Apical Vesicles Bearing Inositol 1,4,5-trisphosphate Receptors in the Ca²⁺Initiation Site of Ductal Epithelium of Submandibular Gland

mAbs against IP₃R1, 2, and 3 were designated KM1112 (IgG₁), KM1083 (IgG_2a), and KM1082 (IgG₁), respectively. They were raised against synthetic peptides corresponding to the COOH-terminal regions of the human IP₃Rs, the sequences of which are specific to each type. The specificities of these mAbs were characterized previously (Sugiyama et al., 1994). A polyclonal antibody (pAb) against an 11–amino acid peptide (αpep6) of the COOH terminus of mouse IP₃R1 was also used (Nakade et al., 1994). The mAb against IP₃R3 (IgG_2a clone 2) was raised against amino acids 22– 230 of the human IP₃R3 (Transduction Laboratories, Lexington, KY). The secondary antibody was an FITC-conjugated, anti–mouse IgG pAb that was adsorbed against rat serum (Vector Laboratories, Burlingame, CA). Subclass-specific secondary antibodies (FITC-conjugated goat anti– mouse IgG₁ pAb and Texas red–conjugated goat anti–mouse IgG_2a pAb) were purchased from Southern Biotechnology Associates, Inc. (Birmingham, AL).

After being anesthetized with Nembutal, 8-wk-old rats were perfused with PBS followed by 4% paraformaldehyde in PBS. SMGs were dissected and postfixed for 2 h at 4°C. Cryosections (8 μm) were permeabilized with 0.05% Triton X-100 in PBS and treated with PBS containing 2% normal horse serum for 30 min at room temperature. The sections were then incubated with antibodies against IP₃R2 (2 μg/ml) and IP₃R3 (10 μg/ml). After washing, the sections were incubated with FITC-conjugated anti– mouse IgG. For double immunofluorescence staining of IP₃R2 and IP₃R3, we used Texas red–conjugated goat anti–mouse IgG_2a and FITC-conjugated goat anti–mouse IgG₁ secondary antibodies for KM1083 and KM1082, respectively. As a control, the antibody solutions were preadsorbed with a 10-fold excess of the peptide used to generate the primary mAb.

Frozen sections (12 μm) were prepared by cryostat sectioning, thaw-mounted onto gelatin-coated glass slides, and fixed for 15 min in 4% paraformaldehyde. By reverse transcriptase PCR, we generated two DNA probes, corresponding to the nucleotide position of 2920–4404 and 4609– 5996 of the rat IP₃R2 cDNA sequence (Sudhof et al., 1991). Each fragment was cloned into the PCRII vector (Invitrogen Corp., Carlsbad, CA). Antisense and sense RNA probes were prepared by standard procedure using these two plasmid DNAs as templates. In situ hybridization was performed essentially as described by Furuichi and colleagues (1993).

SMGs of 8-wk-old male rats were removed and homogenized in a buffer containing 0.25 M sucrose, 1 mM EDTA, 0.1 mM PMSF, 10 μM leupeptin, 10 μM pepstatin A, 1 mM 2-mercaptoethanol, and 50 mM Tris-HCl, pH 7.4. The homogenates were centrifuged at 1,000 g for 5 min at 4°C. The supernatants were centrifuged at 105,000 g for 60 min at 2°C to sediment the membrane proteins. The pellets were resuspended in 8 M urea, 1 mM EDTA, 0.1 mM PMSF, 10 μM leupeptin, 10 μM pepstatin A, 1 μM 2-mercaptoethanol, and 50 mM Tris-HCl, pH 7.4. The membrane fractions dissolved in the sample buffer (4% sodium dodecyl sulfate, 12% glycerol [wt/ vol], 50 mM Tris, 2% 2-mercaptoethanol [vol/vol], 0.01% Serva blue G, pH 6.8) were subjected to 5% Tricine–SDS-PAGE (Schagger and Jagow, 1987) containing 6 M urea, followed by electroblotting onto nitrocellulose membranes. The membranes were blocked with 5% skim milk and incubated with 40 ng/ml of each of the IP₃R mAbs. The bound mAbs were detected with the ECL system (Amersham Corp., Arlington Heights, IL). As a control, the antibody solutions were preadsorbed with a 10-fold excess of the peptide used to generate the primary mAb.

We have developed a new technique that improves on previous preembedding techniques and has resulted in better images with good preservation (Katsumata et al., 1996). Briefly, the cryosections on glass slides were incubated first in a blocking solution, then with primary antibody, KM1083, overnight, and finally with anti–mouse IgG antibody conjugated with 0.8-nm gold colloids. The sections were rinsed with three changes of PBS at each step. The sections were postfixed with 2.5% glutaraldehyde, 0.2% tannic acid, and 1% OsO₄ in 0.1 M phosphate buffer for 30 min at room temperature, followed by a triple-distilled water rinse. Next, they were soaked in a solution for silver enhancement followed by three changes of distilled water. They were then embedded in a 1% chitoisan solution, dehydrated, and then processed routinely. Finally, the thin sections were stained with uranyl acetate and lead citrate and examined with an electron microscope (model JEX-1200EX; JEOL Ltd., Tokyo, Japan).

Wistar rats (8 wk old) were killed under deep anesthesia. The SMGs were excised immediately, trimmed of the capsule connective tissue, and cut into small pieces. They were digested with collagenase (2 mg/ml) in Eagle's minimal essential medium containing 25 mM Hepes, pH 7.4, for 45 min at 37°C with constant shaking and gentle pipetting (Segawa et al., 1985). The dissociated SMGs were then resuspended in a physiological salt solution (PSS) containing 135 mM NaCl, 5.8 mM KCl, 1.8 mM CaCl₂, 0.8 mM MgSO₄, 0.73 mM NaH₂PO₄, 11 mM glucose, 20 mM Hepes, pH 7.4, and 2 mM glutamine. The resuspended cells were incubated in the dark at 37°C for 30 min in the presence of 5 μM Indo-1-AM (Molecular Probes, Eugene, OR) or 10 μM Fluo-3-AM (Molecular Probes). After washing with PSS, they were mounted onto poly-l-lysine–coated cover glasses by Cytospin (Shandon, Cheshire, UK) centrifugation at 200 rpm and placed in a thermostatically controlled chamber at 37°C. Fluorescence was detected with a confocal laser scanning microscope (model RCM8000; Nikon Corp., Tokyo, Japan) with a 40× water immersion objective. Ducts and acini on the chamber were stimulated by perfusion with carbachol, dissolved in PSS. Ca²⁺-free solution was prepared by replacing 1.8 mM CaCl₂ by 1 mM EGTA.

We investigated the localization of three distinct types of IP₃Rs throughout rat tissues with type-specific mAbs. Immunofluorescence using anti-IP₃R2 mAb (KM1083) showed that IP₃R2 was localized predominantly over the ducts in the rat SMG (Fig. 1,A). These ducts are composed of columnar cells lining the lumen. All staining was completely blocked by preincubation of the antibody with the antigen peptide (Fig. 1,B). At a higher magnification (Fig. 1,C), the strong immunosignals were concentrated in the apical cytoplasm. However, diffuse staining was also detected throughout the cytoplasm in the ductal cells. Heterogeneity was noted in the intensity of immunostaining among the duct cells. Extraordinarily strong immunostaining was found in certain ductal cells that were scattered throughout the granulated, striated, and exocretory ducts in intralobules. It was most remarkable in interlobular ducts (Figs. 1,A and 2,A). Weaker immunostaining with KM1083 was observed as well in apico-lateral regions of the acinar cells, corresponding to the intercellular canaliculi, the specialized lumen structure in the acini of rat SMG (Fig. 1, A and C).

Our in situ hybridization study confirmed that IP₃R2 expression in the SMG ducts was highest among all the tissues we analyzed. Silver grains representing the expression of IP₃R2 mRNA were concentrated in cells of the duct system (Fig. 1 D). The other components of the gland, including the secretory acini and the connective tissues, showed labeling at a level comparable to those on the control sections hybridized with a sense (control) probe (data not shown).

The localization of two other IP₃R types, IP₃R1 and IP₃R3, in the rat SMG duct were also analyzed. The Abs against IP₃R1, KM1112 mAb and αpep6 pAb, did not give significant immunostaining in the duct system. We used two mAbs for detection of IP₃R3: KM1082 and IgG_2a clone 2, which recognized different sequences of IP₃R3 (see the Materials and Methods). Both anti-IP₃R3 mAbs gave weak and scattered immunosignals in apical regions of interlobular ductal epithelium and apico-lateral regions of the acinar cells (data not shown). Double immunostaining with KM1083 and KM1082 in the same section was also performed. Fig. 2, A and B, shows double-stained interlobular ducts (arrows), indicating that IP₃R3 coexists with IP₃R2 in the apical region. IP₃R3 was not detected in intralobular ducts (arrowheads). Because the immunosignal of KM1082 was relatively weak, a high concentration of KM1082 and a long photographic exposure were required to clearly localize IP₃R3, resulting in relatively high background. Staining from KM1082 outside of the apical region of the duct was also obtained in control sections (data not shown), so it was most likely nonspecific staining.

We also analyzed the expression of IP₃Rs in rat SMG by Western blotting using mAbs KM1112 for IP₃R1, KM1083 for IP₃R2, and KM1082 for IP₃R3. A band of ∼240 kD was detected using each mAb (Fig. 3). As shown previously (Monkawa et al., 1995), the lower molecular–mass bands in the lanes of three IP₃Rs are proteolytic fragments of IP₃Rs. In our hands, these bands increased in intensity, and the bands of IP₃Rs were decreased by incubating the samples at 37°C for 6 h, suggesting proteolysis (data not shown). The 240-kD bands were abolished by preadsorption of the antibody with the antigen peptide, indicating that these antibodies specifically recognize each type of IP₃R protein expressed in the rat SMG.

The ultrastructural localization of IP₃R2 in the duct cells was then examined by immunoelectron microscopy using the modified, preembedding silver enhancement method. As shown in Fig. 4, the silver grains were distributed mostly around small vesicles located in the apical cytoplasm of certain, but not all, duct cells. In this micrograph, relatively thin cytoplasm and short microvilli identified this as a “dark” cell (Sato and Miyoshi, 1988). Some principal cells in the striated and excretory ducts also possessed apical vesicles positive for IP₃R2. In the apical vesicles, silver grains were found mostly on the cytoplasmic side of the vesicle membrane, consistent with the fact that the epitope of the antibody is predicted to be on the cytoplasmic face. Other organelles, such as the rough ER, nuclei, and mitochondria, were mostly devoid of grains. No labeling was detected on the plasma membrane. These results suggested that among the duct cells, there was a specific population of cells that possess these apical vesicles bearing IP₃R2.

Next we investigated the spatial and temporal distribution of Ca²⁺ in the ducts during stimulation by the phosphoinositide-coupled agonist, carbachol. We treated the rat SMG with collagenase to obtain dispersed pieces of ducts, which were then loaded with a Ca²⁺ indicator Indo-1. The lumen of the dissociated ducts was visualized by confocal imaging (Fig. 5,A). In the presence of extracellular Ca²⁺, the ducts were continuously exposed to 10 μM carbachol. Fig. 5,A (a typical result of 20 such experiments) illustrates the spatial development of the cytoplasmic Ca²⁺ signal within the ducts. At rest (time = 0; Fig. 5,A, a), [Ca²⁺]_i was relatively low, although some localized moderate increases in [Ca²⁺]_i were detected in the basolateral periphery. The first increase of [Ca²⁺]_i was observed in the cytoplasm close to the apical plasma membrane of the cell (indicated by an open square in Fig. 5,A, a) with a time lag of 1 s after stimulation (Fig. 5,A, c). This rise in [Ca²⁺]_i then spread, in a gradient fashion, toward the basolateral area. Fig. 5 B shows a series of digital images (every 67 ms) showing the progression of the carbachol-evoked [Ca²⁺]_i waves in the cell. Within about 1 s after the initial apical rise in [Ca²⁺]_i, a complete elevation of [Ca²⁺]_i throughout the whole cytoplasm was observed. The front of the [Ca²⁺]_i transients moved at about 20 μm/s.

3–4 s after the beginning of the carbachol stimulation (Fig. 5,A, h–k), Ca²⁺ transients occurred also in the apical area of certain other cells. The [Ca²⁺]_i signals propagated toward the basolateral portions with a pattern similar to that seen in Fig. 5,B. When the carbachol-evoked calcium responses were fully developed (Fig. 5 A, o–s), [Ca²⁺]_i levels appeared to be homogeneous across the cytoplasm of these cells. But even at this stage, the [Ca²⁺]_i proved to change dynamically, as shown by the following detailed experiment.

In an attempt to analyze the rapid Ca²⁺ response with greater precision, we used the nonratiometric Ca²⁺ indicator, Fluo-3, and observed [Ca²⁺]_i by video-rate confocal microscopy. We were able to see in greater detail [Ca²⁺]_i fluctuations induced by 10 μM carbachol in the presence of extracellular Ca²⁺. Fig. 6,B gives a representative [Ca²⁺]_i transient observed in a cell; the time courses of [Ca²⁺]_i measured at three points across the cytoplasm of the cell (Fig. 6,A) are shown. In the apical Ca²⁺ initiation site, a short-lasting Ca²⁺ spike occurred, rapidly peaking at 250 ms, and decaying over 400–600 ms, with a frequency of around 0.5 Hz. Subsequently, the oscillatory [Ca²⁺]_i rises moved toward the basolateral area. The peaks of the [Ca²⁺]_i transients at the apical point (square) preceded the peak at the middle point (triangle) by 0.19 ± 0.07 s. The distance between the two points was 4.7 μm. Thus, the velocity of the [Ca²⁺]_i wave in the cell was calculated to be 26 ± 7 μm/s (n = 7), which is similar to that characterized in Fig. 5 B. In the basolateral periphery (circle), small and continuous [Ca²⁺]_i transients were observed but were not statistically different from background noise (fluorescence intensity before stimulation, 66.5 ± 1.6 arbitrary units; after stimulation, 77.2 ± 2.8 arbitrary units, n = 12). These results indicated that the Ca²⁺ wave was confined to the apical-middle region.

To determine if extracellular Ca²⁺ participated in the [Ca²⁺]_i transients, measurements were performed without extracellular Ca²⁺. Fig. 6,C shows the time course of [Ca²⁺]_i at three points as indicated in Fig. 6 A. The Ca²⁺ wave moved from the apical to the basolateral region of the cell with the same velocity, duration, and frequency as seen with extracellular Ca²⁺, but with decreased amplitude. Thus, extracellular Ca²⁺ is not likely to be important for triggering the Ca²⁺ waves and oscillations. In the absence of extracellular Ca²⁺, the Ca²⁺ waves disappeared ∼50 s after the beginning of stimulation. However, the waves continued during stimulation when Ca²⁺ was present (data not shown).

In response to carbachol, Ca²⁺ spikes and subsequent Ca²⁺ waves took place only in certain ductal epithelial cells. Cells that were highly responsive to carbachol and those that were not were positioned randomly along the lumen (Fig. 5,A, o–s). We observed very heterogeneous [Ca²⁺]_i transients in 20 independent ductal preparations. We carried out immunohistochemistry on the ducts that had been used for the Ca²⁺ imaging experiments using KM1083 and an anti–mouse IgG secondary antibody conjugated to FITC. Fig. 7,A shows a [Ca²⁺]_i transient induced by carbachol in the apical cytoplasm in a cell. Likewise, in this same cell, a strong immunopositive signal of KM1083 colocalized with the rise in apical [Ca²⁺]_i (Fig. 7 B). In five out of six cells from three experiments, the Ca²⁺ initiation site also possessed significant amounts of IP₃R2.

We investigated the intracellular and intercellular localization of IP₃Rs and correlated this with the initiation site of Ca²⁺ transient induced by agonist stimulation in ductal cell of the rat SMG. The Ca²⁺ signaling initiated primarily in the apical region of “pioneer cells,” which expressed relatively high levels of IP₃Rs and evoked rapid and wavelike [Ca²⁺]_i transients after carbachol stimulation.

We showed that the distributions of IP₃R2 and IP₃R3 were predominantly in the apical poles of rat SMG duct cells (Figs. 1 and 2). The abundance of IP₃R2 mRNA was also shown by in situ hybridization (Fig. 1); among all the tissues tested, the SMG ducts gave highest hybridization signals. IP₃R3 was immunolocalized using two mAbs recognizing different epitopes (Fig. 2 B). Recently, Lee et al. (1997a) reported the localization of IP₃Rs in rat SMG ducts; IP₃R1 and IP₃R2 were localized close to the luminal and lateral membranes, and IP₃R3 around the nuclear envelope. These results differed from ours in the localizations of IP₃R1 and IP₃R3: (a) We did not detect any significant positive signals for IP₃R1 in the duct cells by either immunofluorescence or in situ hybridization (data not shown). The anti-IP₃R1 mAb, KM1112, has been shown to properly cross-react with rat IP₃R1 in tracheal and bronchiolar epithelia (Sugiyama et al., 1996), even though the peptide sequence used as an antigen to produce KM1112 was based on the human IP₃R1. Also, the 11 COOH-terminal amino acid sequence is common between rat and human IP₃R1s (Mignery et al., 1990; Yamada et al., 1994). So the fact that we did not detect IP₃R1 using KM1112 was not likely due to insufficient affinity of KM1112 to IP₃R1 in rat tissue. (b) In our study, IP₃R3 was not localized to the nuclear envelope but was colocalized with IP₃R2 on the apical side of the cell.

Previous electron microscopic studies showed that the ductal epithelium of rat SMG consists of several cell types, i.e., light, dark, tuft, and basal cells, most of which contain apical vesicles whose function remains unclear (Sato and Miyoshi, 1988). In the present study, using preembedding, silver-enhanced immunoelectron microscopy, we identified vesicular structures bearing IP₃R2 in the apical pole of some SMG duct cells (Fig. 4), which suggested that the apical vesicles might function as an intracellular Ca²⁺ pool. These vesicles were quite small (average size 20 nm in diameter) and did not appear similar to elements of the typical ER in Purkinje cell, which are IP₃-sensitive Ca²⁺ pools (Mignery et al., 1989). The characteristic shape of these apical vesicles resembles that of the “calciosome,” which is believed to be an IP₃-regulated Ca²⁺ pool as differentiated portions of the ER in nonmuscle cells (Volpe et al., 1988; Pozzan et al., 1994). However, calciosomes were distributed throughout the cytoplasm in close apposition to the ER, which is quite different from the apical vesicles we have described in this study. It is interesting whether the apical vesicles we have described are also specialized ER components or something distinct from ER.

We found that only certain ductal cells expressed high levels of IP₃Rs in the apical region of the cell (Figs. 1 and 2). By three-dimensional, scanning confocal microscopy, we confirmed that the heterogeneity was clear at each level through the z-axis (data not shown). Our immunoelectron microscopy showed that certain “dark” and “principal” cells also had a large amount of IP₃R2 (Fig. 4; Segawa et al., 1996), so it is possible that any type of cell that expresses large amount of IP₃Rs may function as a pioneer cell (see below discussion).

Recently, upregulation of IP₃R expression during apoptosis has been reported (Khan et al., 1996), and it may be argued that the SMG duct cells exhibiting high IP₃R immunoreactivity are apoptotic. However, this seems unlikely since in our immunoelectron micrographs, we did not observe condensation of nuclear chromatin or any other apoptotic morphology (Kerr et al., 1972). Furthermore, we found hardly any DNA fragmentation in SMG duct cells (data not shown) by the TUNNEL method (Gavrieli et al., 1992).

The ryanodine receptor and ER Ca²⁺ pump are also expressed in the luminal pole of the SMG duct (Lee et al., 1997a,b). It would be interesting to know whether these molecules are also expressed heterogeneously among the ductal epithelial cells.

Using real-time, confocal microscopy, we succeeded in visualizing [Ca²⁺]_i changes with high temporal and spatial resolution. The changes of [Ca²⁺]_i caused by carbachol were very rapid and dynamic. Our Ca²⁺ imaging demonstrated that Ca²⁺ spikes, local waves, and oscillations originated in the apical region of the duct cells (Figs. 5 and 6). Spatially restricted repetitive Ca²⁺ spikes described as “puffs” were also reported in mouse pancreatic acinar cells (Thorn et al., 1996) and in Xenopus oocytes (Yao et al., 1995; Berridge, 1997). When one compares Ca²⁺ spikes in rat SMG ducts with those in acinar cells and oocytes, the short-lasting Ca²⁺ spike in the SMG duct resembles the Ca²⁺ puffs in oocytes with respect to their short duration. However, the frequency of oscillations in SMG ducts was much faster than that in acinar cells or oocytes. Ca²⁺ puffs and spikes in pancreatic acinar cells are thought to be due to the opening of small clusters of IP₃Rs, but the relationship between the initiation site of Ca²⁺ spikes and the distribution of IP₃Rs has not been studied intensively in these cells. Here, we demonstrated the high-level expression of IP₃R2 and IP₃R3 at the Ca²⁺ wave initiation site (Fig. 7). Recently, Lee et al. (1997a) showed a correlation between the pattern of IP₃R expression and the initiation sites of Ca²⁺ waves in pancreatic and submandibular acinar cells. Our results, together with the data from Lee et al. (1997a), suggest that the apical concentration of IP₃Rs might be a general phenomenon in many other polarized epithelial cells.

While IP₃ signals are generated from the basolateral plasma membrane, the IP₃-induced increase in [Ca²⁺]_i occurs at the apical pole in these polarized cells, and unfortunately, the functional significance of the differential subcellular localizations of these two second messengers (between which cross talk is known to occur) is still unclear. Also, it is noteworthy that there is a considerable difference in the distance that Ca²⁺ and IP₃ can diffuse. In Xenopus oocytes, the diffusion lengths of IP₃, buffered Ca²⁺, and free Ca²⁺ were measured to be 17, 3.6, and 0.08 μm, respectively (Allbritton et al., 1992). In pancreatic acinar cells, IP₃ was shown to act as a long-range messenger (Kasai and Petersen, 1994). In the case of SMG duct cells, therefore, IP₃Rs highly concentrated in the apical Ca²⁺ pools would be activated by IP₃ traveling from the basolateral region and would contribute to a polarized or gradient Ca²⁺ signal generated in the apical region, which might be prerequisite for efficient ductal functions.

In the present study, we demonstrated that the duct cells of the rat SMG did not respond to carbachol uniformly (Fig. 5). Pioneer cells showed rapid and dynamic [Ca²⁺]_i changes, while the less-reactive cells showed a slow and diffuse [Ca²⁺]_i transient. Other factors that might have produced such Ca²⁺ responses were considered, e.g., damage of cells caused by collagenase digestion or heterogeneous distribution of agonist receptor. However, in light of the fact that the distribution of the IP₃Rs completely overlapped with the Ca²⁺ initiation site (Fig. 7), we believe these to be unlikely explanations. The population of the cells highly expressing IP₃Rs in the duct was virtually identical to that of the cells highly reactive to carbachol. These observations indicated that the differential expression level of IP₃Rs was deeply correlated to the discrete Ca²⁺ response in individual duct cells.

The remaining problem is the characterization of the pioneer cell. The SMG duct consists of several cell types, so one of those previously identified may be functioning as the pioneer cell. Also, we have not shown conclusively that the pioneer cells controlled the change in [Ca²⁺]_i in the less-reactive cells in the manner of classical “pace maker” cells. However, our Ca²⁺ imaging suggested that the Ca²⁺ wave that originated in the pioneer cells spread to the less-reactive cells. It should be easy to determine whether second messengers are diffusing from the pioneer cells to neighboring less-reactive cells or whether the spreading Ca²⁺ wave is independent of the pioneer cells.

The authors are grateful to Mr. O. Katsumata for technical assistance of the immunoelectron microscopy, the staff in Nikon Instech for the real time confocal microscopy, and Dr. David Zacharias for critical reading.

[Ca²⁺]_i

intracellular calcium concentration

IP₃

inositol 1,4,5-trisphosphate

IP₃R

IP₃ receptor

PSS

physiological salt solution

SMG

submandibular gland