In the exocrine pancreas, the relationship between structure and function, as well as between normal and pathological functioning, can be easily understood if presented in a systematic and logical manner. In this chapter, we explain pancreas physiology. We start by explaining the embryological and ontogenetic development of the pancreas and describe the basic anatomical characteristics of the mature gland, i.e. the macro- and microscopic structure, its vascular supply and innervation. These form the foundation necessary to understand the mechanisms of acinar and ductal cell secretion and their regulation, which are covered in the middle part, with an emphasis on the ionic part of the pancreatic juice. In the last part, we focus on the enzymatic part of the pancreatic juice and its role in digestion of all main groups of energy-rich nutrients, i.e. carbohydrates, proteins and lipids. Two main sources of additional information will help the reader grasp the main concepts in pancreas physiology: figures summarize and combine various concepts encountered in the main text, and clinical boxes contain examples of how a given piece of knowledge can be relevant to understand some diseases.
The main aim of this review of pancreas physiology is to facilitate the understanding of other chapters of this book. It is divided into three main sections that deal with the development and the functional anatomy of the pancreas, with the two-compartment model of exocrine pancreas and the regulation of exocrine secretion and with the role pancreas plays in intestinal digestion of nutrients. Together, these topics shall provide a solid ground to understand etiopathophysiology of the most common pancreatic diseases, their symptoms and crucial clinical characteristics, as well as some key diagnostic and therapeutic principles.
2. Functional anatomy of the pancreas
This chapter is a brief review of human pancreas development and anatomy, with a special emphasis on the exocrine pancreas from both a physiological and a clinical point of view. In other words, this chapter presents developmental and structural basis to understanding pancreas physiology, its blood and lymphatic vasculature, innervation and the integrative regulation of its function, as well as the clinical symptoms and patterns of spreading in cases of malignancy.
2.1. Embryological and ontogenetic development of the pancreas
All parenchymal cell types of the pancreas (acinar, ductal and endocrine cells) are derived from primitive endodermal cells of the foregut [1, 2]. In humans, between the 26th and 28th day of gestation, two endodermal diverticula evaginate from the duodenum, thus forming the dorsal and the ventral pancreatic anlage [3–6]. The dorsal pancreatic bud lies in the dorsal mesentery opposite and above the liver bud. The ventral pancreatic bud develops in the ventral mesentery below the liver bud and connects with the bile duct. During further development, both the ventral bud and the bile duct rotate clockwise, as viewed in the craniocaudal direction, until they reach the dorsal pancreatic bud. Parenchyma of the two buds merges during the 7th week of gestation. The ventral pancreas gives rise to the ventral or lower part of the head of the pancreas that involves also the processus uncinatus, whereas the dorsal pancreas gives rise to the rest of the future pancreas, i.e. the dorsal or upper part of the head, the neck, the body and the tail . Together with parenchyma, the ducts of the primitive pancreas also merge. Ducts of the ventral pancreas and the proximal part of the dorsal pancreas give rise to the main pancreatic duct (of Wirsung). The distal part of the duct of the dorsal pancreas may either obliterate or give rise to the accessory pancreatic duct (of Santorini). In the latter case, the accessory duct drains into the duodenum in the smaller papilla of Santorini that is located orally relative to the larger papilla of Vater .
During endoscopic retrograde cholangiopancreatography (ERCP), in approximately 3% of people the so-called anomalous pancreaticobiliary junction (APBJ) can be found. In this variation, the pancreatic duct joins the bile duct a few centimetres proximally from the duodenal wall. Due to a reflux and stasis of a mixture of bile and pancreatic juice in the bile duct and gallbladder, the incidence of gallbladder and bile duct carcinoma is increased in these people [8, 9]. In addition to APBJ, a number of other conditions result from defects in the embryological development of the pancreas, such as the annular pancreas and pancreas divisum, that are reviewed elsewhere . The dual embryological origin of the pancreas also reflects in the smaller size and a tighter arrangement of the lobules in the ventral pancreas (lower head and the uncinate process), as well as a different cellular make-up of islets of Langerhans and vascular supply (see below) .
In newborns, the total weight of pancreas is around 3 g and the volume of the exocrine pancreas increases approximately linearly to 20 years of age [11–13]. During the period of 20–60 years of age, the volume remains stable and then decreases beyond 60 years of age  (Figure 1).
2.2. Macro- and microscopic anatomy of the pancreas
The human pancreas is a large solitary retroperitoneal organ with well-defined outer borders located at the level of the L1 and L2 vertebrae. The gland is 14–18 cm long, 2–9 cm wide and 2–3 cm thick, weighing 50–125 g [11, 14–17]. It is surrounded by a fibrous capsule from which connective tissue septa extend into the gland dividing its parenchyma into distinct lobes and lobules. In contrast to the outer borders, there are no clear-cut macroscopic borders between the major parts in which the pancreas is usually divided for descriptive purposes: the head, the body and the tail. Generally, the left border of the superior mesenteric vein (SMV) is regarded as the border between the C-shaped head aligned with the upper duodenum on the right and the body located underneath the stomach and extending roughly horizontally in the medial plane on the left. The mid-point of the body and tail combined is then arbitrarily defined as the border between the body and the tail, with the tail usually ranging 1.5–3.5 cm in length [14, 17, 18]. Some authors define a fourth and a fifth part, the inferomedial uncinate process that lies beneath the SMV and the superior mesenteric artery (SMA), and the isthmus or neck, which is an approximately 2 cm wide part of the pancreas situated anterior to the SMA and the point where the SMV and the splenic vein (SV) join to form the portal vein [14–17].
Together with the mesenchyma, the exocrine part of the parenchyma amounts to 96–99% of the total pancreas volume (TPV) [14–16]. Each lobe contains several smaller lobes called lobules. In humans, the lobules are 1–10 mm in diameter . The borders between adjacent lobules are incomplete and thus the whole parenchyma is a continuous unit . Each morphologically recognizable lobule is also a single-functional glandular unit draining into a single duct. In turn, each lobule is supplied by 2–9 arterioles, thus each glandular lobule comprises a few so-called vascular or primary lobules, each of which, by definition, receives a single artery . The remaining 1–4% of TPV contributes the endocrine parenchyma in the form of approximately a million islets of Langerhans, each of which measures around 100 μm and contains approximately a thousand endocrine cells of at least five different types .
From a pathophysiological point of view, as a basic microcirculatory unit the primary lobules resemble the liver units of Rappaport, in that different types of ischemic injuries involve different parts of primary lobules. In more proximal obstruction of a pancreatic artery (due to vasoconstriction in shock, for instance), the most peripheral parts of the primary lobule undergo necrosis, whereas in more distal obstruction (due to blockage of a terminal arteriole in malignant hypertension for instance) the most central parts of a lobule undergo necrosis [19, 21].
Finally, each lobule is composed of acini that are dome-shaped clusters of pyramid-shaped acinar cells. Exocrine secretions from apical poles of acinar cells flow into the lumen of the so-called intercalate duct. Intercalated ducts drain into intra-lobular ducts, these in turn into larger inter-lobular ducts and these finally converge into the main pancreatic duct. The main pancreatic duct empties into the duodenum together with the common bile duct. The end parts of both ducts constitute the so-called hepatopancreatic ampulla (of Vater). The ampulla communicates with the duodenal lumen via the major duodenal papilla (of Vater). The pancreas may have one accessory duct (of Santorini) that leads into the duodenum independently from the main duct and about 2 cm ventroproximally to it [8, 20, 22, 23]. Smooth muscle fibres in the wall of the distal part of the common bile duct, the main pancreatic duct and the papilla form a sphincter (of Oddi) , whether or not the smooth muscle fibres in the wall of the distal accessory duct form a functional sphincter remains a matter of debate .
Impaction of a gall stone in the ampulla is a specific cause of pancreatitis. Somewhat complimentary to the situation in APBJ, Opie proposed that the impaction creates a common channel between the pancreatic and the common bile duct and that the entry of bile into the pancreatic excretory system triggers the inflammation in pancreatitis .
2.3. Vascular supply of the pancreas
The regional blood flow to the pancreas approximates 1% of the cardiac output, 90% of which is directed to the exocrine part . The arterial supply is derived from the celiac artery and the SMA [15, 27–30]. The neck, body and the tail of the pancreas (i.e. the major part of the dorsal pancreas) are irrigated by pancreatic branches of the splenic artery (SA) and by the dorsal pancreatic artery (DPA) that branches off near the origin of celiac, hepatic or splenic artery. DPA separates into two main branches: the right branch anastomoses with the anterior superior pancreaticoduodenal artery (PDA, see below) and the left branch gives rise to the transverse pancreatic artery (TPA, also termed the inferior pancreatic artery). TPA runs at the inferior border of the body and tail, usually anastomosing with the pancreatica magna artery, which is the largest pancreatic branch of the splenic artery [30, 31]. The head and the uncinate process are supplied by an anterior and a posterior arcade [32–35]. The anterior arcade is formed by the anterior superior PDA, and the posterior arcade is formed by the posterior superior PDA . The anterior and posterior superior PDA anastomoses with the anterior and posterior inferior PDA, respectively, both stem from the SMA [34, 35]. The uncinate process and the lower head of the pancreas (i.e. the ventral pancreas) are thus supplied by the SMA.
The venous drainage is anatomically less constant and roughly follows the arterial pattern. The splenic vein collects blood from the neck, the body and the tail via multiple small braches [17, 29]. The blood from the head of the pancreas is drained via two arcades. The anterior venous arcade is formed by the anterior superior and inferior pancreaticoduodenal veins (PDV) draining into the superior mesenteric vein. The posterior arcade consists of the posterior superior and inferior PDV. The posterior inferior PDV drains blood into the superior mesenteric vein, whereas the posterior superior PDV drains directly into the portal vein [15, 28, 29]. A number of anastomoses connect the veins and are typically more irregular than arterial anastomoses .
The smallest intra-lobular vessels are collectively termed the microvasculature of the pancreas . A physiologically important relationship exists between the endocrine and exocrine tissue at the level of the microvasculature. In the human pancreas, the majority of islets of Langerhans are situated within exocrine lobules and the islet capillaries lead blood to a second capillary network surrounding acini. This arrangement of the two capillary networks in series is named the insulo-acinar portal system and forms an important basis for endocrine influences upon the exocrine pancreas [37–41]. The venous blood from inter-lobular islets flows directly into the inter-lobular veins and this type of flow is named the insulo-venous system. Noteworthy, from both the insulo-venous and -acinar system, the venous blood is ultimately passed to the portal vein [27, 42] (Figure 2).
The lymphatic system of the pancreas is usually divided into an internal and an external system . The former has been described to some extent only in rodents and is reviewed in detail elsewhere [14, 43]. In brief, the internal lymphatic system arises in the form of blind-beginning intra-lobular vessels distributed in intra-lobular septa, close by smallest blood vessels and ducts, but at a certain distance from acinar cells, with every lobule possessing many such vessels [43, 44]. Intra-lobular vessels drain into inter-lobular vessels running close by inter-lobular blood vessels and ducts in inter-lobular septa. The largest inter-lobular vessels, also called collecting vessels, reach the surface of the gland and drain into the external system .
An insufficient removal of extracellular fluid and pancreatic enzymes by the lymphatic overflow system from the interstitium may play an etiological role in pancreatitis. The interstitium and the lymphatic vessels are involved in the inflammatory damage and fibrosis, further hampering the lymphatic drainage and initiating a vicious cycle [43, 44].
The external system consists of large surface lymphatic vessels and regional lymph nodes. Due to the clinical importance of the external system, especially in carcinoma, it has been studied extensively in humans [17, 43, 45–50]. The external lymphatic vessels can be grouped into roughly seven different groups, each of which is associated with a corresponding group of blood vessels. The superior vessels close by the splenic artery and the inferior vessels close by the TPA drain the tail and the left part of the body. The anterosuperior, posterosuperior, anteroinferior and posteroinferior pancreaticoduodenal vessels (close by the arteries of the same name), as well as the gastroduodenal vessels, drain the head of the pancreas and the right part of the body. In general, authors also agree on the anatomical position of the lymph nodes to which the aforementioned vessels drain and on which nodes are most commonly affected in carcinoma of different parts of the pancreas. In contrast, there is much confusion with regard to the nomenclature of the nodes, with a descriptive [17, 46, 51] and a numerical system [49, 52]. In brief, the main groups (with their notation according to the numerical system in parentheses) that collect lymph from the tail and the body are the splenic and gastrosplenic nodes that lie within and superior to the splenic hilum (10), as well as the suprapancreatic (11) and infrapancreatic (18) nodes that lie close by the splenic and inferior pancreatic artery, respectively. The main groups that collect lymph from the head are the hepatic (8) and hepatoduodenal (12), as well as the superior anterior (17a), superior posterior (13a), inferior anterior (17b) and inferior posterior (13b) pancreaticoduodenal nodes. In addition to these nodes that encircle the pancreas, the paraaortic (16), celiac (9), superior mesenteric (14) and the middle colic nodes (15) lie close by the abdominal aorta and its trunks [43, 53]. The nodes tightly surrounding the pancreas and the nodes around the aorta probably do not correspond to first and second barriers of spread, respectively, since they both receive lymph directly from the pancreas as well as from other nodes . Nodes indexed by numbers 1–7 probably do not drain the pancreas . Noteworthy, the centrifugal path from the aforementioned nodes is via cisterna chyli and the thoracic duct .
Lymph node involvement is associated with a poor prognosis and is present in approximately four out of five patients with pancreatic cancer [17, 53]. The largely asymptomatic nature of cancer growth, with jaundice, duodenal obstruction and pain as the most common symptoms appearing late in the course of the disease, probably contributes to the fact that the tumours are detected at an advanced stage. Due to the numerous anastomoses between lymphatic vessels and the fact that obstruction of lymphatic vessels brought about by cancer growth and spread may further alter the already unpredictable routes of drainage, it is extremely difficult to exactly predict the spreading pattern of pancreatic cancer [17, 43, 49]. Tumours originating in the tail and the body most frequently spread to nodes 8, 11, 16 and 18 and only nodes 17 have not been involved in any of the cases [53, 54]. Tumours from the head of the pancreas most frequently spread to nodes 13, 17, 14 and 16, with only nodes 10 and 15 being spared in all cases [49, 50, 53]. It seems that the dual embryological origin of pancreas also influences the spreading pattern of cancer of the head. Tumours from the lower (ventral) head spread to nodes around the SMA (14), in contrast tumours from the upper (dorsal) head spread to nodes around the common hepatic artery (8) and in the hepatoduodenal ligament (12), which is in accordance with the arterial supply (see above) .
2.4. Innervation of the pancreas
The pancreas is innervated by sympathetic, parasympathetic and afferent nerve fibres that enter and exit the pancreas together with vessels and follow them also within the pancreatic tissue [36, 56–59]. The somata of preganglionic sympathetic neurons innervating the pancreas reside in the lateral horn of the C8-L3 spinal cord segments and project to paravertebral sympathetic ganglia. Alternatively, some axons do not terminate at synapses within the paravertebral ganglia but continue within splanchnic nerves to synapse within the celiac ganglia and the superior mesenteric ganglion [36, 56, 57]. The tail and the body of the pancreas are supplied by nerve fibres that originate in the celiac plexus and follow the splenic artery and TPA . The majority of nerve fibres to the pancreas supply the head . They originate in the anterior and posterior hepatic plexus. The fibres that enter the uncinate process originate in the superior mesenteric ganglion .
As already mentioned, lymph node involvement is one of the most important prognostic factors in pancreaticobiliary tract carcinomas. In general, lymph node metastasis is established by lymphatic invasion; however, tumour cells were shown to be able to spread into the hilum of lymph nodes via neural invasion. The knowledge of patterns of neural architecture may improve curative procedures . Moreover, embryological development of the pancreas served as a useful template for patterns of extrapancreatic nerve plexus invasion of pancreatic head carcinoma .
The efferent autonomous nerves in the pancreas have release sites that are not in close contact with cells and thus probably influence many targets at a time [58, 64]. In the exocrine pancreas, the sympathetic terminals contact predominantly the intra-pancreatic ganglia, blood vessels and ducts. Stimulation of sympathetic fibres indirectly inhibits the exocrine secretion by inhibiting intra-pancreatic ganglia and by decreasing supply of fluid via vasoconstriction .
The somata of the parasympathetic preganglionic neurons reside in the dorsal motor nucleus of vagus and the nucleus ambiguus [36, 56]. The majority of their axons join the vagus and some the splanchnic nerves and reach the neural plexuses around arteries where they intermingle with sympathetic fibres . The preganglionic parasympathetic neurons finally reach intra-pancreatic ganglia together with vessels supplying them [36, 56]. The parasympathetic ganglia that reside within the inter-lobular septa, lobules and also close to islets receive input not only from parasympathetic preganglionic fibres, but also from other pancreatic ganglia, sympathetic fibres (see above), the myenteric plexus, as well as the sensory fibres (see below) . Postganglionic fibres innervate acinar and ductal epithelial cells, ductal smooth muscle cells and vascular plexuses, as well as other ganglia. These fibres mediate parasympathetic stimulation of secretion from acinar and ductal cells, constriction of ducts, as well as an increase in fluid supply by vasodilation [36, 61].
In the pancreas, sympathetic and parasympathetic afferent fibres can also be found. They contain substance P (SP) or calcitonin gene-related product (CGRP) as neurotransmitters. Sympathetic afferents that innervate both the exocrine and the endocrine tissue join the sympathetic splanchnic nerves and transmit noci- and mechano-receptive sensory information to somata within the dorsal root ganglia and further on to preganglionic sympathetic neurons in the lateral horn of the spinal medulla and probably higher centres .
Pancreatic sympathetic innervation is altered in chronic pancreatitis and pancreatic cancer and may contribute to the neuropathic pain and visceral neuropathy in these states [65, 66]. Dorsal root ganglion sympathetic afferent neurons send collaterals to efferent ganglia, representing a neuroanatomical substrate for intrapancreatic monosynaptic vegetative reflexes. For example, SP and CGRP released at intra-pancreatic ganglia inhibit exocrine secretion. Intra-pancreatic ganglia are also contacted by vagal afferents .
Somata of vagal afferent neurons reside within the nodose ganglia. They innervate the blood vessels, ducts, acini and islets. However, their centripetal pathways are not well known .
3. Integrative physiology of the exocrine pancreas
In this chapter we elucidate the mechanisms by which the exocrine pancreas secretes pancreatic ductal fluid and digestive enzymes under physiological conditions, the regulatory mechanisms that govern its function, and describe the response of pancreatic secretion to a meal. Moreover, this chapter offers some insight into the pathophysiological background of pancreatic diseases related to exocrine pancreas secretion.
3.1. Composition of pancreatic fluid
In humans, the secretion of a neutral, isotonic, Na+, Cl− and H+-rich fluid, active digestive proteins, as well as zymogens by the pancreatic acinar cells, and of an alkaline, isotonic and HCO3−-rich fluid by the pancreatic ductal cell yields between 1 and 2.5 L of pancreatic fluid per day, which contains around 20 g of digestive enzymes [67–69]. More than 20 different enzymes are secreted by the acinar cells , and some of them are precursor enzymes, such as trypsinogen and chymotrypsinogen. The enzymes released from the acinar cells in an active form are lipases, colipases, A-amylases, collagenases, elastases, ribonucleases and phospholipases A [70, 71].
Human pancreatic fluid contains up to 150 mmol/L of HCO3−. Its concentration increases with pancreatic fluid flow rate, and reaches its peak at 30–50% of maximal flow [72, 73]. The Cl− concentration relates inversely with the pancreatic fluid flow rate and maintains an isotonic osmolality with respect to HCO3−. The composition of cations remains fairly constant, irrespective of pancreatic fluid flow rate, with 140 mmol/L Na+, and 10–15 mmol/L K+. The sum of HCO3− and Cl− concentration closely matches the sum of Na+ and K+ concentration. Electrolytes, such as Ca2+, Mg2+, Zn2+, PO43− and SO42−, are also present, but at minimal concentrations [74–76].
3.2. Regulation of acinar cell secretion
Secretion of digestive enzymes from acinar cells is primarily mediated by acetylcholine (ACh) release from vagal nerve endings and by the intestinal hormone, cholecystokinin (CCK). In addition to its primary role in ductal secretion (see below), the hormone secretin also influences acinar cell function, as does the vasoactive intestinal peptide (VIP) [77, 78].
The pancreas is innervated by postganglionic nerves, which receive input from the preganglionic motor neurons that stem from the dorsal motor nucleus of the vagus (DMV) . ACh mediates its effects on acinar cell secretion via M1 and M3 muscarinic receptors , with M3 muscarinic receptors playing a predominant role . The pancreas is also innervated by sensory vagal afferents, which project to the solitary nucleus, where information is integrated and relayed to the preganglionic motor neurons of the DMV, and the two together constitute the so-called dorsal vagal complex . M1 and M3 muscarinic receptors are linked to the Gq/11 family of G-proteins and cause hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP2) by phospholipase C, yielding inositol 1,4,5,-trisphosphate (1,4,5-IP3) and 1,2-diacylglycerol (DAG) [82–84]. DAG goes on to phosphorylate various proteins via protein kinase C activation, while IP3 mobilizes Ca2+ from internal stores to stimulate amylase secretion [82, 85].
Digestive elements of fats and proteins, such as fatty acids with acyl chains longer than 12 carbon atoms, and amino acids, phenylalanine and tryptophan, are prime secretagogues for CCK secretion by the intestinal I cells. Carbohydrates, on the other hand, do not exhibit particular potency. The mechanisms of CCK action on pancreatic acinar cells seem to exhibit species-specificity . It was thought that the main, if not exclusive, influence of CCK on human acinar cells was mediated via interaction with cholinergic nerves by presynaptic modulation of vagal output . It has been recently shown, however, that CCK can activate human pancreatic acinar cell secretion both directly  and indirectly through CCK-A receptors on vagal afferents [77, 79]. Mechanisms of CCK-relayed digestive enzyme secretion have not yet been fully clarified . The CCK-B subtype of CCK receptors seems to be the predominant form in the human pancreas, while the presence of the CCK-A subtype has been difficult to demonstrate . Similar to the M1 and M3 muscarinic receptors, CCK-B receptors are coupled to the Gq/11 family of G-proteins, and follow a similar signalling pathway to elevate intracellular calcium .
An alternative pathway of pancreatic acinar fluid and protein secretion is mediated by secretin and vasoactive intestinal peptide (VIP). By stimulating their respective Gα G-protein-coupled receptors an increase in cAMP is observed, which in turn increases PKA activity, leading to secretion [78, 89].
3.3. Molecular mechanisms of acinar cell secretion
In response to stimulation with ACh and CCK, acinar cells secrete an isotonic, plasma-like, protein-rich fluid containing Na+, Cl− and H+, which is later modified by the ductal cells to form the final fluid [69, 70, 74]. Both secretagogues stimulate mechanisms that cause Ca2+ oscillations in the cytosol of acinar cells, which is a signalling mechanism for both fluid and enzyme secretions [90–92].
The fluid secretion of acinar cells is a result of ion transport across the basolateral and apical membranes, as well as paracellular transport mechanisms [90–92]. The driving force for acinar cell fluid secretion stems from the Na+/K+-ATPase located on the basolateral membrane and from the trans-cellular ion gradient it creates. The Na+/K+/2Cl− co-transporter NKCC1 on the basal side is responsible for approximately 70% of the Cl− uptake for subsequent secretion by the pancreatic gland. The Ca2+- and voltage-activated K+ maxi-K channel on the basolateral membrane and others set the acinar cell membrane potential close to the K+ diffusion potential. The membrane potential, in turn, serves as the electromotive force for Cl− exit at the apical membrane. Along with the above-mentioned basolateral transport proteins, the Na+/H+ exchanger NHE1 and the AE2 isoform of the Cl−/HCO3− exchanger family also play a role in basolateral Cl- uptake. Both NKCC1 and NHE1 also provide Na+ for the Na+/K+-ATPase and regulate intracellular pH, keeping it at about pH = 7.2 [90, 93]. Luminal secretion of Cl− occurs via a voltage- and Ca2+-activated Cl− channel TMEM16A/Ano1. As Cl− flows through the cell into the lumen of the acinus, Na+ follows via the paracellular pathway. The subsequent osmotically driven water flow is mediated by aquaporin AQP1 [70, 74, 93].
Digestive enzymes are stored in zymogene granules at the apical membrane of acinar cells and are released by way of exocytosis. Fusion of granules with the apical plasma membrane releases their contents into the acinar lumen and later on into the small intestine . As with fluid and electrolyte secretion from pancreatic acinar cells, Ca2+ ions are the key messenger in triggering and controlling a series of events termed stimulus-secretion coupling, i.e. pathways that regulate digestive enzyme secretion from acinar cells. Upon stimulation with secretagogues, a spike in intracellular Ca2+, released from intracellular Ca2+ stores, causes fusion of zymogen granules with the plasma membrane [94, 95]. Physiological stimulants can evoke various intracellular Ca2+ patterns: (i) global Ca2+ oscillations, (ii) Ca2+ waves that flow across the cell and (iii) local calcium spikes . Local apical Ca2+ spikes, which occur with lower levels of stimulation, as well as global Ca2+ spikes will increase the permeability of Ca2+-dependent Cl− channels, resulting in fluid secretion. It seems, however, that physiological stimulation yields zymogene granule fusion only when a global Ca2+ spike is observed .
In acute pancreatitis, a condition, which most frequently occurs due to alcohol abuse and biliary disease, the pro-enzymes stored in acinar cells become activated prematurely, causing autodigestion with inflammation and necrosis of the pancreatic tissue. Under normal conditions, intracellular Ca2+ is a key secondary messenger in pancreatic acinar cell secretion. Recently, however, a body of evidence suggests Ca2+ is a key initiator of pancreatitis. Noxious stimuli, such as alcohol, long-chain fatty acids and bile acids, provoke extensive Ca2+ release from intracellular stores, causing a prolonged and global Ca2+ elevation. This kind of abnormal calcium signalling in turn activates trypsinogen that causes pancreatic autodigestion [97–99].
3.4. Regulation of ductal cell secretion
Secretory control of pancreatic ductal cells exhibits great complexity as it involves a variety of receptors on both the basolateral and apical membranes. Activation of these receptors can be a stimulatory or an inhibitory factor in HCO3− and fluid secretion .
The most important secretagogue for HCO3− secretion from pancreatic ductal cells is the peptide hormone secretin . The primary stimulus for secretin release from the neuroendocrine S cells in the proximal duodenum is intra-duodenal pH below 2–4.5, which occurs upon entry of acidic chyme from the stomach. Fatty acids and bile salts are also stimuli for secretin release [78, 100–102]. Upon activation of the secretin receptor on the basolateral side, which is coupled to adenylyl cyclase, increase and accumulation of cAMP are observed. cAMP activates PKA, which in turn phosphorylates the CF transmembrane conductance regulator (CFTR) in the apical membrane of ductal cells  and the basolateral Na+-HCO3− co-transporter NBCe1-B . Possible alternative routes of cAMP/PKA pathway activation are the release of VIP from vagal nerve terminals, with subsequent VIP receptor VPAC1 activation and beta-adrenergic receptor activation . Vagal nerve fibres release VIP together with the main neurotransmitter ACh. Ductal cells express M2 and M3 muscarinic receptors located on the basolateral membrane . Stimulation with ACh and CCK causes an increase in intracellular Ca2+ concentration by stimulating G-protein coupled receptors that activate the phospholipase C pathway, which activates the calcium-activating chloride channels and possibly also the apical Cl−/HCO3− exchanger, triggering ductal secretion [67, 105]. The effect of cholecystokinin (CCK) in humans is that of a potentiator of secretin effects on ductal HCO3− and fluid output. The enhancing effects of secretin most likely occur by stimulation of vagal afferent fibres . This indicates a synergistic relationship between the Ca2+ and cAMP pathways . Ductal cells also express several types of purinergic receptors and intra-luminal application of ATP and UTP results in enhanced HCO3− secretion [73, 106]. Luminal ATP causes stimulation of HCO3− and pancreatic fluid which quantitatively approaches 75% of maximal secretin stimulation. In contrast, on the basolateral side, ATP inhibits both spontaneous and secretin-evoked secretion by as much as 50% .
Substance P, 5-HT, AVP, and the afore-mentioned basolateral ATP fall in the category of potential inhibitory factors in pancreatic HCO3− secretion. Although molecular mechanisms of inhibition are not yet fully understood, their role is most likely curtailment of luminal hydrostatic pressure, which precludes enzyme leakage into the pancreatic parenchyma and discontinuation of secretion after a meal .
3.5. Molecular mechanisms of ductal cell secretion
One of the earlier models of ductal cell HCO3− secretion proposed by Ashton, Argent and Green presupposes intracellular generation of HCO3− from CO2, and hydration by carbonic anhydrase. The dissociated proton is transported by a Na+/H+ exchanger, located on the basolateral membrane, and the HCO3− ions are transported into the lumen by a HCO3−/Cl− exchanger that is driven by the luminal Cl− availability. The luminal Cl− gradient is maintained by a cAMP-activated Cl− channel, regulated by secretin. Since the exit of HCO3− in this model is electrogenic, it is accompanied by outflow of K+ ions through the cAMP activated maxi-K channels. This model, while providing an explanation for much of what is observed in pancreatic duct cell secretion in many species, is however, limited to a maximum luminal HCO3− concentration of about 70 mmol/L. Human pancreatic duct cells, on the other hand, create a luminal HCO3− concentration as high as 140 mmol/L and above [67, 68, 108, 109].
In attempts to bring the mechanisms that account for such a high HCO3− secretion to light, several models have been proposed [110, 111]. As new information about the identity and properties of ion transporters and channels as well as cellular mechanisms of their action were discovered, a revised two-step model as described below has been suggested.
In the proximal duct, HCO3− is actively transported and accumulated in the cytosol of ductal cells by a 1Na+-2HCO3− co-transporter NBCe1-B on the basolateral membrane, which is driven by the Na+ gradient. Secretion of HCO3− on the luminal side occurs by way of a 1Cl−/2HCO3− exchanger SLC26A6, while the CFTR provides a recycling path for Cl− ions. HCO3− secretion drives the translocation of Na+ ions to the lumen by a paracellular pathway. These two processes create a driving force for water efflux by AQP1 . The proximal duct absorbs a part of the Cl− and secretes up to 100 mmol/L of HCO3− and provides much of the aqueous part of the pancreatic fluid. By the time the fluid reaches the distal segments of the duct, the lumen is Cl− depleted to approximately 30 mmol/L and due to the active CFTR, the intracellular Cl− concentration drops to 10 mmol/L or less. Low concentration of Cl− activates the WNK1-OSR1/SPAK pathway, which results in two events. First, the permeability of CFTR changes in favour of HCO3−, making CFTR a route for HCO3− secretion. Second, the SLC26A6 is inhibited, which favours HCO3− accumulation in the lumen since its active state would most likely result in reabsorption, not secretion as in the proximal lumen [67, 74, 93, 112, 113] (Figure 3).
A mutation in the CFTR encoding gene that codes for the chloride and bicarbonate channel involved in pancreatic ductal cell fluid secretion results in a disease called cystic fibrosis, where anomalous fluid secretion results in dysfunction in several organ systems such as the lung, gastrointestinal tract, liver, male reproductive tract and pancreas. Reduced or even absent CFTR function causes a change in ductal fluid composition – decreased pH and fluid volume, and hyper-concentration of fluid components – that is thought to lead to obstruction. As the disease progresses, acinus plugging and dilation provoke epithelial injury and destruction, with inflammation, calcium deposits and fibrosis. These pathological processes lead to indigestion and malnutrition [114, 115].
3.6. Meal-response of pancreatic secretion and the inter-digestive phase
The basal pancreatic exocrine secretion rate reaches approximately 20% of the maximum capacity for enzyme secretion in humans. This basal secretion could be explained by an intrinsic characteristic of the pancreas, stimulation of the gland by low levels of CCK or secretin, or by ACh release . Inhibition of ACh and CCK input reduces pancreatic enzyme secretion by about 50% .
The basal pancreatic duct HCO3− secretion amounts to only 1–2% in comparison with secretion stimulated with exogenous secretin. Since secretin is the primary regulator of HCO3− secretion, basal HCO3− mainly parallels plasma secretin levels, along with cholinergic input .
In the long term, this is, however, not the full extent of regulation. Inter-digestive pancreatic exocrine function is cyclically coupled with fasting motility phases termed the inter-digestive migratory motor complex (MMC), although they do follow different trends for overall daytime and night time secretory and motor activity [117, 118]. Upon ingestion of a meal, this behaviour is interrupted within minutes. Postprandial enzyme secretion reaches peak levels within the first hour and decreases followed by a stable phase of secretion, only to return to inter-digestive levels in 3–4 h .
The meal response of pancreatic secretion can be divided into four phases, i.e. cephalic, gastric, intestinal and inter-digestive phases. The names correspond to the origin of the predominant form of pancreatic secretion control. There is however significant overlap and integration between these regulatory mechanisms in response to a meal. Average pancreatic secretion stimulated by a meal amounts to about 50–60% of the maximum output of the gland [120–122].
The cephalic phase contributes approximately 25% of the meal-response secretion. It is caused by input of visual, olfactory, gustatory and tactile (mastication) sensory modalities. Sham feeding (a process, in which people are allowed to see, smell, taste but not swallow the food) causes pancreatic secretion that is rich in enzymes but low in concentration and volume of HCO3− . It amounts to a response that is up to 50% of maximum secretory capacity. The main mediator of the cephalic phase is the vagus nerve with ACh as the dominant neurotransmitter .
The gastric phase accounts for about 10% of the meal-response secretion and starts with arrival of food into the stomach. It is regulated by enteropancreatic, vagovagal reflexes, which are stimulated by gastric distension .
The majority of the meal-response pancreatic secretion occurs in the intestinal phase. The secretory output amounts to 50–60% of maximal capacity . It starts as the acidic chyme from the stomach, passes into the intestine, where components of the chyme, HCl, bile and bile salts elicit hormonal and neural responses. Hormonal influence on pancreatic secretion in the intestinal phase is mainly the result of CCK and secretin and the amplification of the secretory response by neural influences of the enteropancreatic reflex . Passing of stomach contents into the intestine lowers the luminal pH, which is a signal for the duodenal S cells to release secretin. It, in turn, functions as the main secretagogue for HCO3− secretion from the ductal cells and also as an enhancer of enzyme secretion from the acinar cells [77, 102].
4. Role of the exocrine pancreas in digestion
Pancreas plays a crucial role in digestion. Its exocrine part secretes enzymes that are involved in digestion of carbohydrates, proteins and lipids. In this section, we will briefly review the digestive processes in general, and specifically point out the contribution of the pancreatic juice.
4.1. Assimilation of carbohydrates
In the western diet, the mean daily intake of carbohydrates is about 300 g that yield about 1200 kcal in metabolism . Starch, the plant storage polysaccharide, constitutes by far the largest percentage of the carbohydrate intake (70%). About 30% of starch is composed of amylose (a straight polymer of glucose), the remainder of amylopectin (a branched polymer of glucose) [125, 126]. These different constituents of starch require different enzymes for their cleavage. Further 20% of carbohydrates in the food contribute refined sugars (e.g. sucrose, fructose and glucose) and approximately 10% the disaccharide lactose from various sources .
Digestion of carbohydrates takes place in two subsequent steps that are separated spatially: (i) digestion in the lumen and (ii) at the enterocyte brush border. Acinar cells of salivary glands (parotid, sublingual and submandibular) and of exocrine pancreas all produce and secrete the closely related enzyme α-amylase [128, 129]. α-amylase is secreted in an active form and has an optimal pH for enzymatic activity at pH = 7.0 . It is an endo-enzyme that cleaves internal α-1,4 glycosidic bonds, but not α-1,6 bonds, terminal α-1,4 bonds, or α-1,4 bonds that are next to the branches in the molecular structure . Amylopectin is thus cleaved to maltose and α-limit dextrins, whereas amylose is cleaved to maltose and maltotriose. The reactions seem to yield also a small percentage of free glucose .
The digestion of carbohydrates starts with intra-luminal digestion in the oral cavity. This appears not to have an important physiological role, since the salivary α-amylases are mostly inactivated in the acid milieu of the gastric lumen . The activity of salivary α-amylases is partly rescued by occupying the active site of the enzyme with the substrate . The digestion continues in the duodenum with the activity of pancreatic α-amylase, which has a rather neutral pH optimum, brought about by alkalinization of acidic chyme from the stomach by duodenal bicarbonate secretion. Contribution of the salivary amylase in starch degradation remains controversial and may become quantitatively more important when mastication time is prolonged [129, 133].
Following the intra-luminal digestion, membrane-bound enzymes of the enterocyte brush border degrade oligosaccharides produced by the luminal digestion to monosaccharides that are then absorbed into enterocytes . The apical membrane of enterocyte brush border contains four major enzymes that act on the luminal side: (i) lactase, (ii) glucoamylase (maltase), (iii) isomaltase and (iv) sucrase . The latter two are located on the same polypeptide chain with two distinctive active sites and are often referred to as the sucrase-isomaltase complex. An essential enzyme for starch digestion is the isomaltase, which is the only enzyme capable of degrading the α-1,6 bond, whereas the other (sucrase, glucoamylase and lactase) are involved in degrading internal α-1,4 bonds. The disaccharides sucrose and lactose are digested by the sucrase and the lactase. The final products of luminal and brush border digestion are the monosaccharides glucose, galactose and fructose, which are then absorbed through the apical membrane via SGLT-1 (glucose and galactose) and GLUT5 (fructose) transporters and through the basolateral membrane via GLUT2 [136–138]. Additionally, GLUT5 transporter in the basolateral membrane may serve as an alternative exit route for fructose (Figure 4).
4.2. Assimilation of proteins
Intake of proteins in the western diet amounts to approximately 70–100 g/day, which accounts for 300–400 kcal . In contrast to the carbohydrates, protein digestion starts in the stomach, as virtually no significant proteolytic enzymes are found in saliva. The chief (zymogenic, peptic) cells of the gastric glands synthesize and secrete the pro-enzyme pepsinogen, the inactive precursor of pepsin, which is a proteolytic enzyme specifically suited to act in the acidic gastric milieu . At pH < 5 in the gastric lumen, pepsinogen is spontaneously converted into the active form, pepsin, by cleavage of an N-terminal peptide. At pH values >5.0 and >7.5 pepsin inactivates reversibly and irreversibly, respectively . Pepsin has its pH optimum at pH = 1.5–2.5. Pepsin functions as an endopeptidase, yielding oligopeptides and amino acids [141, 142].
Bulk proteolysis occurs in the small intestine. The pancreatic acinar cells secrete into the duodenum five major proteolytic proenzymes: trypsin, chymotrypsin, elastase and carboxypeptidase A and B [143–145]. First, the proenzyme trypsinogen is activated by the membrane-bound enterokinase to its active form trypsin . The specific expression of enterokinase serves to limit proteolysis to the lumen of the small intestine. Trypsin, in turn, activates additional trypsin molecules and also converts the other four proenzymes, i.e. chymotrypsinogen, proelastase, as well as procarboxypeptidase A and B to their active forms. Carboxypeptidase A and B are ectopeptidases and cleave amino acids at the C-terminus, whereas trypsin, chymotrypsin, and elastase are endopeptidases that cleave polypeptides at specific sites resulting in 2–6 amino acid oligopeptides . The oligopeptides from the lumen are further cleaved by both brush border-bound and intracellular cytosolic peptidases .
The absorption of oligopeptides and free amino acids differs. The oligopeptides are absorbed through the apical membrane with the PepT1, a H+/oligopeptide co-transporter driven by an H+ gradient generated by the NHE3 (Na/H exchanger type 3) [147, 148]. The amino acid absorption at the apical and basolateral membrane involves at least seven and five different transport systems, respectively; however, a detailed description of these systems is beyond the scope of this chapter [149, 150] (Figure 5).
4.3. Assimilation of lipids
In the western diet, 60–100 g of lipids are ingested daily, which account for about 500–900 kcal . Most of the ingested lipids are in the form of triacylglycerol (TAG, 90–95%), the rest are in the form of phospholipids (PL, 5%) and cholesterol (C, < 0,5%).
Enzymes for digestion of lipids act in the watery environment of intestinal lumen, and due to their hydrophilic character, they act on the surface of ingested lipids organized in amphiphilic droplets. In the process of emulsification, larger droplets are broken into smaller ones, thereby increasing surface to volume ratio, and, consequently, enzyme efficacy . Emulsification starts with food preparation, and continues in the mouth with mastication and in the gastric and intestinal lumen with churning of the ingested food. Lipid droplets organize such that the core is composed of hydrophobic TAG, whereas amphiphilic PL, C and free fatty acids (FFA) are on the surface [152, 153]. This organization, together with some proteins and carbohydrates stabilizes the products of emulsification.
Digestion of lipids exhibits a large functional redundancy . It starts in the gastric lumen catalysed by the lingual and gastric lipase, together termed pre-duodenal lipase [155, 156]. The former is synthesized by acinar cells of the salivary glands and the latter by gastric chief cells. pH optimum for pre-duodenal lipase is around pH = 4, quite appropriate for the acidic gastric milieu [155, 157]. The enzyme cleaves the first ester bond in the TAG producing diacylglycerol (DAG) and an FFA [158, 159]. It is resistant to cleavage by pepsin but not by pancreatic proteases , and therefore functions mostly in the gastric lumen where it digests up to 15% of ingested lipids in a healthy individual. However, they might contribute to some extent to digestion of lipids in the duodenum [151, 157, 160, 161].
Digestion of lipids continues in the intestinal lumen by three major enzymes secreted by pancreatic acinar cells: the (i) pancreatic, (ii) nonspecific (carboxylic ester) lipase and (iii) phospholipase A2. The pancreatic lipase is active in the presence of a colipase (which is in turn activated by trypsin), in an alkaline pH, in the presence of Ca2+ and bile salts [162, 163]. The pancreatic lipase cleaves the first and the third ester bond in TAGs yielding 2-monoacylglycerol (MAG) and FFAs [159, 164]. Phospholipase A2 acts on the second ester bond in glycerophospholipids, giving rise to lysophospholipid (LPL) and FFAs . It is secreted in an inactive form and requires for its activity an alkaline pH and bile salts. The specificity of the nonspecific lipase is low, notably, and it is able to cleave cholesterol esters and MAG . The quantitative contribution of the nonspecific lipase is relatively low compared with the pancreatic lipase. During infancy, the pancreatic lipase related protein 2 contributes to digestion of lipids .
The luminal enzyme activity outlined above, in conjunction with an ever-ongoing emulsification, results in a multi-lamellar envelope developing around the droplet, and consisting of FFAs, bile acids, C, MAG and LPLs [166–168]. The multi-lamellar envelope bursts from the droplets in the form of vesicles that are eventually transformed to mixed micelles (especially under the influence of bile salts) that then finally serve as the main vehicle for absorption of lipids. FFAs enter the enterocyte by (i) collision with the plasmalemma and crossing of the plasmalemma by the flip-flop mechanism, or (ii) by diffusion of non-ionic fatty acids or (iii) by carrier-mediated transport. The latter probably involves the fatty acid binding protein (FABP), the fatty acid transporter protein type 4 (FATP4) and CD36 [169, 170]. MAG, LPLs and C probably enter the enterocyte by means of transport proteins or by simple diffusion. Finally, the FFAs, MAG, LPLs and C are re-esterified within the enterocyte, together with apolipoproteins assembled into chylomicrons, exocytosed into the extracellular space on the basolateral side of the plasma membrane, and reach the systemic circulation via the lymphatic circulation. In contrast, monosaccharides and amino acids reach the systemic circulation via the portal vein [151, 171] (Figure 6).
Exocrine pancreatic insufficiency (EPI) due to chronic pancreatitis or cystic fibrosis will eventually result in digestive malfunctioning . However, because of the redundancy of the digestive processes described above, the effect is noticed rather late in the course of the disease . In fact, malabsorption will not present itself until the exocrine pancreas function falls to <10% . Among all the ingested nutrients, assimilation of lipids is most dependent on normal pancreas function. This seems unexpected due to a large redundancy of the lipid digestion, in fact the pre-duodenal lipase output may even increase in the EPI  and the pre-duodenal lipases can rescue up to 80% of fat digestion [176, 177]. However, as the destruction of the pancreas progresses, the ductal cells fail to neutralize acidic gastric juice leading to intra-intestinal acidification resulting in bile salt precipitation . Bile salts are necessary for the mixed micelle formation, and this strongly hampers assimilation of lipids. Protein and carbohydrate digestion may have greater digestion potential during EPI due to the fact that the digestion is also initiated independently of pancreas, and continued by the brush-border peptidases and oligosaccharidases that are pancreas-independent. An analogue of the brush-border enzymes is missing in the digestion of lipids; therefore, it is not surprising that in EPI, lipid malabsorption is the most overwhelming problem causing many of the clinical symptoms and signs, leading to weight loss, steatorrhea, abdominal discomfort and a deficit in the lipid-soluble vitamins (A, D, E, K) [178, 179]. The digestive function can largely be rescued and nutrient malabsorption ameliorated by pancreatic enzyme replacement therapy (PERT), a therapy that involves oral administration of enzyme mixtures consisting of lipases, amylases and proteases [179–181].
Despite the ever-growing reductionist record on cell biology of constitutive parts of pancreas (exocrine, ductal and endocrine), whole organ integrative physiology detailing organ development, blood and lymphatic vascularization, innervation and direct links to pancreatic role in intestinal digestion, is rarely part of a single review. The present chapter goes even further and gives examples where this knowledge can provide a basis to understand etiopathophysiology of most common pancreatic diseases, including malignancy. We created a series of images to merge the different cell biological, anatomical and physiological layers to explain modern pancreas function and possible causes for dysfunction.