**4. Functional role of NIC**

Since the description of NIC, researchers have been interested in questions about the functional role of NIC and the mechanisms of their formation. These two interrelated problems remain unresolved today. The idea of a regulatory role of the nervous system in endocrine secretion is commonly accepted now [8]. The pancreas is innervated by sympathetic and parasympathetic nerve fibers [8, 39]. Moreover, in the pancreatic islets of humans and rodents, there are the afferent (sensory) nerve fibers [7, 40, 41]. Many studies have demonstrated a role for the nervous system in the regulation and synchronization of hormone secretion from endocrine cells [7, 8, 17–19, 42, 43]. Stimulation of sympathetic nerves increases the release of glucagon and reduces the release of insulin and somatostatin [10, 41, 44]. Parasympathetic stimulation increases the release of insulin, glucagon, somatostatin and pancreatic polypeptides in various species [7, 8, 10, 44, 45]. Sensory nerves are also involved in the regulation of hormone secretion by pancreatic endocrine cells. The chemical destruction of sensory nerves (capsaicin treatment) in mice increases insulin secretion in response to glucose, compared to control. Consequently, sensory fibers may exert an inhibitory effect on insulin secretion [46].

Thus, the various forms of NIC that we observed in the human fetal pancreas are similar in their morphological organization to the NIC, which were found in the fetal and adult pancreas

The amount of NIC gradually decreases at birth. In the pancreas of children and adults, NIC are less abundant than in the fetal pancreas [22]. Our quantitative data indicate that the largest number of NIC I was observed in the early and middle fetal periods, during the active morphogenesis of pancreatic islets, whereas at birth (in the late fetal period) and in the adult, NIC II became more prevalent [35]. It should also be noted that NIC I and NIC II in which a single or few endocrine cells were located inside ganglia or in nerve bundles were found only in the fetal pancreas. We did not find these types of NIC in the adult pancreas, probably due to an insufficient number of fields of observation. Therefore, we could not exclude that these types of NIC can be present in the adult pancreas, but they are rare. NIC I in which pancreatic islets were associated with ganglia were more numerous in the fetal pancreas and were occasionally found in the pancreas of children [22] and adults [35]. Among the NIC II, at all investigated stages of development, as well as in children and adults, interactions between thin nerve fibers and endocrine cells located separately or inside the islets prevailed [35].

To identify whether glial (Schwann) cells cover the periphery of islets in humans, as in other mammals, we used immunohistochemical labeling with antibodies against S100 protein and GFAP as well as electron microscopy. We found small S100-positive cells with thin processes at the periphery of some islets in humans [37, 38]. The same small oval, triangular or elongated cells with long thin processes were observed in the fetal pancreas using electron microscopy [38]. The processes of these cells were often cover or surround nerve fibers passing into islets [38]. In contrast to mice and rats [3, 13], these cells were immunonegative to GFAP. However, according to their ultrastructural characteristics and integration with nerve fibers, these small S100-positive cells with thin processes that we detected in the human pancreas correspond to the glial (Schwann) cells observed at the periphery of islets in other mammals [3, 13, 14, 34]. It should be noted that, in humans, S100-positive glial cells are present only in some islets in small numbers and their processes do not cover endocrine cells, as has been described in other

Taken together, our findings indicate that, in the human pancreas, NIC are more abundant and variable in their morphological organization in the prenatal period, i.e. during the active morphogenesis of pancreatic islets. Based on these findings, we suggest that the nervous system may be involved in the development of the human endocrine pancreas. In the next part of this chapter, we discuss the existing points of view on the possible functional role of NIC.

Since the description of NIC, researchers have been interested in questions about the functional role of NIC and the mechanisms of their formation. These two interrelated problems remain unresolved today. The idea of a regulatory role of the nervous system in endocrine secretion is commonly accepted now [8]. The pancreas is innervated by sympathetic and parasympa-

of other mammals [26–29].

8 Challenges in Pancreatic Pathology

mammals [3, 13, 14, 32, 34].

**4. Functional role of NIC**

Both the parasympathetic and sympathetic nervous systems impact the postnatal development of the endocrine pancreas and pancreatic plasticity in adult animals [17, 43]. For example, a decrease in the proliferation of β-cells has been detected in mice and rats after vagotomy [47, 48].

However, the concept of the regulatory role of the nervous system in the control of hormone secretion and endocrine cell proliferation does not explain the presence of endocrine cells inside ganglia or in nerve bundles. Simard [25] proposed that these endocrine cells may secrete hormones directly into nervous tissue. However, histological and cytological analysis performed by Fujita [26] has shown that endocrine cells in the NIC are similar in their mode of secretion to endocrine cells located in pancreatic islets, because their secretory granules accumulate on the side of the cell facing the capillaries.

In the second half of the twentieth century, there were two widespread concepts: APUD (amine precursor uptake and decarboxylation) [49, 50] and "paraneuron" [51]. It is well known that endocrine cells of the pancreas and neural cells express many common proteins, such as S100, glutamic acid decarboxylase (GAD), NSE, NPY and so on [3, 52–54], and have similarities in their developmental control mechanisms (for review, see [55, 56]). Similarities between endocrine cells and neurons are also confirmed by phylogenetic data. Endocrine cells (insulin-, glucagon-, somatostatin- and PP-secreting) were found in the brain in some invertebrates and lower vertebrates [57]. In the "APUD" and "paraneuron" concepts, these similarities were explained by the common embryonic origin of pancreatic endocrine cells and neurons from the neuroectoderm [49–51]. It has also been proposed that pancreatic islets can be regarded as modified ganglia because of the gradation between pure ganglia, mixed forms representing NIC, and pure islets [29]. This hypothesis was disproved in a series of classic experiments with quail-chick chimeras and in cell culture studies in which the endodermal origin of endocrine cells was established [58–60].

Today, it is well known that pancreatic endocrine cells differentiate from epithelial progenitors. In human fetal pancreas, epithelial ductal cells express numerous transcriptional factors that regulate endocrine cell differentiation [61–63], and differentiating endocrine cells transiently retain epithelial markers and are often associated with the ductal epithelium [61, 62]. The structures of the nervous system originate from the neuroectoderm [64]. The data concerning the mechanisms of the formation of NIC are very limited. Studies on rodents (mice and rats) have demonstrated that the innervation of islets develops in the early postnatal period [31, 65]. In other mammals (cats, dogs and rabbits) [26] including humans, NIC have been detected in the pancreas during prenatal development. The morphogenetic mechanisms underlying the integration between structures of the nervous system and endocrine cells remain unclear.

In his work, Van Campenhout [24] found that all primary islets (Laguesse islets) form NIC; he first suggested that the nervous system may be involved in the development of the endocrine pancreas. He proposed that NIC form through the budding of islet cells from the primitive ducts followed by their migration into adjacent neural tissue and that the differentiation of islet cells may occur under the influence of nervous components of these complexes (cited from [26]). Consequently, it was shown that both NIC and non-innervated islets can be detected in the fetal pancreas. In our studies, we also found NIC and non-innervated islets in the human fetal and adult pancreas [22, 35].

Nonetheless, genetic studies on mice have confirmed that the nervous system may regulate the differentiation of endocrine cells. In mice deficient for Phox2b or Foxd3 gene expression (Phox2b−/− or Foxd3−/−), neural crest cells and their derivatives are absent in the pancreas. It was shown that, in such mutant embryos, the total β-cell mass and β-cell proliferation had increased [66]. Furthermore, β-cells in the mutant embryos were immature, since the expression levels of MafA and Pdx1 mRNA in β-cells were decreased, and insulin granules had abnormal morphology and were decreased in number [67]. Taken together, these data demonstrate that signals from the neural crest negatively regulate β-cell proliferation and positively regulate β-cell maturation [66, 67]. Moreover, in the developing mouse pancreas, neural crest cells and their derivatives are located in close proximity to endocrine β- and αcells, suggesting that the regulation of β-cell mass and their maturation may occur through juxtacrine and paracrine signals from the nervous system [67, 68]. Similar results, demonstrating that co-culturing pancreatic islets with neural crest stem cells promotes the regeneration of functional β-cells, were observed in vitro [69].

A recent study has also shown an important role of sympathetic innervation in the establishment of pancreatic islet architecture and functional maturation during development. The absence of sympathetic innervation during development resulted in altered islet architecture, reduced insulin secretion and impaired glucose tolerance in adult mice [70].

Several studies have demonstrated that the structures of the nervous system interact with endocrine cells through the homophilic binding of cell adhesion molecules (NCAM and SynCAM), which are expressed on the surface of pancreatic endocrine cells and neural crestderived cells in both rodents [68, 71] and humans [72]. Contacts between forming islets and NCAM-positive nerve fibers have been observed in the human fetal pancreas [72]. Therefore, it has been proposed that autonomic nerves may facilitate the outpouching of endocrine cell clusters to form islets through the homophilic binding of CD56 (NCAM) molecules on both of these tissues [72].

In our studies, we have demonstrated close integration between the structures of the nervous system and endocrine cells in the human pancreas, which was more frequently observed during prenatal development. We suggest that such integration may be necessary for the development of the endocrine pancreas in humans [22, 35]. It is possible that the nervous system may regulate endocrine cell mass and their maturation, as has been shown in mice. It is also possible that the contacts with structures of the nervous system may be necessary for the migration of epithelial progenitors into forming islets. In this case, different types of NIC may represent various stages of pancreatic islet morphogenesis.
