Early observations of pancreatic ultrastructure during hemorrhagic necrosis in the rat with closed duodenal pouch]

Many problems are still unanswered in the pathogenesis of acute clinical and experimental pancreatic necrosis. A new technique which can be performed in the rat seems a suitable model for reflux pancreatic necrosis without artificial pressure changes in the ductal system. A closed duodenal loop is obtained with ligation proximal and distal to Vater's ampulla and a gastroenteroanastomosis is associated to avoid intestinal obstruction. All the rats die with hemorrhagic pancreatic necrosis in 36 hours. After 12 hours from the operation ductal and acinar lumina are enlarged. In the centroacinar and intercalated duct cells some lysosomes and mitochondria with clear matrix and reduced cristae are detected. Intercellular junctions in ducts and acini have normal morphology. In the basal cytoplasm of acinar cells some prominent autophagic vacuoles are detectable. After 24 hours in the acinar cells autophagic vacuoles are greatly increased and basal cytoplasmic degeneration often occurs, with plasmalemma and basal lamina interruptions. Intercellular junctions are apparently unaffected until cell necrosis sets in. In blood capillaries endothelial cells are swollen, fibrin thrombosis, hemorrhage and leucocyte infiltration are often detectable. As lysosomal activity occurs also in different kinds of experimental pancreatic necrosis, it could be a common pathogenetic factor, responsible for hydrolytic enzyme activation and for vascular damage in the early stages of hemorrhagic pancreatic necrosis.     

Distribution of Na+,K+-ATPase in rat exocrine pancreas as monitored by K+-NPPase cytochemistry and [3H]-ouabain binding: a plasma membrane protein found primarily to be ductal cell associated

We investigated the distribution of Na+,K+-ATPase in rat exocrine pancreas. By use of enzymatic dissociation techniques, pancreatic acini (containing acinar cells and centroacinar ductal cells in a ratio of about 10:1) and all major classes of pancreatic ducts were isolated and analyzed for the presence of Na+,K+-ATPase using K+-NPPase cytochemistry and [3H]-ouabain binding assays. Ultrastructural analysis demonstrated a basolateral localization of ouabain-sensitive enzyme activity in all classes of pancreatic ducts, although the degree of activity varied among the various classes. Qualitative analysis (scale of 0 to + + +) indicated the following enzyme distribution: centroacinar ductal cells (+); intralobular ducts (+ +); interlobular ducts (+ + +); main duct (+ +). In contrast, no reaction product was associated with pancreatic acinar cells even when observed adjacent to enzyme-positive centroacinar ductal cells. Parallel experiments monitoring [3H]-ouabain binding supported the cytochemical studies. When expressed as femtomoles [3H]-ouabain/microgram DNA, the following values were obtained: whole pancreas, 100.3; ducts (pooled intralobular and interlobular), 337.0; acini, 48.2. The acinar value is complicated by the fact that acini contain both acinar and centroacinar cells, but in light of the cytochemical observations we suggest that most of the [3H]-ouabain binding is due to the few ductal cells present in acini. The results suggest that Na+,K+-ATPase is primarily associated with the ductal epithelium of the exocrine pancreas and is differentially distributed among the different classes of ducts.     

Sox9+ ductal cells are multipotent progenitors throughout development but do not produce new endocrine cells in the normal or injured adult pancreas

One major unresolved question in the field of pancreas biology is whether ductal cells have the ability to generate insulin-producing β-cells. Conclusive examination of this question has been limited by the lack of appropriate tools to efficiently and specifically label ductal cells in vivo. We generated Sox9CreER(T2) mice, which, during adulthood, allow for labeling of an average of 70% of pancreatic ductal cells, including terminal duct/centroacinar cells. Fate-mapping studies of the Sox9(+) domain revealed endocrine and acinar cell neogenesis from Sox9(+) cells throughout embryogenesis. Very small numbers of non-β endocrine cells continue to arise from Sox9(+) cells in early postnatal life, but no endocrine or acinar cell neogenesis from Sox9(+) cells occurs during adulthood. In the adult pancreas, pancreatic injury by partial duct ligation (PDL) has been suggested to induce β-cell regeneration from a transient Ngn3(+) endocrine progenitor cell population. Here, we identify ductal cells as a cell of origin for PDL-induced Ngn3(+) cells, but fail to observe β-cell neogenesis from duct-derived cells. Therefore, although PDL leads to activation of Ngn3 expression in ducts, PDL does not induce appropriate cues to allow for completion of the entire β-cell neogenesis program. In conclusion, although endocrine cells arise from the Sox9(+) ductal domain throughout embryogenesis and the early postnatal period, Sox9(+) ductal cells of the adult pancreas no longer give rise to endocrine cells under both normal conditions and in response to PDL.     

Endocytosis of native and cationized ferritin by intralobular duct cells of the rat parotid gland

Summary The ability of the intralobular ducts of the rat parotid gland to take up protein from the lumen was examined after retrograde infusion of native and cationized ferritin. At high concentrations (3–10 mg/ml), cells of both intercalated- and striated ducts avidly internalized the tracers. No differences were noted in the mode of uptake or fate of native or cationized ferritin. Large, apical ferritin-containing vacuoles up to 5 m in size were present in cells of the intercalated ducts after infusion for 15 min. Small, smooth-surfaced spherical or flattened vesicles and tubules containing ferritin were also observed, often in association with the large vacuoles. Ferritin uptake increased with increasing infusion time, up to 1 h. Uptake by the striated ducts was less consistent than by the intercalated ducts, and occurred mainly in small vesicles and tubules. Secondary lysosomes became labeled with ferritin in both cell types. Ferritin was not observed in the Golgi saccules, nor was it discharged from the cells at the basolateral surfaces. At low concentrations (0.3–1 mg/ml), uptake was reduced, especially by cells of intercalated ducts, and differences were noted in the behavior of the two tracers. Cationized ferritin was internalized mainly into vesicles and tubules of cells of striated ducts; little uptake of native ferritin occurred at low concentrations. These results demonstrate that the ductal cells of the salivary glands are capable of luminal endocytosis of foreign proteins. They also suggest that in addition to modifying the primary saliva by electrolyte reabsorption and secretion, and secretion of various glycoproteins, the ductal cells are able to reabsorb proteins secreted by the acinar cells.     

Distribution of aquaporin 1 in the rat pancreatic duct system examined with light- and electron-microscopic immunohistochemistry

The pancreatic duct is the major site for the secretion of pancreatic fluid, but the pathway of water transport in this system is not known. Recently, intense signal for mRNA of aquaporin 1 (AQP1) water channels was detected in isolated rat interlobular ducts. Therefore, we performed light- and electron-microscopic (EM) immunohistochemistry for AQP1 in the rat pancreatic ducts. AQP1 immunoproducts were not observed in the acinar cells, centroacinar cells or intercalated ducts. In the smaller intralobular ducts less than 10 microm in diameter (the lumen plus duct cells), most cells were immunonegative. AQP1-positive cells appeared in intralobular ducts 10-15 microm in diameter. In small and medium-sized interlobular ducts 15-70 microm in diameter surrounded by periductal connective tissue 2-40 microm thick, most cells were AQP1 positive with various degrees of immunoreactivity. In the larger interlobular ducts, the expression of AQP1 was variable, ranging from immunopositive to negative. In the main pancreatic duct, most cells were negative for AQP1. EM immunohistochemistry of the intralobular and small interlobular ductal epithelial cells showed that the AQP1 immunoproducts were more abundant in the basolateral membrane than in the apical membrane, though they were present in both membranes. In the medium-sized interlobular ducts, AQP1 immunoproducts were distributed densely along the apical, lateral interdigitation and basal membrane of the epithelial cells. In the various sizes of interlobular ducts, immunoproducts were associated not only with the plasma membrane, but also with the caveolae and vesicle-like structures. Secretin did not induce any significant difference in AQP1 expression and cellular and subcellular localization. These results indicate that the expression and subcellular localization of AQP1 vary considerably depending on the duct size, which may reflect water transport characteristics in the different divisions of the pancreatic duct system.     

Aggregation of macrophages in the tips of intestinal villi in guinea pigs: their possible role in the phagocytosis of effete epithelial cells

The immunoreactivity of a monoclonal antibody against cell suspensions from guinea pig adrenal glands was examined at light- and electron-microscopic levels. In addition to the cell surface membrane of adrenocortical cells, the antibody labeled specific sites in the pancreas, liver and testis, but did not label any of the other tissues examined. In the pancreas, microvilli-like processes and the cell surface membrane of centroacinar cells were immunoreactive to the antibody. The microvilli of interlobular duct cells and pancreatic duct cells were also immunoreactive. In the liver, bile canalicular microvilli of hepatocytes were exclusively labeled. Membrane structures of cell organelles, mainly mitochondria, in testicular Leydig cells were also labeled. Immunoblot analysis showed that the monoclonal antibody bound to two common bands at molecular weights of approximately 62 kDa and 110 kDa in the pancreas, liver, testis, and adrenal gland. The two bands reacted with the digoxigenin-conjugated lectin, Sambucus nigra agglutinin (SNA), which recognizes sialic acid linked (2–6) to galactose. Reaction patterns of SNA in the pancreas, liver and testis were similar to those of the monoclonal antibody; pancreatic centroacinar cells and interlobular duct cells, hepatocyte bile canaliculi and testicular Leydig cells were densely stained with SNA. Thus, the monoclonal antibody recognizes two common membrane glycoproteins containing sialic acids in the pancreas, liver, testis and adrenal cortex.     

Fine reconstruction of the pancreatic ductular system at the onset of pancreatitis

The three-dimensional structure of the pancreatic ductular system (from the intercalated duct to the intercellular secretory canaliculus) is controversial and unclear. The aim of this study is to reveal the three-dimensional structure of the pancreatic ductular sysytem at the onset of pancreatitis. One day following rat pancreatic duct ligation, dilated lumina from the pancreatic ductular system were reconstructed by light microscopic and scanning electron microscopic examination of pancreatic tissue serial sections. The existence of the intra-acinar duct, which is formed only by centroacinar cells and interconnects the adjacent central lumina in an acinus, was demonstrated. The intercellular secretory canaliculi, which are the terminal parts of the pancreatic ductular system, anastomose and end blindly in the intercellular space located between adjacent lateral surfaces of the acinar cells. The intercalated ducts, the intra-acinar ducts, the central lumina, and the intercellular secretory canaliculi are arranged together in a complex connecting and branching system. However, there were no anastomoses found among the central lumina or acini.     

Evidence for a membrane carbonic anhydrase IV anchored by its C-terminal peptide in normal human pancreatic ductal cells

The high concentration of HCO₃ ^– ions (150 mM) in the human pancreatic ducts raises the question of the membrane proteins responsible for their secretion in addition to the Cl^–/HCO₃ ^– exchanger. In this study, we investigated the expression of carbonic anhydrase IV (CA IV), a possible candidate. Experiments were carried out on specimens of normal human pancreas obtained from brain-dead donors (n=9) as well as on isolated human ductal cells. Two antibodies were generated: CA IV NH₂ antibody directed against the NH₂ terminal of human glycosyl phosphatidylinositol (GPI)-anchored CA IV and CA IV COOH antibody directed against the COOH terminal of the same protein before its association with a GPI in the rough endoplasmic reticulum. A 35-kDa CA IV was detected in the homogenates of human pancreas. Immunocytochemistry demonstrated the expression of CA IV in centroacinar cells and in intercalated, intralobular, and interlobular ductal cells. The immunoreactivity observed with the CA IV COOH antibody was mainly localized on luminal membranes of ductal cells. Treatment of purified plasma membranes with phosphatidylinositol-phospholipase C indicated that the CA IV expressed in pancreatic ducts was not GPI-anchored. Its detection in the same extracts by the CA IV COOH antibody indicated that it was anchored by a hydrophobic segment at the carboxy terminal. Taken together, these results suggest that normal human pancreatic ductal cells express a 35-kDa CA IV anchored in their luminal plasma membrane by a hydrophobic segment of the COOH terminus. In view of its localization and its mode of anchorage in luminal plasma membranes, this CA IV may participate in the maintenance of luminal pH.The first two authors have contributed equally to this work     

Histochemical localization and analysis of blood group-related antigens in human pancreas using immunostaining with monoclonal antibodies and exoglycosidase digestion

We examined the distribution of blood group-related antigens using an indirect immunoperoxidase method with monoclonal antibodies (MAb) directed to A, B, H, Lewis a (Lea), Lewis b (Leb), Lewis x (Lex), and Lewis y (Ley) antigens and Type 1 precursor chain in human pancreas. Effects of prior digestion with exoglycosidases on MAb stainings were simultaneously investigated. A, B, H, Leb, and Ley antigens were detected in acinar cells and interlobular duct cells but not in centroacinar cells, intercalated duct cells, and islet of Langerhans cells. The expression of these antigens in acinar cells was not dependent on Lewis type and secretor status of the tissue donors, whereas that in interlobular duct cells was strictly dependent on secretor status. The distribution pattern of these antigens in acinar cells was not homogeneous, i.e., cells producing H antigens expressed both Leb and Ley antigens but not A or B antigens, whereas those producing A or B antigens did not secrete Leb and Ley as well as H antigens. Digestion with alpha-N-acetylgalactosaminidase or alpha-galactosidase resulted in the appearance of Leb and Ley antigens as well as H antigen in acinar cells producing A and/or B antigens. Type 1 precursor chain was not detected in pancreatic tissues from secretors but appeared in acinar cells producing H antigen after alpha-L-fucosidase digestion, which also disclosed Lex but not Lea antigen in acinar cells expressing both Leb and Ley. In some non-secretors, MAb against Type 1 precursor chain reacted with acinar cells without enzyme digestion. Although Lea antigen was not detected in acinar cells, it was found in centroacinar cells, intercalated duct cells, and interlobular duct cells from all individuals examined except two Le(a-b-) secretors. After sialidase digestion, Lex antigen appeared in centroacinar and intercalated duct cells from some individuals. Sialidase digestion also elicited reactivity with MAb against Type 1 precursor chain in islet of Langerhans cells from some individuals. These results demonstrate the complexity in the pattern of expression and regulation of blood group-related antigens in different cell types of human pancreas. Such complexity may largely be ascribed to differences in individual genotypes and in gene expression patterns of different cell types.     

Visualization of CD44 and CD133 in Normal Pancreas and Pancreatic Ductal Adenocarcinomas: Non-overlapping Membrane Expression in Cell Populations Positive for Both Markers

Tumor-initiating cells of pancreatic ductal adenocarcinoma (PDAC) have been isolated based on expression of either CD133 or CD44. The authors aimed to visualize pancreatic cells simultaneously expressing both these cell surface markers by employing the same antibodies commonly used in cell-sorting studies. Normal and diseased pancreatic tissue, including 51 PDAC cases, were analyzed. CD44 and CD133 expression was determined by immunohistochemical double staining on formalin-fixed material and subcellular protein distribution evaluated by immunofluorescence/confocal microscopy. In the normal pancreas, CD44 and CD133 were coexpressed in the centroacinar regions but in non-overlapping subcellular compartments. As expected, CD44 was found mainly basolaterally, whereas CD133 was present on the apical/endoluminal membrane. This was also the case in chronically inflamed/atrophic pancreatic tissue and in PDAC. In some malignant ducts, CD44 was found at the apical cell membrane adjacent to but never overlapping with CD133 expression. CD44 level was significantly associated with the patient’s lymph node status. In conclusion, a CD44+/CD133+ cell population does exist in the normal and neoplastic pancreas. The preferentially centroacinar localization of the doubly positive cells in the normal parenchyma suggests that this population could be of particular interest in attempts to identify tumor-initiating cells in PDAC. This article contains online supplemental material at http://www.jhc.org. Please visit this article online to view these materials.     

Ultrastructural Distribution of Calcium in Cutaneous Electroreceptor Organs of Teleost Fish

The calcium distribution in the ampullary electroreceptor and the type B electrore-ceptor organ (gymnarchomast) of Gymnarchus niloticus (Gymnarchidae) and in the tuberous organ of Apteronotus leptorhynchus (gymnotidae) was studied. Endogenous calcium appeared as electron-dense precipitates when the cutaneous organs were pre-fixed with phosphate-buffered glutaraldehyde and postfixed with osmium tetroxide plus potassium bichromate. Calcium precipitates were localized in both intracellular compartments of sensory cells and afferent nerve fibers. In contrast to sensory cells, small amounts of calcium precipitates were found in the cytoplasm of accessory cells. In sensory cells, electron-dense deposits were apparent mainly in synaptic vesicles near synaptic ribbons, inside vacuoles of the endoplasmic reticulum, and between the layers of the nuclear membrane. Very few deposits were found in mitochondria. Precipitates were also observed within the axons of afferent nerves and between the layers of the myelin sheath. The synaptic cleft was devoid of calcium. Calcium deposits have a specific cellular distribution in electro-receptor organs of teleost fish.     

Pancreatic passenger leukocytes in organs obtained from cadaver donors for allotransplantation

In previous studies we have demonstrated that human normal pancreata were populated by interstitial mononuclear cells constituted by dendritic cells (70%), macrophages (30%) and few B cells (less than 1%). Furthermore, the endocrine parenchyma and the centroacinar cells appeared negative for the detection of HLA-A,B,C molecules. In the present report, 15 pancreata from cadaver donors were studied by means of immunohistochemical methods. 5 out of 15 organs were found to contain a larger amount of interstitial hematic cells that were constituted mainly by macrophages and B cells, together with a small percentage of granulocytes and T cells. In such "infiltrated" pancreata both endocrine parenchyma and centroacinar cells displayed a positivity for HLA-A,B,C antigens. As HLA-class I antigens have importance in allo-recognition and rejection responses, our data suggest that pre-transplant biopsies may be useful for a better evaluation of the pancreatic tissue immunogenicity.     

Localization of calcium in roots and microsomal membranes of corn by direct pyroantimonate precipitation

Many cell membrane systems, including microsomal vesicles of corn, are able to regulate calcium levels both in vivo and in vitro, often in an ATP-dependent, calmodulin-stimulated fashion. The purpose of this study was to determine calcium distribution in meristematic cells of intact tissue and microsomal vesicles from corn roots using direct pyroantimonate-osmium fixation. In root cells, precipitates were localized in mitochondria, plastids, the nucleus, endoplasmic reticulum, Golgi apparatus, and along the plasma membrane. Plasma membrane-enriched microsomal vesicles isolated from corn roots incubated in media to permit calcium transport before pyroantimonate-osmium fixation show internal precipitates associated with the membrane and in the lumen of the vesicles. De-staining of the sections with 1 mM EDTA or EGTA removed precipitate from the sections, confirming the presence of calcium in the antimonate precipitates. These data support biochemical data that this same membrane preparation exhibited ATP-dependent calcium sequestration that was stimulated by calmodulin, as measured by retention of 45Ca. This provides evidence that these membranes are responsible for ATP-requiring, calmodulin-stimulated calcium transport in the intact cell.     

Reduced expression of the membrane skeleton protein beta1-spectrin (SPTBN1) is associated with worsened prognosis in pancreatic cancer

Spectrins are members of the superfamily of F-actin cross linking proteins that are important as scaffolding proteins for protein sorting, cell adhesion, and migration. In addition, spectrins have been implicated in TGF-beta signaling. The aim of the present study was to analyze the expression and localization of beta1-spectrin (SPTBN1) in pancreatic tissues. mRNA levels of SPTBN1 in cultured pancreatic cancer cell lines, as well as in normal pancreatic tissues (n=18), chronic pancreatitis (n=48) and pancreatic cancer tissues (n=66) were analyzed by real time quantitative RT-PCR. Localization of SPTBN1 in pancreatic tissues was determined by immunohistochemistry. SPTBN1 staining was assessed semi-quantitatively in 55 cancer tissues and survival analysis was carried out using the Kaplan-Meier method. Median SPTBN1 mRNA levels were 6.0-fold higher in pancreatic cancer tissues compared to the normal pancreas (p<0.0001) and 2.2-fold higher compared to chronic pancreatitis tissues (p=0.0002). In the normal pancreas, SPTBN1 was present in the cytoplasm of normal ductal cells and occasionally in pancreatic acinar and centroacinar cells. In pancreatic cancer tissues, SPTBN1 was present in the cytoplasm of pancreatic cancer cells. Low SPTBN1 protein expression indicated a tendency for worsened prognosis with a median survival of 14.0 months, versus 23.8 months for patients whose tumors expressed moderate/high levels of SPTBN1. In conclusion, reduced SPTBN1 expression correlated with shorter survival of pancreatic cancer patients, suggesting a tumor suppressor function of this gene, as has already been shown for other malignancies of the gastrointestinal tract.     

Fine structure of the parotid gland in the crest-tailed marsupial-rat (Dasyuroides byrnei).

The parotid gland of Dasyuroides byrnei was examined by light microscopy, and transmission and scanning electron microscopy. The acini were composed predominantly of seromucous cells with a few mucous cells. The seromucous cells were light or dark cells containing acidophilic spherical granules of moderate to high electron density and had well-developed cytoplasmic organelles-ordinary mitochondria and large mitochondria with tubular cristae, RER with vesicular or tubular elements, and Golgi apparatus with lamellae, vesicles and vacuoles. The mucous cells had basophilic amorphous granules of low electron density, like those of ordinary mucous cells. The intercalated ducts were composed of simple cuboidal light cells having a few electron-dense granules. The striated ducts consisted of tall columnar light cells containing numerous vesicles and mitochondria with tubular cristae, the same as found in acinar seromucous cells.     

Close association of centroacinar/ductular and insular cells in the rat pancreas

Close contacts between endocrine insular cells and exocrine acinar, centroacinar and ductular cells occur frequently in the rat pancreas as seen by both light and electron microscopy. Islets of Langerhans are surrounded incompletely by a thin connective tissue capsule or mantle but numerous exocrine-endocrine cell contacts occur at the periphery, which is irregular with considerable "intermingling" of the two cell types. Centroacinar and ductular cells are seen to be in contact with all endocrine cell types but most commonly insulin-secreting B-cells. The basal surface of centroacinar cells in the region of contact may be extensive, sometimes with overlap of basal processes of these cells and their lateral extension between acinar and insular cells. The areas of contact contain no connective tissue or basal lamina and show no surface specializations. The presence of both the "open" and "closed" type of enteroendocrine cells within acini is confirmed, some also being in contact with centroacinar cells. The functional significance of these exo-endocrine cell contacts is discussed in terms of the endocrine-acinar portal system, possible direct paracrine secretion, compartmentalization within the islet, and the known effects of islet hormones on exocrine secretion. Also relevant is the developmental origin of islets from ductal tissue and the cellular origin of some tumours, e.g., insulinomas, from duct cells.     

Electron-dense precipitates in glomus cells of rat carotid body after fixation in glutaraldehyde and pyroantimonate-osmium tetroxide mixture as possible indicators of calcium localization

Summary An attempt was made to study the subcellular localization of calcium in carotid body glomus cells of adult rats using fixation with glutaraldehyde followed by treatment with a mixture of pyroantimonate and osmium tetroxide. Precipitates were seen as electron-dense particles (EDP) in the glomus cells, mostly within membrane-bound organelles, such as dense-cored vesicles, mitochondria, small clear vesicles, multivesicular bodies, and especially in lysosomes. However, EDP were also seen in the nuclei and in the free cytoplasm of the glomus cells and even outside them.Preincubation of carotid bodies in media containing calcium and either high potassium or calcium-ionophore A 23187 resulted in a marked increase in the general precipitation pattern, there being an increased amount of EDP both in the glomus cell nuclei and in the cytoplasm. Dense-cored vesicles more often showed precipitates than those in the controls. Some dense-cored vesicles contained multiple precipitates, typically located in the electron-lucent area between core and vesicle membrane.Extensive diffusion of ions probably occurred during fixation before precipitation, making the localization of calcium and other precipitating cations unreliable. However, it is possible that precipitates, which were regularly seen in the dense-cored vesicles, may reflect the content of bound calcium. The possible significance of calcium in glomus cell function is discussed, and the need for more adequate methods is emphasized.The present study has been supported by grants from the Finska Läkaresällskapet and the Sigrid Jusélius Foundation, Helsinki, FinlandWe wish to express our gratitude to Dr. Robert Hamill of Eli Lilly Co. for kindly providing us with the ionophore A 23187. Technical assistance by Mrs. S. Huhtaniitty and Mrs. T. Stjernvall is gratefully acknowledged     

Distribution of gamma-glutamyl transpeptidase in human pancreas: immunohistochemical study with a monoclonal antibody.

We studied in detail the distribution pattern of gamma-glutamyl transpeptidase (gamma-GT) in human pancreas, using the immunoperoxidase technique and a monoclonal antibody to human kidney gamma-GT. Positive reaction was confined exclusively to the luminal surface of the centroacinar cells and the epithelia of the intercalated, intralobular, and interlobular ducts. The immunoreaction was stronger in the intercalated and intralobular ducts than in the interlobular ducts. The acinar cells did not show any reaction. The islets of Langerhans were heavily surrounded by the ducts, and normal islet cells showed no reaction. The course of the ducts, from the acinar lumina to the interlobular ducts, was delineated by using reaction sites positive for gamma-GT as markers. The courses of the ducts, which comprise the pathway for pancreatic juice, were found to vary widely in their connections with each other, especially between the intralobular and interlobular ducts. At least three separate routes could be identified.