Rat anterior pituitary cells in vitro can partly reconstruct in vivo topographic affinities.код для вставкиСкачать
THE ANATOMICAL RECORD PART A 272A:548 –555 (2003) Rat Anterior Pituitary Cells In Vitro Can Partly Reconstruct In Vivo Topographic Afﬁnities TAKAHIRO NODA, MOTOSHI KIKUCHI, SACHIKO KAIDZU, AND TAKASHI YASHIRO* Division of Histology and Cell Biology, Department of Anatomy, Jichi Medical School, Tochigi, Japan ABSTRACT Hormone-producing cells in the rat anterior pituitary gland are not randomly distributed; rather, there are speciﬁc topographic afﬁnities among ﬁve cell types (Noda et al., Acta Histochem. Cytochem. 2001;34:313-319). In this study we reconstructed these afﬁnities, at least partially, in primary monolayer culture. Pituitary cells collected from adult male rats were enzymatically dispersed and cultured for 72 hr at a density of 1 ⫻ 105 cells/cm2. We double-immunostained cells using antibodies against hormones, and then used confocal laser microscopy to examine the ability of the cells to attach to each other. We also statistically analyzed the afﬁnity of all combinations of the ﬁve types of hormone-producing cells. We observed clusters by electron microscopy to identify junctional complexes between the cells. Confocal laser microscopy indicated that the features and attachment patterns of hormoneproducing cells in vivo were similar to those in vitro. Statistical analyses revealed that the rates at which the ﬁve types of hormone-producing cells attached to growth hormone (GH)-, prolactin (PRL), and luteinizing hormone (LH)-producing cells were unequal, which suggests there are speciﬁc topographic afﬁnities. The speciﬁc rates of adrenocorticotropic hormone (ACTH)-producing cell attachment to GH cells, LH to PRL cells, and PRL to LH cells were high, whereas that of PRL attachment to PRL cells was low. In addition, the rates correlated with the data from our previous in vivo study. Ultrastructural observations revealed few junctional complexes between hormone-producing cells. These results indicate that anterior pituitary hormone-producing cells can attach to speciﬁc types of cells by means of speciﬁc and/or nonspeciﬁc adhesion factors, and can reconstruct the topographic nature of the pituitary gland. Anat Rec Part A 272A:548 –555, 2003. © 2003 Wiley-Liss, Inc. Key words: anterior pituitary gland; topographic cell afﬁnity; primary monolayer culture; immunohistochemistry; confocal laser microscopy The anterior pituitary is a multifunctional gland that comprises a mixture of hormone-producing cells. Immunohistochemical studies of the distribution of such cells in the pituitary of various animals have shown that the localization of each cell is species-speciﬁc (Doerr-Schott, 1980). In rodents, ﬁve types of hormone-producing cells appear in a temporally and spatially speciﬁc fashion (Watanabe and Daikoku, 1979) via signals from the ventral diencephalon and oral ectoderm (Scully and Rosenfeld, 2002), but localize almost at random within the entire adult pituitary gland (Nakane, 1970; Baker and Gross, 1978). These cells, like folliculo-stellate (F-S) cells, are distributed with speciﬁc topographic afﬁnity in the rat anterior pituitary. Nakane (1970) found high topographic afﬁnity between growth hormone (GH)- and adrenocorticotropic hormone (ACTH)-producing cells. Siperstein and Miller (1970) and Yoshimura and Nogami (1981) pub© 2003 WILEY-LISS, INC. lished images of an ACTH cell adjacent to a GH cell. Nakane (1970) and Nogami and Yoshimura (1982) found that individual large, oval luteinizing hormone (LH)-/follicle-stimulating hormone (FSH)-producing cells are fre- Grant sponsor: Jichi Medical School, “The Research Award to JMS Graduate Student”. *Correspondence to: Takashi Yashiro, Division of Histology and Cell Biology, Department of Anatomy, Jichi Medical School, 3311-1 Yakushiji, Minamikawachi, Tochigi 3290498, Japan. Fax: ⫹81-285-445243. E-mail: firstname.lastname@example.org Received 5 December 2002; Accepted 12 February 2003 DOI 10.1002/ar.a.10065 AFFINITIES AMONG PITUITARY CELLS IN VITRO quently surrounded by cup-shaped prolactin (PRL)-producing cells. Gon (1987) observed that F-S cells form clusters, and Soji and Herbert (1989) showed images of clustered F-S cells surrounding the lumen. In addition, PRL cells (Nogami and Yoshimura, 1982; Shirasawa et al., 1985) or thyroid-stimulating hormone (TSH)-producing cells (Yashiro et al., 1981) sometimes form clusters. These morphological features should provide a suitable environment for cell-to-cell communication in the pituitary gland. However, when these features are identiﬁed by electron microscope, their interpretation is subjective according to the viewer. The interpretation of these data become standardized by observing a wide area of the gland by light microscopy after double-immunostaining the cells, and by statistical analysis (Noda et al., 2001). Enzymatically dispersed pituitary cells can form clusters in primary monolayer cultures as well as in vivo. Topographic afﬁnity in clusters merits further study, as few reports have examined this issue. Wilﬁnger et al. (1984) showed that LH cells locate in close proximity to PRL cells when seeded at high density. Allaerts et al. (1991) demonstrated that PRL and LH cells juxtapose in enriched suspension culture. However, the topographic afﬁnities among all types of hormone-producing cells in vitro have not been reported, so the relationship between in vivo and in vitro afﬁnities has remained obscure. We therefore used confocal microscopy to study the afﬁnities of all types of double-immunostained rat anterior pituitary hormone-producing cells in primary monolayer culture. We also examined cell-to-cell junctions in clusters by electron microscopy. MATERIALS AND METHODS Cell Culture Male Sprague-Dawley rats were purchased from Japan SLC, Inc. (Shizuoka, Japan). When the rats were 8-10 weeks old (and weighed 250-300 g), they were perfused with Ca2⫹- and Mg2⫹-free (CMF) Hank’s solution under Nembutal anesthesia. The anterior pituitary glands were excised and minced into small pieces that were incubated in CMF Hank’s solution containing 1% trypsin (Life Technologies, Inc., New York, NY) and 0.2% collagenase (Nitta Gelatin Inc., Osaka, Japan) for 20 min at 37°C. Thereafter, the pieces were incubated in the same solution containing 5 g/ml of DNase I (Boehringer-Mannheim, Mannheim, Germany) for 5 min at 37°C. After the digest was washed with CMF Hank’s solution, it was incubated in the same solution containing 0.3% ethylenediaminetetraacetic acid (Wako Pure Chemicals, Osaka, Japan) for 5 min at 37°C. Dispersed cells were separated from debris by centrifugation, rinsed and resuspended in CMF Hank’s solution by pipetting, and then ﬁltered through nylon mesh (Becton Dickinson Labware, Franklin Lakes, NJ). The ﬁltered cells were plated on eight-well glass chamber slides (1 cm2/well; Asahi Techno Glass, Chiba, Japan) at a density of 1 ⫻ 105 cells/cm2 in 400 l of Medium 199 with Earle’s salts (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (Sigma, St. Louis, MO). They were then cultured for 72 hr at 37°C in a humidiﬁed atmosphere of 5% CO2 and 95% air. Time-Lapse Observation of the Cluster Formation The cells were cultured under the conditions described above, in a CO2 gas culture chamber (Sankei Co. Ltd., 549 Tokyo, Japan) with a thermostat (Kokensha Engineering Co. Ltd., Tokyo, Japan), and examined by phase contrast microscopy (Diaphot-TMD; Nikon Corporation, Tokyo, Japan). Cluster formation was monitored using a time-lapse VTR system (Sankei Co. Ltd.) for 72 hr. Immunoﬂuorescence Staining Cultured cells ﬁxed with 10% buffered formalin for 10 min at room temperature were ﬁrst immersed in phosphate-buffered saline (PBS, pH 7.2) containing 1% Triton X-100 (Sigma), and then in normal goat serum for 1 hr at 30°C to reduce nonspeciﬁc immunoreactivity. Incubation with primary antibodies for 2 hr at 30°C was followed by a wash with PBS containing 1% Triton X-100, an incubation with secondary antibodies for 1 hr at 30°C, and another wash with PBS containing 1% Triton X-100. The specimens were immunostained with another species of antibody, as described above. Antibodies Rabbit polyclonal antibodies against rat GH (diluted 1:1,600), rat PRL (1:100), and ovine LH ␤ subunit (1: 3,200) were donated by Dr. K. Wakabayashi (Biosignal Research Center, Gunma University, Gunma, Japan). Rabbit polyclonal antibody against porcine ACTH 1-39 (1:1,600) was donated by Dr. F. Nakamura (Hokkaido University, Sapporo, Japan). Rabbit polyclonal antibody against rat TSH ␤ subunit (1:3,200) and guinea pig antibody against rat LH ␤ subunit (1:1,000) were obtained from the National Institutes of Health (Bethesda, MD). Guinea pig polyclonal antibody against human GH (1:200) was purchased from Biogenesis Ltd. (Poole, UK). Mouse monoclonal antibodies against rat PRL (1:400) and human ACTH C-terminal (1:200) were purchased from Chemicon International Inc. (Temecula, CA) and Cymbus Biotechnology Ltd. (Hampshire, UK), respectively. Nonspeciﬁc bindings of each primary antibody were checked using nonimmune rabbit or guinea pig serum and mouse IgG. Goat polyclonal antibody against rabbit IgG labeled with Texas Red (1:50) was purchased from ICN Pharmaceuticals Inc. (Aurora, OH). An antibody against guinea pig IgG labeled with ﬂuorescein isothiocyanate (FITC, 1:100) and mouse IgG labeled with FITC (1:100) were purchased from EY Laboratories Inc. (San Mateo, CA). The absence of cross reactions in double immunostaining was conﬁrmed in preliminary experiments. All antibodies were diluted with PBS containing 0.1% NaN3. Observation of Topographic Cell Afﬁnity The specimens were mounted with Vectashield (Vector Laboratories, Burlingame, CA), and clusters were observed with a confocal laser microscope (Yokogawa Electric Co., Tokyo, Japan) at wavelengths of 488 and 568 nm for FITC and Texas Red, respectively, to conﬁrm the frequency with which the speciﬁc cell types attached. The same ﬁeld was captured at both wavelengths and merged with the use of a computer program (MetaMorph, Universal Imaging Co., Downingtown, PA). The images were further processed with Adobe Photoshop 6.0 (Adobe Systems, San Jose, CA). Numerical Expression of Cell Attachment Panels for measuring relative attachment preferences were established from cell clusters at randomly selected 550 NODA ET AL. Fig. 1. Phase contrast microscopy observation of cluster formation. Cells in culture were video-recorded at a ﬁxed point under phase contrast microscopy. a-h: Extracted images acquired 0, 5, 10, 20, and 30 min; and 1, 3, and 24 hr, respectively, after the cells were plated. Bar: 50 m. levels. The circumference of a speciﬁc cell type in a cluster, and the length of attaching lines to counter-immunostained cells were determined by use of the computer software. Relative attachment preference was deﬁned as the rate of cell attachment, using the following formula: alcohols, and embedded in epoxy resin (Quetol 812; Nissin EM Co., Tokyo, Japan). The cells were sectioned into ultrathin slices with a Reichert-Nissei Ultracut S (Leica Inc., New York, NY), stained with uranyl acetate and lead citrate, and then observed under a Hitachi H-7600 electron microscope. rate of cell attachment (%) ⫽ length of line attached to counterstained cells/ circumference of stained cells ⫻ 100 Data Analysis Cultured cells from three or four animals in each experiment were statistically analyzed. Clusters were randomly selected under scanning microscopy to measure the rates of cell attachment. Proﬁles of immunostained GH (n ⫽ 922), PRL (n ⫽ 814), ACTH (n ⫽ 314), TSH (n ⫽ 237), and LH cells (n ⫽ 343) were obtained from over 30 clusters. The rates of attachment of one type of cell to other types are difﬁcult to compare because the varying numbers and shapes among cell types directly affect attachment preferences. Therefore, we normalized the attachment rates of various types of cells to one speciﬁc type of cells in a panel to eliminate this problem and to generate comparable values when the relative attachment preference was not speciﬁc. The attachment preferences of hormone-producing cells are presented as means ⫹ S.E.M. They were statistically analyzed by use of the nonparametric Kruskal-Wallis analysis of variance (ANOVA) of ranks, followed by a test for outliers. Electron Microscopy Observation For observation of junctional complexes in clusters, cultured cells were ﬁxed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.2 for 30 min at 4°C. After the cells were washed repeatedly with 0.1 M cacodylate buffer, they were postﬁxed with 1% OsO4 in 0.1 M cacodylate buffer for 30 min at 4°C, dehydrated in a series of graded RESULTS Microscopy Observation of Topographic Cell Afﬁnity After the anterior pituitary cells were cultured for 72 hr, they assumed solitary, paired, or clustered states. Most clusters comprised less than a dozen cells. Figure 1 shows cluster formation by pituitary cells. As the cells settled to the surface of the culture dishes, they appeared to form cell aggregates without migration in 1 hr. The aggregates gradually compacted to form clusters, and thereafter the number of cells remained constant. To determine whether cells in clusters proliferate, we used immunocytochemistry with a mouse anti-PCNA monoclonal antibody (ICN Biomedicals Inc., OH). All of the cells were negative for anti-PCNA antibody, indicating that cluster formation is not related to cell proliferation (data not shown). All ﬁve types of hormone-producing cells (GH, PRL, ACTH, TSH, and LH) formed clusters. With the exception of the PRL cells, the morphology of the cells did not signiﬁcantly differ from that in vivo. The PRL cells were rounder and larger than those in vivo. We used confocal laser microscopy to determine how the cell types formed clusters. Figures 2-4 are extracted two-dimensional images that show the topographic afﬁnities of all combinations of the hormone-producing cells. The following is a summary of our ﬁndings based on observations of hundreds of cells of each type. Round GH cells attached to all types of cells, but more frequently to GH cells (Fig. 2). Figure 2 is a typical image of a GH cell surrounded by an ACTH cell in a cluster. The topographic relationship between round PRL cells and LH cells was speciﬁc (Fig. 3). Some PRL AFFINITIES AMONG PITUITARY CELLS IN VITRO 551 Fig. 2. Immunoﬂuorescence-stained GH (green) and ACTH cells (red) in primary monolayer culture. Polygonal ACTH cells enclose a GH cell via cell projections. One ACTH cell totally envelopes a GH cell (arrow). Bar: 50 m. Fig. 3. Immunoﬂuorescence-stained PRL (green) and LH cells (red) in primary monolayer culture. An LH cell is enveloped by two or three PRL cells (arrow). Bar: 50 m. Fig. 4. Immunoﬂuorescence-stained ACTH (green) and LH cells (red) in primary monolayer culture. Bar: 50 m. cells entirely enveloped an LH cell. Other PRL cells attached to ACTH or TSH cells, but not as frequently as to LH cells. ACTH cells were stellate, polygonal, or irregularly shaped, and frequently enveloped GH cells (partially or entirely) with their cell projections (Fig. 2). ACTH cells seldom attached to ACTH, TSH, or LH cells (Fig. 4). TSH cells assumed various shapes (round or oval, stellate, polygonal, or irregular) and mainly attached to GH and PRL cells. LH cells, which were always round or oval, frequently attached to GH or PRL cells (Fig. 3). Two or three PRL cells often surrounded a single LH cell. arrangements, respectively, among the cell types. The values in Fig. 6 were not analyzed, since signiﬁcant heterogeneities were absent (as described above). We investigated whether the values of these 25 combinations in vitro (Figs. 5 and 6) correlated with the in vivo values found in our previous study (Figs. 9-13 in Noda et al., 2001). Statistical analyses showed a signiﬁcant overall correlation between the rates in vivo and those in vitro (P ⬍ 0.05, r ⫽ 0.67; Pearson’s correlation coefﬁcient test; Fig. 7). Statistical Analysis of Topographic Cell Afﬁnity Electron Microscopy Observation of the Clusters The rates of attachment between the types of hormoneproducing cells were statistically analyzed using a computer program, and the results are shown in Figs. 5 and 6. According to the Kruskal-Wallis test, the rates of attachment of the ﬁve types of cells to GH, PRL, and LH cells (Fig. 5a-c) were signiﬁcantly different, indicating that there were speciﬁc topographic afﬁnities among cell types. The rates of attachment of ACTH to GH cells (Fig. 5a), LH to PRL cells (Fig. 5b), and PRL to LH cells (Fig. 5c) were signiﬁcantly high, and that of PRL to PRL cells was signiﬁcantly low (Fig. 5b). This indicates close or remote To determine whether there were junctional complexes between the cells, we observed the clusters by electron microscopy. The cultured cells formed clusters connected by cell bodies or projections (Figs. 8 and 9). Some intermediate junctions and desmosomes were identiﬁed between the F-S cells, especially surrounding the folliculo-lumen (Fig. 8). On the other hand, junctional complexes including gap junctions were scarce between hormone-producing cells, as well as between hormone-producing and F-S cells, among hundreds of clusters (Fig. 9). 552 NODA ET AL. Fig. 5. Rates of attachment to (a) GH, (b) PRL, and (c) LH cells. The average percentage of the circumference of all cell types attached to (a) GH, (b) PRL, and (c) LH cells is graphed. Vertical lines represent the S.E.M. The rates of cell attachment of ﬁve types of cells to each type are signiﬁcantly different among types (Kruskal-Wallis test, P ⬍ 0.01). Values of (a) ACTH cells to GH cells, (b) LH cells to PRL cells, and (c) PRL cells to LH cells were signiﬁcantly higher (P ⬍ 0.01, asterisk), and that of PRL cells was signiﬁcantly lower (P ⬍ 0.01, asterisk) among ﬁve tested groups. Fig. 6. Rates of attachment to (a) ACTH and (b) TSH cells. The average percentage of the circumference of all cell types attached to (a) ACTH and (b) TSH cells is graphed. Vertical lines represent the S.E.M. The Kruskal-Wallis test revealed no signiﬁcant differences among types. DISCUSSION Previous electron microscopy studies have shown that various hormone-producing cells in the rat anterior pituitary have speciﬁc afﬁnities for each other (Nakane, 1970; Siperstein and Miller, 1970; Yashiro et al., 1981; Yoshimura and Yogami, 1981; Nogami and Yoshimura, 1982; Shirasawa et al., 1985). In an earlier work we numerically expressed topographic cell afﬁnity from data acquired using confocal laser microscopy (Noda et al., 2001). Furthermore, our statistical analysis objectively conﬁrmed that some combinations (such as GH and ACTH cells) locate in juxtaposition (Nakane, 1970; Siperstein and Miller, 1970; Yoshimura and Nogami, 1981), and that PRL and LH/FSH cells (Nakane, 1970; Nogami and Yoshimura, 1982) have a high afﬁnity for each other. Conversely, cells that do not attach (such as ACTH and TSH types (Nakane, 1970)) have a speciﬁc and low afﬁnity for each other. Anterior pituitary cells in primary culture tend to form clusters (Fig. 1). Our results indicate that cluster formation issues from contingent encounters and circumstantial attachment of the cells—not from active migration. Furthermore, cells in clusters were immunohistochemically negative for anti-PCNA antibody, which suggests that cluster formation is caused only by cell attachment. A few investigators previously described the topographic afﬁnity AFFINITIES AMONG PITUITARY CELLS IN VITRO Fig. 7. Correlation between rates of cell attachment in vivo and in vitro (n ⫽ 25). Statistical analysis showed a signiﬁcant positive correlation between rates in vivo (Figs. 9-13 in Noda et al., 2001) and those in vitro (P ⬍ 0.05, r ⫽ 0.671; Pearson’s correlation coefﬁcient test). between clustered cells. Allaerts et al. (1991) studied the topographic afﬁnity of hormone-producing cells in vitro and found that enriched suspension cultures of PRL and LH cells have an afﬁnity for each other. Wilﬁnger et al. (1984) showed that LH cells position themselves in close proximity to PRL cells after they are plated at high density, which enhances cell-to-cell interaction in clusters. However, none of these reports examined all combinations of known hormone-producing cells. The present work is the ﬁrst to describe the ability of anterior pituitary cells to reconstruct speciﬁc cell attachments in primary monolayer culture. The morphological features of clustered hormone-producing cell types in vitro are essentially identical to those in vivo, with the exception of PRL cells, which are larger and mainly oval in vitro (Fig. 3) but polygonal or irregularly shaped in vivo (Nogami and Yoshimura, 1982; Shirasawa et al., 1985). Wilﬁnger et al. (1984) and Snyder et al. (1976) reported that PRL cells changed round shape in vitro, and Orgnero de Gaisán et al. (1997) suggested that these morphological changes are the result of the disruption of the normal paracrine relationship and hypothalamic regulatory system. Our observations showed that the shape of the PRL cells changed in vitro, but the number did not increase. Characteristic cell attachment in vivo between speciﬁc cell types was notably reconstructed in vitro (Figs. 2-4). For instance, the attachment of ACTH to GH cells via cell projections (Fig. 2), and the envelopment of an LH cell by two or three PRL cells (Fig. 3) were similar to our previous ﬁndings in vivo (Noda et al., 2001). Statistical analyses of numerically expressed cell afﬁnities showed that the rates of cell attachment varied according to the combination of cell types (Figs. 5 and 6). The rate of cell attachment should vary with the degree of speciﬁcity of their topographic cell afﬁnities. If none of these cells had speciﬁc afﬁnities, the rates among cell types would be similar. Our statistical analyses showed that cell type topographic afﬁnities are speciﬁc in vitro, at least with some combinations of cell types (Fig. 5a-c). However, it is difﬁcult to compare rates among cell types 553 because the numbers and shapes of the cells directly affect the rates. We therefore compared the rates of attachment of various types of cells to one other type of cell. Our previous study (Noda et al., 2001) showed that the afﬁnities of hormone-producing cells are speciﬁc in vivo. We examined the correlation between the rates of cell attachment in vitro and the corresponding values in vivo (n ⫽ 25; Figs. 9-13 in Noda et al., 2001). Statistical analyses showed a signiﬁcant overall correlation between the rates in vivo and those in vitro (Fig. 7). This ﬁnding suggests that the speciﬁc arrangement in vivo was at least partly reconstructed in vitro. Combinations with speciﬁc high or low afﬁnities were essentially the same in vivo and in vitro. These ﬁndings indicate that the hormone-producing cells of the rat anterior pituitary can reconstruct speciﬁc topographic afﬁnities in vitro. Our observations revealed that the cells formed clusters attached to each other at random, and that the clusters became smaller without a decrease in the number of cells (Fig. 1). Therefore, topographic afﬁnity may be reconstructed after cluster formation is mediated by unknown factors, such as junctional complex and cell adhesion molecules. We observed intermediate junctions and desmosomes between F-S cells (Fig. 8), but very few junctions between hormone-producing cells despite high-afﬁnity pairing (GH and ACTH cells in Fig. 9). These ﬁndings are mostly consistent with previous in vivo reports that showed tight junctions, intermediate junctions, and desmosomes (Soji and Herbert, 1989) between F-S cells, but only desmosomes between PRL cells (Saunders et al., 1982). In vitro studies have found tight junctions and gap junctions in aggregates (Wilﬁnger et al., 1984), tight junctions and intermediate junctions between F-S cells (Allaerts and Denef, 1989), and tight junctions in PRL cell/GH cell-enriched and LH cell-enriched aggregates (Van der Schueren et al., 1982). However, these investigations were performed under speciﬁc conditions; some types of cells were concentrated, and the topographic afﬁnity in vivo was ignored. The scarcity of junctional complexes, at least in our culture system, indicates that they are not the cause of reconstructed topographic afﬁnity between hormone-producing cells. Some cell adhesion molecules have been identiﬁed in the rat anterior pituitary, including NCAM (Langley et al., 1987; Berardi et al., 1995) and PB-cadherin (Sugimoto et al., 1996). Spangler and Delidow (1998) observed P- and N-cadherin expression in PRL-producing rat pituitary 253-1 cells, and Heinrich et al. (1999) reported that N-cadherin mediates cell aggregation in rat somatolactotropic GH3 cells. These molecules may serve a function during cluster formation, but there is no evidence that cell adhesion molecules play a role in topographic afﬁnity in the anterior pituitary. Some suggestions have been proposed on the role of cell adhesion molecules in organization of other endocrine tissues, such as the pancreatic islets, that have various types of hormone-producing cells. Some distribution patterns (such as that of B cells concentrated in the center of the islets, and of A, D, and PP cells dispersed around B cell clusters) can be reconstructed according to the expression of NCAM or cadherin in vitro (Cirulli et al., 1994). Cirulli et al. (1994) suggested that these differences in expression among cell types contribute to the characteristic distribution of cells within the islets of Langerhans. Because the topographic afﬁnity in Langerhans islets tends to be between homotypical cells, and cadherins mainly mediate attachment 554 NODA ET AL. Fig. 8. Electron micrograph of a cluster primarily cultured for 72 hr. In addition to hormone-producing cells, several dispersed F-S cells (nonhormone-producing cells) gathered to form a pseudolumen within the central portion. Microvilli are observed in the lumen. A desmosome (arrow) and long intermediate junctions (double arrow) can be seen between F-S cells. * indicates the outside of the cell cluster. Bar: 1 m. Fig. 9. Electron micrograph of a cluster primarily cultured for 72 hr. An ACTH cell, characterized by the row arrangement of secretory granules along the cell membrane, encircles a neighboring GH cell. G, GH cell; A, ACTH cell. Bar: 10 m. Inset shows the region outlined by the rectangle at a higher magniﬁcation. No junctional complexes are evident between these cells. AFFINITIES AMONG PITUITARY CELLS IN VITRO between them, heterotypical topographic afﬁnity in the anterior pituitary might not be caused solely by cadherins, but by some other, unknown factors as well. This issue should be further investigated. The functional signiﬁcance of speciﬁc pairing between hormone-producing cells remains unclear. The attachment of one cell to another is convenient for direct cell-tocell communication, but we did not observe gap junctions between cultured hormone-producing cells in this study. Intracellular paracrine communication does not always require communicating cells to be juxtaposed. However, a report by Abraham et al. (1996) showed that PRL gene expression differs depending on whether a mammotrope is alone, is in contact with one other mammotrope, or adheres to a non-mammotrope. In conclusion, cultured rat anterior pituitary hormoneproducing cells form clusters with a speciﬁc cell-to-cell afﬁnity similar to that in vivo. Some cell adhesion molecules may be involved with cluster formation. The present study provides information that will be helpful in future investigations of intracellular communication. ACKNOWLEDGMENTS We are grateful to K. Inose and M. Yatabe for their excellent technical assistance. This study was partly supported by “The Research Award to JMS Graduate Student” from Jichi Medical School. LITERATURE CITED Abraham JE, Faught JW, Frawley SL. 1996. Intercellular communication: relative importance of cellular adhesion and paracrine signaling to hormonal gene expression. Endocrinology 137: 4050-4053. Allaerts W, Denef C. 1989. Regulatory activity and topological distribution of folliculo-stellate cells in rat anterior pituitary cell aggregates. Neuroendocrinology 49:409-418. Allaerts W, Mignon A, Denef C. 1991. Selectivity of juxtaposition between cup-shaped lactotrophs and gonadotrophs from rat anterior pituitary in culture. Cell Tissue Res 263:217-225. Baker LB, Gross SD. 1978. Cytology and distribution of secretory cell types in the mouse hypophysis as demonstrated with immunocytochemistry. Am J Anat 153:193-215. Berardi M, Hindelang C, Laurent-Huck MF, Langley K, Rougon G, Félix MJ, Stoeckel EM. 1995. Expression of neural cell adhesion molecules, NCAMs, and their polysialylated forms, PSA-NCAMs, in the developing rat pituitary gland. Cell Tissue Res 280:463-472. Cirulli V, Baetens D, Rutishauser U, Halban AP, Orci L, Rouiller GD. 1994. Expression of neural cell adhesion molecule (N-CAM) in rat islets and its role in islet cell type segregation. J Cell Sci 107:14291436. Doerr-Schott J. 1980. Immunohistochemistry of the adenohypophysis of non-mammalian vertebrates. Acta Histochem Suppl Band XXII S:185-223. 555 Gon G. 1987. The origin of ciliated cell cysts of the anterior pituitary. An experimental study in the rat. Virchows Arch A 412:1-9. Heinrich AC, Lail-Trecker RM, Peluso JJ, White AB. 1999. Negative regulation of N-cadherin-mediated cell-cell adhesion by the estrogen receptor-signaling pathway in rat pituitary GH3 cells. Endocrine 10:67-76. Langley KO, Aletsee CM, Gratzl M. 1987. Endocrine cells share expression of N-CAM with neurones. FEBS Lett 220:108-112. Nakane KP. 1970. Classiﬁcations of anterior pituitary cell types with immunoenzyme histochemistry. J Histochem Cytochem 18:9-20. Noda T, Kaidzu S, Kikuchi M, Yashiro T. 2001. Topographic afﬁnities of hormone-producing cells in the rat anterior pituitary gland. Acta Histochem Cytochem 34:313-319. Nogami H, Yoshimura F. 1982. Fine structural criteria of prolactin cells identiﬁed immunohistochemically in the male rat. Anat Rec 202:261-274. Orgnero de Gaisán E, Maldonado C, Diaz Gavier FM, Aoki A. 1997. Diversity of pituitary cells in primary cell culture. An immunocytochemical study. Ann Anat 179:453-460. Saunders LS, Reifel WC, Shin HS. 1982. Desmosomes between mammotrophs suggest the existence of a functional syncytium. Acta Anat 114:74-80. Scully MK, Rosenfeld GM. 2002. Pituitary development: regulatory codes in mammalian organogenesis. Science 295:2231-2235. Shirasawa N, Kihara H, Yoshimura F. 1985. Fine structural and immunohistochemical studies of goat adenohypophyseal cells. Cell Tissue Res 240:315-321. Siperstein RE, Miller JK. 1970. Further cytophysiologic evidence for the identity of the cells that produce adrenocorticotrophic hormone. Endocrinology 86:451-486. Snyder J, Wilﬁnger W, Hymer CW. 1976. Maintenance of separated rat pituitary mammotrophs in cell culture. Endocrinology 98:25-32. Soji T, Herbert CD. 1989. Intercellular communication between rat anterior pituitary cells. Anat Rec 224:523-533. Spangler RP, Delidow CB. 1998. Co-regulation of pituitary tumor cell adhesion and prolactin gene expression by glucocorticoid. J Cell Physiol 174:115-124. Sugimoto K, Honda S, Yamamoto T, Ueki T, Monden M, Kaji A, Matsumoto K, Nakamura T. 1996. Molecular cloning and characterization of a newly identiﬁed member of the cadherin family, PB-cadherin. J Biol Chem 271:11548-11556. Van der Schueren B, Denef C, Cassiman JJ. 1982. Ultrastructural and functional characteristics of rat pituitary cell aggregates. Endocrinology 110:513-523. Watanabe GY, Daikoku S. 1979. An immunohistochemical study on the cytogenesis of adenohypophyseal cells in fetal rats. Develop Biol 68:557-567. Wilﬁnger WW, Larsen JW, Downs RT, Wilbur LD. 1984. An in vitro model for studies of intercellular communication in cultured rat anterior pituitary cells. Tissue Cell 16:483-497. Yashiro T, Nogami H, Yoshimura F. 1981. Immunohistochemical study of the postnatal development of pituitary thyrotrophs in the rat, with special reference to cluster formation. Cell Tissue Res 216:39-46. Yoshimura F, Nogami H. 1981. Fine structural criteria for identifying rat corticotrophs. Cell Tissue Res 219:221-228.