The Neuropeptide Pituitary Adenylate Cyclase–Activating Polypeptide and Islet Function

The synthesis and secretion of adenohypophysial hormones are regulated by various neuropeptides with different specificity and mechanisms. In 1989, Miyata et al. (13) isolated PACAP, a novel peptide from ovine hypothalamus with a hitherto unmet ability to increase the pituitary content of cAMP. PACAP was originally discovered as a COOH-terminally amidated peptide of 38 amino acids (13), PACAP38 (molecular weight 4.5 kDa), but was later found to also exist in a COOH-terminally truncated 27–amino acid long-form equivalent to PACAP38(1–27) and thus called PACAP27 (molecular weight 3.0 kDa) (14). In addition, PACAP27 is amidated at its COOH-terminal end. In all tissues examined, PACAP38 is the predominant form of PACAP (15,16). The peptide is structurally related to VIP and is therefore a member of the glucagon/VIP family of peptides comprising secretin, helodermin, helospectin, and GLP-1. In fact, PACAP27 displays 68% identity with the full length of VIP (Fig. 1) (14). The amino acid sequence of PACAP is identical in all mammals, and in species such as the chicken, frog, salmon, and tunicate, only 1–3 amino acids are substituted, suggesting that PACAP is highly conserved and has remained almost unchanged during an evolutionary period of ∼700 million years (17).

The human gene for PACAP is located on chromosome 18p11 and consists of five exons (Fig. 2) (18). Translation of the gene yields a 176–amino acid long precursor form, including a signal peptide. A proregion of the native PACAP is followed by a Gly-Gly-Arg sequence for proteolytic processing and amidation preceding the sequence of PACAP38 (amino acids 132–169). The PACAP precursor contains two sets of dibasic cleavage and amidation sites that are located before the structures of PACAP38 and PACAP27, respectively, i.e., also one within the structure of PACAP38. This suggests that the formation of the two PACAP forms is the result of alternative splicing at the respective amidation sites, which seems to be a unique biosynthetic feature for neuropeptides. Within exon 4 of the PACAP gene, there is a sequence encoding for another protein, called PACAP-related peptide (PRP) (18). PRP is 29 amino acids long (amino acids 82–110 of the precursor protein), but its structure is not as well preserved between species as that of PACAP. This makes the PACAP gene similar to the VIP gene, in which the coding of peptide histidine methionine corresponds to the coding of PRP in the PACAP gene (19). Although PRP is expressed concomitantly with PACAP, the function of this peptide is unknown.

PACAP is localized to the central and peripheral nervous systems. In the rat brain, the highest concentration of PACAP is found in the hypothalamus (20), whereas in the human brain, the highest concentration of PACAP is found in the dorsal vagal complex, the bed nucleus of the stria terminals, the median eminence-pituitary stalk, and the periventricular and paraventricular hypothalamic nuclei (21). There is a specific and saturable uptake mechanism of PACAP into the brain through the blood-brain barrier, called peptide transport system-6 (22); thus, it is possible that PACAP may also affect brain function through this route. PACAP has been localized to nerve fibers and terminals in the superficial layer of the dorsal horn and cell bodies in the dorsal root ganglia in rats (23). In these nerve fibers, PACAP is partly colocalized with calcitonin gene-related peptide (CGRP) and substance P, suggesting that PACAP is a sensory neurotransmitter. This is also supported by findings that the tissue abundancy of PACAP is markedly reduced by the sensory neurotoxin capsaicin in several peripheral organs (23,24). In the peripheral nervous system, PACAP-containing nerve fibers have been demonstrated in the adrenals (23), the respiratory tract (25), the urinary tract (24), the gastrointestinal tract (24), and blood vessel walls in several organs (26). In many of these tissues, PACAP is colocalized with VIP (26,27). In addition, PACAP has also been localized to nitric oxide synthase–containing nerves (26). Together these results show that PACAP is a constituent of several types of nerves throughout the organism, with the potential of exerting a complex role in physiology.

It has been shown that the circulating level of PACAP is, as it is for other neuropeptides, very low and most likely <10 pmol/l (28). However, it is difficult to accurately measure circulating PACAP because PACAP38, but not PACAP27, is tightly bound to ceruloplasmin (29). Therefore, until these technical obstacles have been overcome, the importance of circulating PACAP is unknown. To estimate the circulating turnover rate of PACAP in humans, we have compared the steady-state level of PACAP during a continuous intravenous infusion of PACAP27 in relation to the infusion rate and found that the metabolic clearance rate of the peptide is ∼70 ml · kg⁻¹ · min⁻¹ (28). This value is close to that of the entire cardiac output and corresponds to a circulating half-life of <1 min, suggesting an avid and rapid degradation system for circulating PACAP, which is expected from a locally active neuropeptide.

Three PACAP receptors have been cloned (30,31,32,33,34,35)—PAC1-R, VPAC₁-R, and VPAC₂-R (36)—and they all consist of seven transmembrane regions and are coupled to G-proteins. Several transduction mechanisms can be activated upon binding of PACAP and VIP, and so far, adenylate cyclase (AC), phospholipase C (PLC), and calcium have been identified as effector pathways (37,38). However, different ligands have been shown to be variably effective in activating these signaling pathways, depending on which receptor has been activated (Table 1). VPAC₁-R seems to activate AC and PLC signaling pathways, whereas ligand binding to VPAC₂-R predominantly leads to stimulation of AC (26). PAC1-R activation has been found to induce the above-mentioned signaling transduction mechanisms to various degrees. This can be explained by the finding that several different splice variants with different signaling properties can be generated from the PAC₁-R gene (17). Six of the variants are generated by alternative splicing of the exon encoding the third intracellular loop of the receptor, as illustrated in Fig. 3. Two different cassettes of 28 (hip or hop1) or 27 amino acids (hop2), are inserted singularly, together (yielding five different variants), or not inserted (yielding the short form, PAC₁-Rs). In addition, two other variants are generated by the deletion of a 21–amino acid residue in the amino-terminal extracellular domain of the receptor (the very short form, PAC₁-vs) or by the deletion and substitution of two amino acids in the fourth transmembrane domain (the PAC₁-R-TM variant) (37,38,39,40,41). These splice variants are differentially expressed in different tissues, and therefore ligand binding can result in the activation of different signaling pathways, depending on the tissue being studied. PAC₁-R-TM4, cloned from rat cerebellum, does not activate either AC or PLC but instead increases intracellular Ca²⁺ by directly opening membrane-bound l-type Ca²⁺ channels upon activation (40). In contrast, no splice variants seem to exist for VPAC₁-R and VPAC₂-R.

As tools for studies on the pharmacology and physiology of PACAP receptors, several modified forms of PACAP have been generated, and their effects have been evaluated (Table 1). To date, the best antagonistic effect is yielded with NH₂-terminally truncated fragments of PACAP, i.e., PACAP(6–38) and PACAP(6–27). These peptide fragments interact with the PAC₁ and VPAC₂ receptors, without agonistic action (42,43,44), whereas VPAC₁-R has no affinity for the shortened PACAP forms. Instead, a specific antagonist for VPAC₁-R is represented by a modified VIP/GRF hybrid peptide (26). No specific VPAC₂-R antagonist has been generated so far. Interestingly, a natural highly specific agonist for PAC₁-R has been isolated from the saliva of the sand fly, Lutzomia lingipalpis. This agonist, maxadilan, is a 61–amino acid peptide with no significant sequence homology to PACAP (45). Furthermore, M65, a modified fragment of maxadilan (with deletion of amino acids 24–41) acts as an antagonist of PAC₁-R, without affecting VPAC receptors (46). Finally, agonists for VPAC₁-R and VPAC₂-R have also been developed (47,48).

PACAP has several effects in the central and peripheral nervous system as well as in peripheral organs. Several recent excellent reviews have compiled complete descriptions of the effects of PACAP (15,17,26,49,50). However, all effects have been documented after adding the peptide in different experimental conditions. Therefore, despite well-documented actions of the peptide in different systems, its physiological role has not yet been established.

Although PACAP was originally isolated for its ability to potently stimulate AC in cultured rat pituitary cells (13), the role of PACAP in the regulation of pituitary function has not yet been established. In general, however, in vivo studies show that PACAP increases hormone release from the anterior pituitary, particularly that of growth hormone and ACTH (17,51).

Besides the action of PACAP on cortisol release caused by stimulated ACTH release, PACAP also directly stimulates synthesis and release of epinephrine, as demonstrated in cultured chromaffin cells, perifused adrenal glands, and adrenal slices in vivo, and this seems to be mediated by PAC₁-R (17,26).

Several actions of PACAP have been demonstrated in regard to gut motility. For example, the peptide induces relaxation of the smooth muscles in the stomach, small intestine, and colon, as demonstrated in the rat (52), and PACAP in particular has been suggested to mediate the nonadrenergic noncholinergic relaxation of longitudinal muscle in rat distal colon (53). Moreover, PACAP stimulates the secretion from a colonocyte cell line (54) and inhibits gastric acid secretion in conscious rats (55).

PACAP exerts direct effects on the liver, as shown by stimulation of glucose output from the perfused rat liver and from cultured rat hepatocytes by an effect mediated by a cAMP-dependent stimulation of glycogenolysis and is also accompanied by an increase in intracellular Ca²⁺ (56,57).

PACAP is localized to nerves innervating blood vessels throughout the body (17,26). Furthermore, specific PACAP binding sites have been demonstrated on membranes isolated from blood vessels (58). PACAP also induces vasodilatation both in vitro and in vivo in different organs (49); for example, in healthy men PACAP potently induces vasodilatation in the brachial artery (59). The effect of PACAP on blood vessels seems to last longer than the effects of VIP or CGRP (6 h for PACAP vs. 2 h for the other peptides), suggesting a slower dissociation from the receptors than for these two other vasoactive neuropeptides (59).

Immunocytochemistry of pancreatic tissue has revealed that PACAP is localized to nerves throughout the gland, within the exocrine parenchyma, around blood vessels, and within the islets in mice (10,60), rats (10,61), and pigs (9). Also, the human pancreas contains PACAP, and as in other tissues, PACAP38 is the predominant form (28). This suggests that PACAP is a pancreatic neuropeptide in nerves in the endocrine and exocrine portion of the gland. Species differences seem to exist regarding the abundancy of the nerves; the mouse pancreas is more abundantly innervated by PACAP nerves than the rat pancreas (10). In the islets, the PACAP nerves are uniformly distributed over the islet area (10,60), suggesting that PACAP is an islet neuropeptide with the potential ability to affect all islet cells. In addition, nerve cell bodies within intrapancreatic ganglia contain PACAP (9,10,61), indicating that PACAP is also involved in the regulation of pancreatic ganglia function.

The nature of intra-islet PACAP nerves has been studied in detail. In the pig and rat pancreas, PACAP has been colocalized with VIP both in the exocrine portion of the gland and in ganglia and islets (9,10), suggesting that PACAP is a parasympathetic neurotransmitter. The parasympathetic nature of pancreatic PACAP nerves is also evident from results showing that vagal electric stimulation releases PACAP from the pig pancreas and that the pancreatic PACAP content is reduced by vagotomy (9). However, a recent study has demonstrated that localization of PACAP occurs not only in nerve terminals of intrinsic VIP-containing nerves but also in extrinsic CGRP-containing nerves in mouse pancreas and that the pancreatic content of PACAP is reduced by neonatal treatment with capsaicin (62). This suggests that PACAP is also a sensory neuropeptide in the pancreas because CGRP is found in capsaicin-sensitive sensory nerves (1). Therefore, pancreatic PACAP seems to be a neurotransmitter both in the parasympathetic nervous system, together with acetylcholine and VIP, and in the sensory nerves, together with CGRP. This is corroborated by studies in other organs, where PACAP is a parasympathetic (63) as well as a sensory (23,24) neurotransmitter.

In addition, a study on freshly isolated rat islets has localized PACAP to islet endocrine cells by using both immunocytochemistry and reverse transcriptase–polymerase chain reaction (RT-PCR) (64). However, this finding has not been confirmed by immunocytochemistry in other studies (9,10,62) and therefore needs to be viewed cautiously.

By in situ hybridization, mRNA for PAC₁-R and VPAC₂-R has been demonstrated in the exocrine and endocrine parenchyma of mouse and rat pancreas as well as in insulin-producing cell lines (10). The islet receptor mRNA was shown to be evenly distributed within the islets, suggesting that receptors are localized to all islet cell types (10). Immunocytochemistry of rat islets has shown that the PAC₁-R is localized to β-cells as well as to rat α-cells (64). By using semiquantative RT-PCR, it has also been shown that VPAC₁-R is expressed in the rat pancreas (65), whereas only VPAC₂-R has been shown to be expressed in the human pancreas (66). Whether there is a true species difference regarding expression of the different PACAP receptors in islets remains to be finally established. Regarding splice variants of PAC₁-R in islets, one study has demonstrated the expression of the PAC1-TM4 in islet cells (40). Therefore, at least in the rodent pancreas, all three PACAP receptors seem to be expressed, and two of them are localized to islet β-cells.

Several in vitro studies in islets, insulin-producing cells, and the perfused pancreas have shown that PACAP potently stimulates insulin secretion in a dose- and glucose-dependent manner and that PACAP27 and PACAP38 are equipotent in these respects (10,67,68,69,70,71,72,73). In isolated rat islets, one study has demonstrated that PACAP is extremely potent at stimulating insulin secretion, i.e., an effect was observed at a concentration as low as 10 fmol/l (61). However, other studies could not reproduce this extraordinary effect in rat islets or clonal cells. For example, the lowest effective concentration level in the insulin-producing cells (HIT-T15 cells) was 1 nmol/l (10,67) and 0.1 nmol/l (68,69), which is comparable with stimulatory concentrations of other neuropeptides. Nevertheless, a tendency of stimulation of insulin release by PACAP at very low concentrations was observed in mouse islets because in ∼50% of the mouse islet incubations studied, PACAP38 elicited an insulin response at the low dose of 10 fmol/l (10). These divergent findings of the sensitivity to PACAP may be explained by facilitating presynaptic PAC₁-R on the islet PACAP-containing nerve terminals that are variably expressed in different experimental systems. Such receptors would explain why PACAP does not affect insulin secretion at very low doses in clonal β-cells, which lack innervation (10), and why PACAP exerts a clear action at very low-dose levels in freshly isolated islets (61), which contain nerve terminals, but not in overnight-cultured islets (73), which may lack islet nerve terminals (74). However, presynaptic PACAP receptors in pancreatic islets have not yet been established.

Several studies have shown that PACAP and VIP are of approximate equal potency at stimulating insulin secretion in vitro (9,10,68,69). This suggests that VPAC₁-R and VPAC₂-R are the receptors carrying out the insulinotropic effect of PACAP in vitro because these receptors bind PACAP and VIP with equal affinity. However, studies in the perfused pancreas from genetically modified mice lacking PAC₁-R have shown that PACAP-induced insulin secretion is diminished by ∼50% (12). This shows that the PAC₁-R is also involved in the mediation of the insulinotropic effect of PACAP.

PACAP also stimulates glucagon secretion, as has been demonstrated in the perfused rat pancreas (71) and in single rat α-cells (75), although this has not yet been examined in great detail.

Although these actions of PACAP are short-term and of acute nature, long-term actions of PACAP in insulin-producing cells should not be overlooked. In fact, a recent study demonstrated that on a long-term basis, PACAP exerts transcriptional regulation of the insulin gene and on the genes of glucose transporter and hexokinase in a clonal cell line, RIN 1,046-38 (76). However, more studies are required for the long-term effects of PACAP on islet function.

PACAP stimulates insulin secretion in vivo in several species of experimental animals, such as calves (77), dogs (78), and mice (60,79). In both calves and mice, PACAP and VIP stimulate insulin secretion equipotently, and in mice it has also been demonstrated that PACAP27 and PACAP38 are equipotent (79). Recently, we also examined the effect of PACAP on islet hormone secretion in healthy human volunteers (28). The highest tolerable infusion rate of PACAP in humans is ∼3.5 pmol · min⁻¹ · kg⁻¹ because higher doses cause intense flushing of the skin caused by PACAP-induced vasodilatory effect (80). Therefore, full agonistic action of exogenously added PACAP cannot be examined in humans because higher infusion rates would then be required. When intravenously infused at 3 pmol · min⁻¹ · kg⁻¹, PACAP27 was well-tolerated in our study, although all subjects experienced flushing of the facial skin and peripheral paleness (28). In some cases, the flushing was of long duration and remained >12 h after termination of the infusion, which corroborates with the slow dissociation of PACAP from its receptors (59). Systolic and diastolic blood pressures decreased by ∼10 mmHg, and pulse rates increased by 15 beats per min, which is in accordance with the results obtained in calves (77). The results show that PACAP slightly increased serum insulin levels in the overnight-fasted healthy volunteers and potentiated the insulin and C-peptide responses to an intravenous glucose challenge (28). This shows that PACAP also stimulates insulin secretion in humans.

The action of PACAP to stimulate insulin secretion in vivo is probably mainly executed through a direct islet effect. However, because PACAP is expressed in the hypothalamus, it is also possible that neuropeptide affects insulin secretion through a central action when administered in vivo because the autonomic nerves originating in hypothalamic nuclei are of importance for islet function (1). However, this aspect of PACAP has not yet been examined and therefore needs further study.

In addition, glucagon secretion in vivo is stimulated by PACAP, as demonstrated in mice (60,79), dogs (78), and humans (28), whereas in calves, PACAP augments acetylcholine-stimulated glucagon secretion without having any direct effect per se (77). Furthermore, administration of glucose inhibits PACAP-induced glucagon secretion (28,77,78,79), showing that the action is glucose sensitive. Finally, PACAP seems to have vasodilatory action in the pancreas, as demonstrated in vivo in rats (81) and in the perfused rat pancreas (72). The contribution of this vascular effect for the effect on islet hormone secretion remains to be established.

As in other tissues, PACAP potently increases cellular content of cAMP in insulin-producing cells, as evident from studies in clonal HIT-T15 cells in which a marked increase in cAMP content is apparent after only a 2-min incubation with PACAP38, and PACAP-induced insulin secretion is inhibited by H89, an inhibitor of protein kinase (PK)-A (67,69,82). In some cell systems, it has been shown that PACAP acts not only through AC but also through the PLC pathway, with an activation of PKC and an increased formation of intracellular inositol phosphates (IPs) (26). However, the insulinotropic effect of PACAP is not diminished by downregulation of PKC by the phorbol ester 12-o-tetradecanoylphorbol-13-acetate (69), which suggests that PKC is not of major relevance for the insulinotropic action of PACAP. Similarly, PACAP only slightly increases IP₃ formation in insulin-producing cells, suggesting that this pathway is also of limited relevance (67).

In both islet and clonal HIT-T15 cells, PACAP increases the cytoplasmic concentration of calcium by an effect that is abolished by either removal of extracellular calcium or calcium channel blockade (61,67,69). This suggests that increased cytoplasmic calcium is caused by the opening of membrane-bound calcium channels. Electrophysiological studies in HIT-T15 cells have shown that PACAP also increases cytoplasmic calcium by releasing calcium from intracellular stores, which is an action dependent on a small prior influx of calcium through membrane-bound channels (83). However, stimulated calcium influx by PACAP is mainly of importance for its insulinotropic action because the removal of extracellular calcium markedly reduces PACAP-stimulated insulin secretion (67). Because PACAP-induced formation of cAMP in HIT-T15 cells is not markedly affected by the removal of extracellular calcium at the same time (K.F., B.A., unpublished observations), it is assumed that the formation of cAMP (through activation of PKA) induces calcium uptake, as previously observed (84). This increase in cytoplasmic calcium subsequently transduces the potent insulinotropic response of PACAP.

PACAP has also been shown to open membraneous sodium channels, causing an inward membrane current (83), and studies using fluorophor-labeled clonal cells have displayed an increase in the cytoplasmic sodium concentration after administration of PACAP (85). Furthermore, removal of extracellular sodium impairs the insulinotropic action of PACAP without affecting the formation of cAMP (85), suggesting that the uptake of sodium is involved in PACAP-induced insulin secretion. The increase in cytoplasmic sodium by PACAP has been shown to be abolished by the PKA inhibitor H89, suggesting that the sodium channel is regulated by cAMP and PKA (85). The influx of sodium might contribute to the opening of the voltage-dependent calcium channels through a depolarizing effect, thus increasing cytoplasmic calcium and thereby potentiating insulin secretion (83). However, PACAP has been shown to increase cytoplasmic calcium in the absence of extracellular sodium (85), suggesting that sodium is also of importance for the events of exocytosis distal to calcium influx. However, this needs further investigation. One study has shown that wortmannin, which inhibits phosphatidylinositol 3-kinase (PI 3-K), partially inhibits PACAP-induced insulin secretion in HIT-T15 cells, suggesting that this pathway is also involved in the mediation of the insulinotropic effect of PACAP (69). However, it was simultaneously shown that PACAP-induced insulin secretion is not accompanied by increased PI 3-K activity, implying that a PI 3-K–independent, yet wortmannin-sensitive, signaling pathway is involved in the insulinotropic action of PACAP (69). This pathway may involve distal mechanisms that affect the exocytotic expulsion of granules. Hence, although more studies are required for elucidating the exact β-cell mechanism of PACAP, most evidence suggests that PACAP induces insulin secretion mainly by increasing cellular cAMP and activating PKA in association with the uptake of extracellular cations (Fig. 4).

Several approaches have been used to relieve the physiological role of PACAP in the regulation of islet hormone secretion and glucose homeostasis. One strategy has been to incubate isolated islets with PACAP antisera to immunoneutralize endogenous PACAP. In freshly isolated rat islets, this approach resulted in ∼50% inhibition of insulin secretion in the presence of a high concentration of glucose (64,73). It is likely that this effect of PACAP is caused by immunoneutralization of neurally released peptide because no effect of the antisera was observed in isolated islets cultured for 48 h before the experiments (73), a time point when all nerve terminals have disappeared (74). However, although these results suggest the importance of neural PACAP for glucose-stimulated insulin secretion, the physiology of the peptide cannot be deduced. A second more appropriate approach to studying the physiology of PACAP is the generation of a mouse strain deficient in PAC₁-R (PAC₁-R^−/−), by homologous recombination in embryonic stem cells (12). In this model, the 18 exon–containing PAC₁-R gene was disrupted by deletion of a fragment containing exons 8–11. This resulted in expression of a truncated PAC₁-R, which does not bind PACAP, as evident from binding experiments in mouse brain and the dramatic reduction in the cAMP formation in islets after incubation with PACAP (12). These PAC₁-R–deleted mice were viable and developed normally. In addition, fasting glucose and insulin levels were normal, as was the pancreatic insulin content. However, the insulin response to raising the glucose concentration from 4.2 to 16.7 mmol/l in the perfused pancreas was impaired by ∼50% in the PAC₁-R^−/− mice (12). Similarly, the insulin response to exogenouly added glucose in vivo in these mice was accompanied by impaired glucose-stimulated insulin secretion (12), and unpublished data (B.A., K. Persson, unpublished observations) also show that glucose-stimulated insulin secretion in isolated islets is reduced. Therefore, this approach presents evidence that PACAP is of physiological importance for glucose-stimulated insulin secretion. Regarding these PAC₁-R–deleted mice, it is important to remember that the expression of VPAC₁-R and VPAC₂-R are still intact, which enables PACAP to affect islet function through activating these receptors. This explains why PACAP-induced insulin secretion is not abolished but is reduced only by ∼50% in these animals. Future studies on VPAC₁-R– and VPAC₂-R–deleted mice are now of importance for a more complete understanding of the insulinotropic action of PACAP.

The mechanism behind the requirement of PACAP and intact PAC₁-R for a normal glucose-stimulated insulin secretion remains to be established. One mechanism might involve a sequential action of glucose to stimulate insulin secretion, which would be executed directly by an action on β-cells and indirectly through release of PACAP from islet nerve terminals. A second possibility is that intact β-cell PACAP signaling is required for glucose to elicit a full effect, e.g., by promoting full PKA activation. A study in isolated β-cells has supported this latter hypothesis. Thus, it was shown that PACAP at 100 fmol/l restored a calcium response to glucose in β-cells not responding to glucose (64). Hence, in analogy to suggestions for GLP-1 (86), PACAP also seems to make β-cells glucose-responsive or glucose-competent, being a positive modulator of glucose-stimulated insulin secretion.

Because of its localization to parasympathetic nerves in the pancreas, PACAP might be of physiological relevance for neural regulation of islet function, e.g., after meal intake. To study this possibility, a third strategy using pharmacological PACAP receptor antagonism by NH₂-terminally truncated forms of PACAP was used. The efficiency of this approach is evidenced by studies showing that PACAP (6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27) specifically inhibits PACAP27-induced insulin secretion in mice (11). Through this approach, it has been shown that PACAP38(6–38) inhibits vagally induced insulin secretion from the pig pancreas (9) and that PACAP27(6–27) prevents the increase in circulating insulin of gastric glucose gavage in anesthetized mice (11). Hence, vagally and postprandially induced insulin secretion seem partially mediated by PACAP. Similarly, plasma insulin levels after gastric glucose gavage in mice lacking PAC₁-R are reduced compared with those in normal mice (12). Therefore, these results suggest that PACAP is involved in the regulation of early postprandial insulin secretion, which is partially mediated by the vagal nerves during the cephalic phase of insulin secretion. Because PACAP is a neurotransmitter in the gastrointestinal tract (24), PACAP might also affect postprandial insulin secretion through an action on gut motility or release of gastrointestinal hormones, for example. Although this remains a possibility, there is no distinct evidence for this, however, because the rate of increase in circulating glucose after gastric glucose administration in PAC₁-R–deleted mice or in mice given the antagonist PACAP27(6–27) was not different from that in controls (11,12). In addition, the increase in circulating GLP-1 after gastric glucose was exaggerated, not reduced, in PACAP27(6–27)-pretreated mice, a sign of compensation rather than inhibition (11). Therefore, it seems that the action of PACAP to mediate insulin secretion after gastric glucose is executed through an islet effect.

Despite the potent effect of PACAP on insulin secretion, the peptide does not affect or even increase plasma glucose levels after intravenous administration, as demonstrated after acute administration in mice (60,79), dogs (78), and humans (28). Furthermore, after intravenous administration of PACAP together with glucose, despite high insulin levels, the glucose elimination rate was not different from that observed after the administration of glucose alone (79). Therefore, PACAP seems to counteract an augmentation of glucose disposal, which is expected because of the high insulin levels. In fact, it was found that PACAP27 reduced the insulin sensitivity index during the intravenous glucose tolerance test in mice, as calculated by the minimal modeling of insulin and glucose data (79). A potential explanation for this is that glucagon, released by PACAP, counteracts the action of insulin. However, this explanation is unlikely because during the intravenous glucose tolerance test, glucagon secretion is suppressed. Another explanation is that PACAP directly inhibits insulin action. However, this is also an unlikely explanation because PACAP has been shown to stimulate glucose uptake in adipocyte cell line 3T3L1 (87). Two remaining possibilities are that PACAP stimulates the glucose output from the liver, which has been demonstrated in vitro (56,57), and that PACAP indirectly stimulates hepatic glucose output through stimulation of epinephrine release from the adrenals. This latter explanation is supported by the finding that PACAP increases plasma epinephrine levels in mice (79); epinephrine is known to antagonize insulin action (88). This hypothesis is also supported by a recent study showing that after adrenalectomy in mice, which prevents any influence through epinephrine, PACAP potentiates glucose elimination (89). Therefore, the increase in circulating epinephrine after PACAP will counteract the glucose lowering effect of insulin, and this will contribute to reduced net insulin sensitivity in vivo, despite higher insulin levels after PACAP administration.

Because PACAP stimulates glucagon secretion, the peptide may also be involved in the regulation of α-cell function. An important role of these cells is to provide a sufficient glucagon response to hypoglycemia to elicit an adequate glucose counterregulation. In fact, increased glucagon secretion is of major importance in the counterregulatory defense against hypoglycemia (90). It has also been shown that the increase in glucagon secretion during hypoglycemia is partly mediated by the parasympathetic nervous system (91). The localization of PACAP to pancreatic parasympathetic nerves and PACAP-induced glucagon secretion in vivo suggests that PACAP might contribute to the glucagon response to hypoglycemia. Although such studies have yet to be completed, recent preliminary results show that the glucose recovery after insulin-induced hypoglycemia is impaired in PAC₁-R–deleted mice (B.A., K. Persson, unpublished observations). This would suggest a role for PACAP in the counterregulation in hypoglycemia.

To date, most studies regarding the role of PACAP in the regulation of islet hormone secretion have focused on physiological aspects. However, a few recent reports suggest implication of PACAP for development of type 2 diabetes, although this needs to be explored more in detail. A genome-wide nonparametric linkage analysis has revealed linkage between chromosome 18p11, where the gene for PACAP is located, and type 2 diabetes (92). Furthermore, the pancreatic contents of PACAP-immunoreactivity and VPAC₁-R are increased in streptozotocin-induced diabetic rats (65). Finally, daily intraperitoneal injections of PACAP have been shown to reduce circulating glycemia in diabetic GK (Goto-Kakizaki) rats and high-fat–fed mice with impaired glucose tolerance, suggesting a potential of PACAP as an antidiabetogenic agent (93). However, considering the complex actions of PACAP— stimulating insulin secretion as well as glucagon and epinephrine secretion and counteracting the glucose- suppressive action of insulin—the mechanism of this is not completely understood.

The ubiquitously distributed neuropeptide PACAP has turned out to be an important neuropeptide in the endocrine pancreas, where it is expressed in parasympathetic and sensory nerves. A main function of PACAP is to stimulate insulin secretion in a glucose-dependent manner, which is executed mainly through augmentation of the formation of cAMP and stimulation of the uptake of calcium. Although its specific role in islet function as well as its potential involvement for development of type 2 diabetes remains to be establish, a major function of the peptide seems to be related to the requirement of its action for a normal glucose-stimulated insulin secretion and for a normal postprandial insulin response. Therefore, the islet neuropeptide PACAP needs to be taken into account when understanding the complexity of the regulation of islet function.

The studies by the authors have been supported by grants from the Swedish Medical Research Council (no. 6834); Albert Påhlsson, Crafoordska, and Novo Nordisk Foundations; the Swedish Diabetes Association; and the Faculty of Medicine, Lund University.

The authors are grateful to Lilian Bengtsson, Kerstin Knutsson, and Lena Kvist for technical assistance and Dr. Giovanni Pacini (Padova, Italy), Dr. Anton J.W. Scheurink (Groningen, the Netherlands), and Dr. Frank Sundler (Lund, Sweden) for fruitful collaborations.