cell and is placed in the direction of the secretory pole. Prior to the photolytic release of InsP₃ the resting [Ca²⁺] was low (image 1). The first image that exceeds a threshold of 5 % in the secretory pole is shown (image 2) and subsequent frames were taken at consecutive 36 ms intervals.

In global photolysis experiments we limited the photolysis area (UV spot of ∼20 μm at the image plane) to just the cell of interest. Ca²⁺ responses to spot InsP₃ liberation were recorded as the average ΔF/F_o Calcium Green fluorescence ratio within an area of 4 μm diameter centred on the region of the photolysis spot. The latency of the response was measured as the time taken to reach a 5 % increase in fluorescence intensity. The rate of rise of the Ca²⁺ response within this area was obtained from the first derivative. Spot photolysis of InsP₃ was repeated many times (up to ∼10) in the same cell and gave reproducible responses (see Figs 3B and 5A). Our spot photolysis protocol randomized the order of presentation of the different stimulus intensities to a given cell. We allowed a period of more than 2 min between flashes to allow the cell to recover; this was determined by a return to baseline fluorescence levels.

Figure 3. Control spot photolysis experiments.

A, cells filled via the patch pipette with caged fluorescein were used to test the spatial extent of the stimulus. In these experiments exactly the same spatial stimulus parameters were used as for those with caged InsP₃; that is, a spot of ∼2 μm diameter at the image plane. The first image (left) shows an even fluorescence throughout the cell. Also indicated, by the white circle, is the region of the UV spot (FWHM). An image taken 200 ms after spot photolysis (right) shows a spatially limited increase in fluorescence in the region of the UV flash. B, cells filled with caged InsP₃ were given repeated high-intensity UV spot flashes. Two consecutive Ca²⁺ responses from a 4 μm diameter region, centred over the photolysis region are shown in the graph. The similarity of these responses shows reproducibility and along with the ‘dose’-response relationships shown in Figs 5A, and 6C and D suggests that the cells were not significantly damaged by laser light.

Figure 5. The secretory pole Ca²⁺ dose-response to spot InsP₃ liberation in the secretory pole.

A, in a single cell, repeated photolytic liberation of InsP₃, at different intensities of UV laser light, gave rise to a ‘dose’-dependent increase in the rate of rise and amplitude of the Ca²⁺ response. The order of presentation of the different intensity UV flashes was randomized. In this particular case the order was (laser power (a.u.)): 1, 0.25, 0.12, 0.05, 2, 6. The data show the reproducibility of the responses and saturation at the higher stimulus intensities. The lines on the left-hand side of each trace indicate ΔF/F₀= 0. B, the means of the maximal rates of rise of the fluorescence signal are shown plotted against laser power. Mean latency to a 5 %ΔF/F₀ change is plotted against UV laser intensity, showing a decline to a minimum with increasing light intensities. Error bars are s.e.m.

Spot photolysis

As shown in Fig. 1A, an argon-ion UV laser (Coherent), λ= 351 and 361 nm, was used for photolysis of InsP₃. The UV laser intensity was regulated by manually adjusting the laser power setting to between a minimum of 0.05 W and a maximum of 6 W. A 3 mm aperture electronic shutter (Uniblitz, Vincent Assoc., USA) was used to control a 10 ms photolysis period with 1 ms resolution. The UV beam was then passed through a telescope beam expander optic, a converging lens and a 35 μm pinhole in order to produce a ∼2 μm diameter photolysis spot at the focal plane of the objective. The photolysis spot was imaged using the Calcium Green fluorescence excited by the UV beam, and used to determine the spatial profile of the spot. The image shown in Fig. 1B is the average of five 1 W spots from five different cells. Intensity profiles through the averaged spot were well fitted by a gaussian distribution (σ= 0.818 μm) having a full-width at half-maximal intensity (FWHM) of ∼2 μm. The actual total (integrated) spot photolysis power seen at the cell was measured and shown to vary linearly with the range of laser power settings used in the experiments (Fig. 1C for example – a 1 W setting on the laser resulted in 32 μW at the cell).

In order to estimate the relative concentration of InsP₃ that was uncaged at the focal plane by different laser power settings, a computer model of the photolysis spot was constructed. The total number of UV photons delivered at the spot in the 10 ms period was calculated from the integrated spot power for each of the laser settings used. Then, using the fitted 2-D gaussian spatial profile, an image of the photon density was calculated (Fig. 1D). The fraction of InsP₃ that was uncaged (Fig. 1E) was then calculated from these photon density images. To do this each photon density image (J cm⁻²) was multiplied by 1.8 × 10¹⁸ photons J⁻¹ (λ= 357 nm), by the absorption cross-section σ (Lakowitz, 1983) for caged InsP₃ of 2.3 × 10⁻¹⁸ cm² (σ=ε(3.82 × 10⁻²¹), calculated using an extinction coefficient ε= 600 M⁻¹ cm⁻¹; Dr J. Walker, personal communication), and by the quantum yield for caged InsP₃ of 0.6 (Dr J. Walker).

As can be seen in Fig. 1E, for laser settings above 0.05 W, there is a region at the centre of the spot where the photon density is sufficient to uncage all the InsP₃ present, which experimentally corresponds to 10-50 μm InsP₃. The graph in Fig. 1F shows the average, relative [InsP₃] within the 4 μm diameter area of the spot corresponding to the ΔF/F₀ measurement region. The relationship between laser power and the average [InsP₃] is not linear, due to the contribution from the expanding region of photolysis saturation. The size of this region of saturation increases with the laser power.

Since the location and density of InsP₃ receptors within the cells was unknown, randomly different alignments of the spot profile with the location of discrete patches of InsP₃ receptors would be expected to result in different laser power settings evoking the same responses. This is illustrated in Fig. 1G, where the relationship between laser power and the average, relative uncaged [InsP₃] at different radii from the centre of the spot is shown for five annuli of equal area (as indicated by the dotted lines at radii of 1, 1.41, 1.73, 2 and 2.25 μm). If this analysis was correct, a discrete site of InsP₃-evoked Ca²⁺ release, which happened to be positioned under the exact centre of the spot, would respond maximally with as little as a 0.1 W laser flash. In fact we did observe one such event, where the response to 0.1 W was comparable to other responses obtained at higher laser intensities.

Actual calculations as to the amount of InsP₃ liberated are complex and will be affected by degradation rates (Fink et al. 1999) and diffusional effects. To study diffusion, control experiments with the spot release of caged fluorescein were performed (injected to a final concentration of 10-50 μm inside the cell, fluorescent only after uncaging; Molecular Probes). The results showed a rapid rise in fluorescence, localized to the photolysis region, followed by a slow spread of the wavefront of fluorescence (10 μm s⁻¹) across the cell (Fig. 3A). The molecular weights of InsP₃ and fluorescein are similar (649 and 600, respectively) and therefore the patterns of photolytic liberation and diffusion should also be similar.

In order to investigate whether diffusion would be an important factor in determining the free [InsP₃] produced in the region of the photolysis spot, we modelled the dynamics of the process. This model combined the kinetics of the uncaging process with diffusion of both caged and free InsP₃ in 2-D, using software previously described (Zou et al. 1999). Computer simulation showed that with diffusion coefficients for caged and free InsP₃ set in the range 50-250 μm² s⁻¹, while the total amount of InsP₃ liberated in the dynamic model was up to 5 times greater, the [InsP₃] generated locally within the photolysis region was similar to that calculated from the simpler, static model (data not shown; the breakdown of caged InsP₃ was modelled as an exponential process with τ_½= 2 ms, Dr J. Walker personal communication).

We have only considered the variation in [InsP₃] resulting from positioning of the photolysis spot within the plane of focus. Due to the much lower axial resolution of the microscope, the intensity of the spot, as well as the spatial profile, would be expected to vary much more slowly with changes in focus. The analysis of the spot profile could be extended to 3-D. The general nature of the relationship between the laser power setting and the resulting average [InsP₃] evoking Ca²⁺ release would be similar, but probably shifted to lower estimates of [InsP₃]. While modelling in 3-D would be valuable, in the light of the uncertainty in some physical parameters of the behaviour of the caged InsP₃ and the complexity in our optical system we feel a 2-D analysis provides at least a minimal model for understanding the processes. Throughout the paper we have expressed the stimulus intensity as the power setting on the laser (in absolute units (a.u.)). Based on our modelling and on previous characterizations of the caged InsP₃ (Walker et al. 1989), we have estimated that 1 a.u. or greater is equivalent to a rapid elevation in [InsP₃] to an average of 10-50 μm at the focal plane.

Even though there may be errors in the estimates of [InsP₃], due to the limitations of our models, we would expect these errors to be consistent in the different regions of the cell. As such, comparisons between responses in different regions of the cell evoked by the same intensity UV photolysis spots are valid.

RESULTS

Intracellular Ca²⁺ changes were imaged with high temporal resolution using a high-speed, digital-imaging microscope (up to 500 Hz full-frame image capture) in cells microinjected with 70 000 MW dextran-coupled Calcium Green (see Methods). The amount of InsP₃ liberated was controlled by regulating the intensity of the UV flash. An example of a global Ca²⁺ response is shown in Fig. 2, and as previously reported for agonists, the initial response was observed in the secretory pole (defined by the location of zymogen granules) followed by a delayed response in the basal pole (Fig. 2). We determined that this global response was not affected by the addition of 100 μm ryanodine (n = 3, data not shown) and therefore conclude that, consistent with previous reports, ryanodine receptors play no role in InsP₃-mediated responses (Thorn et al. 1994).

Despite the delay in the response, the rapid rise of the Ca²⁺ signal in the basal pole suggests an active mechanism of Ca²⁺ release, rather than passive diffusion of Ca²⁺ from the secretory pole. However, in these relatively small cells it is difficult to isolate regional Ca²⁺ signals during a global response. We therefore aimed to try to study the local mechanisms of Ca²⁺ release by spatially limiting the stimulus. To this end InsP₃ was locally liberated by photolytic breakdown of a caged precursor using a UV laser beam focused down to a narrow spot within the cell (2 μm diameter at the image plane, FWHM). To determine the spatial characteristics of our spot stimulus we carried out control experiments uncaging fluorescein (Fig. 3A). At up to 200 ms after photolytic release the region showing elevated fluorescence was spatially restricted (Fig. 3A, right-hand image). A further control was carried out to test for the repeatability of the InsP₃-evoked Ca²⁺ responses and the possibility of laser damage. Even at the highest stimulus intensity used (6 a.u.) repeated UV spot photolysis yielded very similar Ca²⁺ responses (Fig. 3B). In all our experiments with caged InsP₃ the Ca²⁺ signal was recorded from the whole cell but we restricted our analysis to the average response within a 4 μm diameter region centred on the photolysis spot region.

At UV laser intensities of 1 a.u. or above, spot release of InsP₃ within the secretory pole nearly always elicited a Ca²⁺ response initiated within the photolysis region (Fig. 4, response observed in 12/13 cells), with a dose-dependent increase in the magnitude (to a maximum ΔF/F_o plateau of 33.55 ± 0.89 %, mean ±s.e.m., n = 5) and the rate of rise of the Ca²⁺ response (to a maximum of 0.168 ± 0.037 % ms⁻¹, mean ±s.e.m., n = 8) and a decrease in latency (to a minimum of 79 ± 9.3 ms, mean ±s.e.m., n = 5) (Fig. 5A and B). It is clear in Fig. 5A that the response saturates at the highest stimulus intensities. The maximum amplitude and rate of rise were comparable to those seen in global InsP₃ photolysis experiments (such as in Fig. 2) and in response to the injection of Ins(2,4,5)P₃ (Kidd et al. 1999). This suggests that it is the Ca²⁺ release mechanism that is saturated, rather than any local depletion of InsP₃ that might occur at higher flash intensities.

Figure 4. The Ca²⁺ response to spot InsP₃ liberation in the secretory pole.

The apex of the secretory pole is shown by an asterisk and region of the UV photolysis spot (FWHM of the photolysis spot) is shown by the white circle in the first ratio image. After InsP₃ liberation (time point shown by the arrow on the response graph, 1 a.u. laser power) in the secretory pole, the local Ca²⁺ response, measured in the region of photolysis (4 μm diameter), developed rapidly and with a short latency. The series of whole-cell ΔF/F_o images, collected at 11 ms sampling intervals, shows the origin of the signal beneath the photolysis spot that then spreads as a wave across the cell. The time at which the images were collected is also shown on the graph. In this figure, and in Fig. 6A and B, image 2 was selected as the time point when the ΔF/F_o signal reached a 5 % increase over baseline.

The Ca²⁺ release induced by spot release of InsP₃ in the basal pole region was much more variable (Fig. 6A and B). In about half of the cells (5/12) little active response was observed, even at the highest stimulus intensities used. These responses had long latencies and the Ca²⁺ signal, under the region of the photolysis spot, was sometimes contaminated with a Ca²⁺ signal, back spreading, from a delayed response in the secretory pole (Fig. 6A). In the other cells, however, spot release of InsP₃ in the basal pole did induce a clear response, initiated in the photolysis region, of a magnitude, rate of rise and latency similar to that found in the secretory pole (Fig. 6B). Separate Ca²⁺ traces are shown in Fig. 6C and D for the two cells shown in Fig. 6A and B, respectively. It is clear that the poorly responding cell (Fig. 6A and C) has slow responses with a rise time that shows little dependence on laser power (Fig. 6E, squares). In contrast, the cell in Fig. 6B and D shows a fast component that develops at higher stimulus intensities (Fig. 6E, circles).

Figure 6. Examples of the Ca²⁺ response, in two different cells, to the spot liberation of InsP₃ in the basal pole region.

InsP₃ release in the basal pole region elicited either a weak response (A, 1 a.u. laser power) or a response indistinguishable from that induced in the secretory pole (B, 1 a.u. laser power). C, in a single cell (the same as that shown in A), repeated UV photolysis in the basal pole elicited poor Ca²⁺ responses that showed little ‘dose’ dependence in the rate of rise and long latencies to a ΔF/F₀ of 5 %. The order of presentation of the stimuli was random; in this particular case it was (laser power, a.u.): 1, 0.5, 0.25, 2, 4, 6. D, in a single cell (the same as that shown in B), repeated UV photolysis in the basal pole elicited a ‘dose’-dependent effect on the rate of rise of the Ca²⁺ signal and on the latency. Again the order of presentation of the stimuli was random; in this particular case it was (laser power, a.u.): 1, 0.25, 0.12, 0.05, 0.5. E, graphs show the latency and maximum rates of rise for each of the cells (▪, cell in A and C; •, cell in B and D). (Note, a latency of longer than 1760 ms (at 0.125 a.u. for the cell in A and C) is indicated on the graphs as 1760 ms. Also rates of rise of less than 0.02 % ms⁻¹ were difficult to measure for the cell in A and C and were not included on the graph.)

In the poorly responding cells, moving the photolysis spot to other regions did elicit a rapid response, indicating that these cells were capable of responding. Figure 7 shows an example of this, with data obtained from a single cell giving markedly different Ca²⁺ responses to the same InsP₃ liberation, in different regions of the basal pole.

Figure 7. The Ca²⁺ response to regional spot photolysis of InsP₃ in four places within a single cell.

The response of a single cell to repeated InsP₃ liberation in different regions. The left-hand arrow shows the time point of flash photolytic release of InsP₃. The pseudocolour images were taken at the time point indicated by the arrow marked ‘images’. In the secretory pole (region 1) a vigorous response was observed. Regions 2, 3 and 4 in the basal pole show variability in the responses to the same stimulus. To position the photolysis spot in some regions of the basal pole, parts of the cell sometimes fell outside the camera field of view (images 2 and 4).

Figure 8 compares the relationship between the InsP₃ dose and the rate of Ca²⁺ rise of active responses in the basal and secretory pole. The response is shown plotted against absolute units (Fig. 8A) and also plotted against the relative [InsP₃] calculated from our models (Fig. 8B). The basal pole and secretory pole responses rose to the same maximum (not significantly different, Student's t test) indicating that the flux, within the analysed spot, is similar. However the basal pole responses were shifted to the right (the data at the mid-point of the curve were significantly different, Student's t test, P < 0.05) indicating a lower functional affinity for InsP₃.

Figure 8. Comparison of the InsP₃ dose relationship to the rate of Ca²⁺ rise.

The responses are plotted either against laser power expressed in absolute units (A) or against the relative [InsP₃] calculated from our model (B). The mean secretory pole responses (n = 12 cells, •, error bars are s.e.m.) and the mean of the active basal pole responses (n = 4, ▪, error bars are s.e.m.) are shown. The points indicated by the asterisks were significantly different at P < 0.05 (Student's one-tailed t test), all other points were not. The units of the InsP₃ dose are relative [InsP₃], estimated as described in Methods and illustrated in Fig. 1. From the injected concentrations of caged InsP₃ we would estimate the maximum [InsP₃] to be 10-50 μm.

DISCUSSION

These results provide a functional description of regional InsP₃-evoked Ca²⁺ responses in single epithelial cells. In particular we have addressed the basis of the global Ca²⁺ response in acinar cells. The spot photolytic liberation of InsP₃ showed that some regions of the basal pole were incapable of rapid InsP₃-evoked Ca²⁺ release even at the highest stimulus levels we used. We conclude that these regions either do not contain InsP₃ receptors or have receptors with very different characteristics. However, other sites in the basal pole did show InsP₃-evoked responses, which had similar kinetics of Ca²⁺ release, but with a lower apparent InsP₃ affinity, compared to the secretory pole responses.

Secretory pole Ca²⁺ responses

The local secretory pole spikes and the initiation of global Ca²⁺ waves at the secretory pole have long been postulated to be due to a greater InsP₃‘sensitivity’ in this region (Kasai & Augustine, 1990; Thorn et al. 1993). Experiments on permeabilized cells have shown that it is a property of Ca²⁺ release that gives rise to the polarized Ca²⁺ signal in acinar cells (van de Put & Elliott, 1996; Tanimura et al. 1998). However, attempts to functionally identify the basis of the secretory pole Ca²⁺ release characteristics, compared to that of the basal pole, have not shown any regional differences in response (Toescu et al. 1994; van de Put & Elliott, 1996). Recently, Ito et al. (1999) have measured a difference in the functional affinity for InsP₃ between the secretory pole (estimated EC₅₀≡ 5 μm) and the basal pole (EC₅₀≡ 50 μm) based on global flash photolysis experiments. These experiments provide the first evidence that there are regional differences in InsP₃ sensitivity. However, interpretation of these experiments is compromised due to the potential for contamination of the basal pole Ca²⁺ signal with secretory pole Ca²⁺ responses and also the lack of spatial resolution in the basal pole response.

To overcome these problems our experimental approach used microinjection of a minimum amount of 70 000 MW dextran-coupled Calcium Green. This technique, unlike whole-cell patch clamp, retains native intracellular Ca²⁺ buffers and minimizes disturbances in intracellular Ca²⁺ buffering due to exogenous dyes. We also designed a system capable of delivering a localized InsP₃ stimulus and recording the localized regional Ca²⁺ responses. In this way we were able to determine the regional InsP₃ responses, in isolation, in the absence of the signal contamination that is inevitable during the global stimulation. Our local stimulus and local recording system had the spatial resolution to determine spatially discrete responses and subdivide the basal pole region. Finally, the high temporal resolution of our recordings allowed full characterization of the extremely rapid responses elicited at the higher stimulus intensities.

With the above methods we can now conclude that the secretory pole is a region of higher apparent InsP₃ affinity, although we estimate less than a 5-fold difference (less than a 2-fold difference based on our calculations of relative [InsP₃], Fig. 8B), compared to the 10-fold difference seen by Ito et al. (1999). This discrepancy between our work and that of Ito et al. (1999) is likely to be explained by our greater spatial resolution, which has enabled us to identify discrete regions within the basal pole that are essentially unresponsive to InsP₃.

Global Ca²⁺ responses

Our observation on the isolated basal pole responses provides a functional explanation for the global Ca²⁺ responses seen at higher intensities of stimulation. The apparent difference in InsP₃ affinity between the secretory pole and the active sites of release in the basal pole indicates that the basal pole would have a higher threshold for activation. Also, although we consistently observed responses to local InsP₃ stimuli in the secretory pole we failed to see rapid responses in some discrete regions within the basal pole. Therefore, under conditions where InsP₃ is globally elevated we would expect the release sites in the secretory pole to become more excitable and therefore initially more responsive than those in the basal pole. The basal pole responses would only be recruited at higher levels of InsP₃. In this way two parameters are acting to reduce the responsiveness of the basal pole region: firstly, the functional characteristics of the InsP₃ receptors, which show a lower apparent affinity, and secondly, the more widespread separation of release sites in the basal pole, which would limit their recruitment to higher stimulus intensities.

Our results indicate that the basal pole has heterogeneous InsP₃-evoked Ca²⁺ responses. This region of the acinar cells is rich in endoplasmic reticulum and is also the region containing the nucleus. It has previously been shown that the nuclear envelope is capable of InsP₃-dependent responses (Gerasimenko et al. 1995) and InsP₃ receptors have been shown to be clustered on the envelope membrane (Mak & Foskett, 1997), potentially giving rise to the observed local Ca²⁺ responses (Lipp et al. 1997). It is possible that the regions of active responses we identified arise from Ca²⁺ release from the nuclear envelope rather than the endoplasmic reticulum.

Role of InsP₃ receptor isoforms

This greater apparent InsP₃ affinity in the secretory pole may result from a greater concentration of higher affinity InsP₃ receptors. Pancreatic acinar cells have been shown to predominantly contain type 2 and type 3 InsP₃ receptors (De Smedt et al. 1994; Wojcikiewicz, 1995) and binding studies have shown that type 2 InsP₃ receptors have a higher InsP₃ affinity than type 3 (Newton et al. 1994; Parys et al. 1995). The type 2 receptor would therefore be predicted to show a secretory pole localization. Consistent with this, immunolocalization studies have shown that the type 2 InsP₃ receptor is exclusively located in the secretory pole (Lee et al. 1997; Yule et al. 1997) whereas there is some evidence for the type 3 (Lee et al. 1997; Yule et al. 1997) and type 1 receptors (Yule et al. 1997) in the basal pole regions. This apical clustering of type 2 receptors at the site of initiation of the Ca²⁺ signal has also been seen at Ca²⁺ response initiation sites in other cells, such as the submandibular duct cell (Yamamoto-Hino et al. 1998) and astrocytes (Sheppard et al. 1997). The clustering of InsP₃ receptors, immunolocalized to the secretory pole, does agree with our data of consistent InsP₃-evoked Ca²⁺ responses in this region. However, our data are not consistent with the apparent low density of receptors immunolocalized in the basal pole (Lee et al. 1997; Yule et al. 1997). In our experiments we observed that the maximum rate of rise of the Ca²⁺ response achieved in active regions of the basal pole is the same as in the secretory pole. This suggests regional clustering of InsP₃ receptors in the basal pole at a similar density to that in the secretory pole. More detailed studies of the cellular localization of InsP₃ receptors will be required to resolve this issue, which might arise out of the limitations in immunolocalization techniques or the presence of a novel InsP₃ receptor isoform (Lee et al. 1997).

Of course, the affinity for InsP₃ is only one aspect of InsP₃ receptor behaviour and it has been suggested that a decreased sensitivity to Ca²⁺-dependent inactivation of the type 3 InsP₃ receptor makes this a likely candidate for initiation of the Ca²⁺ signal in the secretory pole (Hagar et al. 1998). Very similar arguments have been made for the type 2 receptor (Ramos-Franco et al. 1998). Further experiments will be required to determine these isoform-specific characteristics (Taylor, 1998). An additional complication is that heteroligomeric InsP₃ receptors may be found in the secretory pole and their behaviour would be dependent on isoform composition (Miyakawa et al. 1999).

Agonist-evoked Ca²⁺ signals

In pancreatic acinar cells it is clear that the agonist-evoked Ca²⁺ pool is critically dependent on the regional properties of InsP₃-evoked Ca²⁺ release that we have described above. Agonist responses are blocked by heparin (Thorn et al. 1994) and agents that act on the ryanodine receptor only modulate, but do not block, the global responses (Nathanson et al. 1992; Thorn et al. 1994). However, in trying to build up a complete picture of agonist-evoked Ca²⁺ signalling we have to include the potential for InsP₃-independent pathways. These include cyclic adenosine 5′-diphosphate-ribose (cyclic ADPribose)-evoked Ca²⁺ release (Thorn et al. 1994) and, more recently, nicotinic acid adenine dinucleotide phosphate (NAADP)-evoked responses (Cancela et al. 1999). These other pathways may play a role in agonist-specific aspects of the Ca²⁺ signal.

Conclusion

In conclusion we have described regional differences in InsP₃-evoked responses in single pancreatic acinar cells. Our results provide a functional explanation for the observation of global Ca²⁺ signals which initiate in the secretory pole and then spread across the basal pole region.