The pancreatic lymphatic system has been poorly studied compared with other organs where lymphatics play additional important roles such as absorption of lipids in the small intestine (Azzali, 1982; Dobbins, 1966; Dobbins and Rollins, 1970; Othani, 1987). Present knowledge of the finer distribution of lymphatics within the pancreas is very limited due to the insufficient number of studies addressed (O'Morchoe, 1997). Nevertheless, the definition of the exact relationships of the lymphatic network with the parenchymal components of the pancreas has to be considered a step that cannot be ignored to outline the physiopathologic involvement of lymphatics in some diseases. The understanding of the lymphatic distribution within the gland might help to define the role that lymphatics could play in some pancreatic pathologies by removing the overflowed pancreatic enzymes (Bockman, 1992; Godart, 1965), or by draining pancreatic hormones (i.e., insulinoma, hyperinsulinemic hypoglycemia). Moreover, the definition of the extent of the lymphatic network within the gland might help to explain the fast lymphatic spreading of pancreatic cancer to the lymph nodes.
The distribution of the lymph vessels inside the pancreas is a subject that has been considered controversial for many years. Lymphatics have been reported as originating from the interlobular connective septa (Laguesse, 1914; Rénji-Vamos, 1955; Rouvière, 1932; Tömböl and Vajda, 1962), from the intralobular connective septa (Bastianini and Fonzi, 1978a,b; Sokolowsky, 1970) or even from the fine interacinar space (Rusznyàk et al., 1960; Zhemchuzhnikova, 1959). More recently, the lymphatic system of the pancreas has been studied in greater detail by transmission electron microscopy in guinea pig (Bertelli et al., 1993; Ji and Kato, 1997) and rat pancreas (Navas et al., 1995) with different results. Because the extension of the intrapancreatic lymphatic network and the relationships of the lymph vessels with the parenchymal structures of the pancreas are subjects that still need to be defined, we decided to undertake a comparative study in rodents to verify eventual differences among species belonging to the same order of animals and to outline the role that the lymphatic system might play in some pancreatic pathologies.
MATERIALS AND METHODS
Three guinea pigs (200–300 g), three white Wistar rats (150–200 g), and three C57 mice (25–30 g) were anesthetized with ether and subsequently sacrificed with the same anesthetic. Pieces of pancreas were fixed in 1% glutaraldehyde or 4% paraformaldehyde, postfixed in 1% OsO4, dehydrated and embedded in Epon 812. To improve the fixation, some animals were previously perfused via the thoracic aorta with 2.5% glutaraldehyde for 30 min. The specimens were postfixed in 1% OsO4, dehydrated and embedded in Epon 812. Consecutive semithin sections (0.5–0.7 μm), stained with 0.1% Toluidine Blue, were observed by a light microscope Zeiss Axioplan and the presumptive lymph vessels were positively identified by transmission electron microscopy (Philips 201) on consecutive ultrathin sections stained with uranyl acetate and lead citrate. After the electron microscopic identification the course of the lymphatics was followed with long series of semithin consecutive sections to define exactly their distribution. On the whole more than 500 semithin sections were performed for each specie of animal.
Lymphatics were easily detected in guinea pig and mouse pancreas. In the rat pancreas, however, they appeared as a rarer finding. At any rate, lymph vessels were observed in the interlobular and intralobular connective septa within the pancreas of all the animals (Fig. 1). Nevertheless, along their course, the association with the different parenchymal components of the gland varied according to the animals.
Relations With the Acinar Tissue
In guinea pig pancreas, the lymphatic network could be spotted even in very small intralobular connective septa (Fig. 1a) from where they also extended long processes amongst the nearest acini (Fig. 1b). Acinar cells and lymphatics were found as close as less than 0.5 μm (Fig. 1b). In mouse and rat pancreas, the relationships were not as close as in guinea pig and lymphatics appeared running within larger intralobular septa at a certain distance from the acinar cells (Fig. 1c). In rat pancreas, pancreatic lobules were provided with just one lymph vessel that was centrally located (Figs. 1d, 3a,b) and appeared to be preferentially satellite of major blood vessels (Fig. 1d).
Relations With the Ductal System
Whereas in the rat pancreas lymph vessels were in relation just with the major pancreatic ducts, in mouse and guinea pig the relationships with ducts of minor caliber were frequent and, at least for guinea pig, very close. In the latter case, in fact, we could find lymphatics near small intralobular ducts (Fig. 1a). Remarkably, close and very extended relationships were detected with the interlobular ducts (Fig. 2). In these cases, lymphatics could run for long tracts in the context of the very thin mantle of fibroblasts surrounding the ducts (Fig. 2b).
Relations With the Endocrine Pancreas
Lymphatics were never observed within the islets of Langerhans but they were found close to the periphery of clusters of endocrine cells in all the species. In rat pancreas, however, lymphatics were rarely observed in proximity to endocrine tissue. Nevertheless, we could detect intralobular lymph vessels approaching small clusters of endocrine cells sited within the connective septa (Fig. 3a,b). On the contrary, guinea pig pancreas (Fig. 3c,d) and mouse pancreas (Fig. 3e,f) showed frequent close relations between lymphatics and islets of Langerhans (<0.3 μm). This was due to the peculiar topography of many islets of Langerhans that could be found embedded within the connective interlobular septa or just facing the stromal septa; in such cases, lymphatics closely associated with the periphery of the islets could be detected (Fig. 3c–f). At any rate, lymphatics running in the neighborhood of the islets (within 15–20 μm) were constantly observed in all the animals that underwent the present investigation (Fig. 1c).
The involvement of lymphatics in many pathologic processes, such as inflammation and cancer, is well-known (O'Morchoe, 1997). Their role, however, may acquire peculiar aspects according to the organ. In the case of the pancreas, for example, lymphatics can be engaged in the removal of large amounts of enzymes released into the interstitium as occurring in severe pancreatitis. In this respect, the definition of the lymphatic distribution within the gland is an unquestionable preliminary step to frame the role of lymphatics in pancreatic pathology.
The lymphatic system of the pancreas has been recently studied in guinea pig and rat pancreas (Bertelli et al., 1993; Ji and Kato, 1997; Navas et al., 1995): in guinea pig, the pancreatic lymphatic network has been observed developing in the intralobular connective septa; remarkably, close relations between absorbing lymph vessels and endocrine tissue were observed (Bertelli et al., 1993). On the contrary, in rat pancreas, the lymphatic network has been reported to be much less developed and no relations with the pancreatic endocrine tissue have been observed so far (Navas et al., 1995). Aim of this study has been to precise the relationships of the lymph vessels with the parenchymal structures of the pancreas comparing three animals belonging to the same order and avoiding to describe once more the ultrastructural features of the lymphatics that are known in detail since the 1960s (Leak and Burke, 1966; Leak, 1976) and that have been reported to be roughly the same in the pancreas as well (Bertelli et al., 1993; Navas et al., 1995). Regardless of these considerations, we summarize the guidelines we followed to positively identify lymphatics. We referred to a vessel as an initial lymphatic when it was devoid of pericytes and was formed by a simple endothelium that lacked a continuous basement membrane and that lined a close and well-defined lumen. We did not consider as a vessel all the structures surrounding a lumen provided with bundles of collagen inside (Aharinejad et al., 1995) but, indeed, we never observed any vessel-like structure like that.
Following such criteria we have clearly shown that the pattern of distribution of the lymphatics is almost the same in all the species of animals examined in this study. Summarizing, we can affirm that lymphatics form a network of vessels in the interlobular areas and that lymph vessels can be found running in some intralobular connective septa as well. Moreover, the lymphatic system can approach the pancreatic endocrine tissue in all the animals. Despite such a general accord in the pattern of distribution, minor differences exist among rat, guinea pig and mouse pancreatic lymphatic systems: the most evident of all is the degree of development that is more striking in guinea pig and mouse. On the whole, lymphatics are a quite common finding in the pancreas of these species where several intralobular connective septa of each lobule are provided with a lymph vessel. In contrast, in rat pancreas, lymphatics are more difficult to be detected, as previously reported (Navas et al., 1995), because they are confined to the major intralobular connective septum that crosses through lobule. In addition, unlike the rat, guinea pig and mouse intrapancreatic lymphatic networks are largely related with the ductal system and the intralobular lymph vessels are in close relation with the acini adjacent the connective septa in guinea pig pancreas. At any rate, pancreas appears to be a gland provided with a rich network of lymphatic vessels in all the species examined. This finding should be considered important for the understanding of the mechanisms leading to the lymphatic spreading of pancreatic cancer.
The relationships between lymphatics and the exocrine parenchyma, either ducts or adenomeres, as showed in guinea pig pancreas, are important in view of the possible role of “safety valve” that lymph vessels could play in pancreatitis. Such hypothesis was first suggested by Dumont et al. (1960) who observed an increase of the levels of lipase and amylase in the lymph of the thoracic duct after contemporaneous injection of morphine, which acts by closing the sphincter of Oddi, and secretin, which is a note secretagogue for the pancreas. Since then, many evidences have been accumulated that confirm how lymphatics represent a pathway for enzymes overflowed in the interstitial tissue in experimental pancreatitis. Godart (1965) showed that tracers can be found in the lymph vessels only 15 min after their injection in the duct system, and the levels of pancreatic enzymes in the lymph of the thoracic duct increase dramatically within the first few hours from the induction of the pancreatitis (Sim et al., 1966; Lange et al., 1986). In such conditions the lymphatic system seems to be able to drain, together with the enzymes, also fragments of cells and erythrocytes (Bockman et al., 1973). Taking together, these data suggest that lymphatics, draining away large amounts of enzymes in the first few hours of pancreatitis, can really play a role in preserving the integrity of the gland, and our observations are consistent with this view especially for guinea pig and mouse pancreas. On the other hand, lymph vessels are early damaged by pancreatitis. Subcapsular lymphatic network, in fact, disappears after acute pancreatitis and are no more detectable in chronic pancreatitis (Reynolds, 1970). Furthermore, the early obstruction of the lymphatics can make worse the pancreatic inflammation not only because they do not fulfill anymore their protective function but also because this condition is able by itself alone to cause a severe pancreatitis as previously showed (Müller et al., 1988). In human, the thick and compact nature of the pancreas should even benefit more from an extensive lymphatic system than the thin, diffuse pancreas of rodents and the impairment of lymphatic drainage could, therefore, represent a predisposition to the recurrent bouts of inflammation typical of chronic pancreatitis (O'Morchoe, 1997).
Lymphatic close relations (<1μm) with the endocrine tissue are more frequent in guinea pig and mouse pancreas, whereas in rat pancreas, even though existing, they can be considered occasional findings. Lymphatics running in the neighborhood of the endocrine tissue (within 15μm) are, however, common in all animals and this is a datum that deserves to be underlined because it is reasonable to hypothesize a lymphatic drainage of minor amounts of hormones. The influence that the islet hormones are able to exert on the peri-insular acini via a paracrine way (Bendayan, 1985; Bendayan and Gregoire, 1987; Owyang, 1993), in fact, shows that they are probably not completely captured by the intrainsular blood vessels and that they can diffuse for a certain distance. In this respect, as previously suggested (Bonner-Weir, 1993), lymph vessels draining interstitial fluid could work to create a gradient of islet hormones across the acinar tissue. This gradient, decreasing from islets toward lymphatics would extend, therefore, beyond islet efferent capillary network (Bonner-Weir, 1993). Moreover, insulin has been previously detected in the lymph of the thoracic duct of rabbit (Daniel and Henderson, 1966), rat (Rasio et al., 1965), dog (Pepin et al., 1970), and human (Rasio et al., 1967) and, even though we cannot exclude other sources of lymphatic insulin such as extrapancreatic islets of Langerhans in the duodenum (Bendayan and Park, 1991, 1997) where lymphatics are more abundant, our observations supply a morphologic evidence of consistent relationships between lymphatics and pancreatic endocrine tissue. Such relations had been previously reported just in guinea pig pancreas (Bertelli et al., 1993). The role of the “lymphatic insulin” should be particularly considered because, even though representing a minor fraction of the entire secretion of insulin, it skips the “first pass” effect of the liver that is able to extract from the blood more than 50% of the insulin during a single hepatic passage (Rojdmark et al., 1978). Thus, even though in physiologic conditions the amount of this fraction is considered little, accounting presumably less then 1% of the total insulin secretion (Henderson, 1974), we can not rule out that alterations of this ratio may occur in pathologic situations (i.e., insulinoma, hyperinsulinemic hypoglycemia). In conclusion, we demonstrated the existence of extensive relationships between lymphatics and all the parenchymal components of the pancreas of all the species studied. Close relations with excretory ducts and islets of Langerhans, however, were more pronounced in guinea pig and mouse pancreas. Such relations should be particularly considered in the progression of pancreatic diseases.
This work received a grant from the University of Siena consisting of intramural funds to A.B.