`Glucagon-Like Peptides
`
`Daniel J. Drucker
`
`Proglucagon contains the sequence of two glucagon-
`like peptides, GLP-1 and GLP-2, secreted
`from
`enteroendocrine cells of the small and large intestine.
`GLP-1 lowers blood glucose in both NIDDM and IDDM
`patients and may be therapeutically useful for treat-
`ment of patients with diabetes. GLP-1 regulates blood
`glucose via stimulation of glucose-dependent insulin
`secretion, inhibition of gastric emptying, and inhibi-
`tion of glucagon secretion. GLP-1 may also regulate
`glycogen synthesis in adipose tissue and muscle; how-
`ever, the mechanism for these peripheral effects
`remains unclear. GLP-1 is produced in the brain, and
`intracerebroventricular GLP-1 in rodents is a potent
`inhibitor of food and water intake. The short duration
`of action of GLP-1 may be accounted for in part by the
`enzyme dipeptidyl peptidase 4 (DPP-IV), which cleaves
`GLP-1 at the NH 2-terminus; hence GLP-1 analogs or the
`lizard peptide exendin-4 that are resistant to DPP-IV
`cleavage may be more potent GLP-1 molecules in vivo.
`GLP-2 has recently been shown to display intestinal
`growth factor activity in rodents, raising the possibility
`that GLP-2 may be therapeutically useful for enhance-
`ment of mucosal regeneration in patients with intesti-
`nal disease. This review discusses recent advances in
`our understanding of the biological activity of the
`glucagon-like peptides. Diabetes 47:159-169, 1998
`
`THE INCRETIN CONCEPT AND p-CELL FUNCTION
`Enhancement of insulin secretion from the islet |3 cell is a prin-
`cipal goal for treatment of patients with NIDDM. The obser-
`vation that sulfonylureas stimulate insulin secretion has pro-
`vided the rationale for the therapeutic use of these agents in
`the treatment of NIDDM. Nevertheless, the sulfonylurea
`stimulation of insulin secretion is not strictly glucose depen-
`dent, and hence hypoglycemia is an undesirable side effect of
`sulfonylurea treatment, particularly in elderly patients.
`The observation that glucose administered via the gas-
`trointestinal tract is associated with a greater stimulation of
`insulin release compared with a comparable glucose chal-
`lenge given intravenously (1,2) prompted a search for the
`responsible "incretins," gut-derived factors that increase glu-
`cose-stimulated insulin secretion (3). The concept of the
`
`From the Department of Medicine, the Toronto Hospital; and the Banting
`and Best Diabetes Centre, University of Toronto, Ontario, Canada.
`Address correspondence and reprint requests to Dr. Daniel J. Drucker,
`The Toronto Hospital, 200 Elizabeth St., CCRW3-838, Toronto, Canada M5G
`2C4. E-mail: d.drucker@utoronto.ca.
`Received for publication 9 September 1997 and accepted in revised
`form 8 October 1997.
`CNS, central nervous system; DPP-IV, dipeptidyl peptidase 4; GIP, glu-
`cose-dependent insulinotropic peptide; GLP, glucagon-like peptide; GLP-1R,
`GLP-1 receptor; GRP, gastrin-releasing peptide; ICV, intracerebroventricu-
`lar; PC, prohormone convertase; PGDP, proglucagon-derived peptide; RT-
`PCR, reverse transcription-polymerase chain reaction.
`
`DIABETES, VOL. 47, FEBRUARY 1998
`
`enteroinsular axis suggested that insulin secretagogues were
`synthesized in and released from the intestinal endocrine sys-
`tem after nutrient ingestion. The isolation and characterization
`of glucose-dependent insulinotropic peptide (GIP) repre-
`sented an important advance in the identification of intestinal
`incretin hormones. GIP is released from enteroendocrine cells
`in the duodenum and proximal jejunum after nutrient intake
`and stimulates insulin secretion in a glucose-dependent man-
`ner (3,4). Nevertheless, immunoneutralization of GIP or
`removal of GIP from intestinal extracts does not result in
`complete elimination of incretin activity, consistent with the
`presence of additional gut-derived factors with insulinotropic
`activity (4).
`After the isolation of the cDNAs and genes encoding
`proglucagon approximately 15 years ago (5-7), two novel
`glucagon-like peptides, GLP-1 and GLP-2, were identified
`COOH-terminal to the glucagon sequence in mammalian
`proglucagon (Figs. 1 and 2). Initial characterization of GLP-
`1 bioactivity using NH2-terminally extended GLP-l(l-37)
`failed to demonstrate effects on blood glucose or insulin
`secretion; however, subsequent experiments using the NH2-
`terminally truncated GLP-l(7-36) amide or GLP-l(7-37) pep-
`tides demonstrated potent effects on glucose-dependent
`insulin secretion, islet cell cAMP formation, and insulin gene
`expression (8-12). The principal aim of this review is to high-
`light recent advances in our understanding of the biology of
`the GLPs. Previous reviews of the biology of GLP-1 and GLP-
`2 offer a detailed introduction to the subject (4,13-15).
`
`GLP-1 BIOSYNTHESIS AND SECRETION
`A single proglucagon gene in mammals gives rise to an iden-
`tical proglucagon RNA transcript that is translated and
`processed differently in brain, pancreatic islets, and intestine
`(Fig. 1) (16). In contrast, vertebrates such as the chicken, fish,
`frog, and lizard may contain two proglucagon genes, and use
`alternative RNA splicing for the generation of proglucagon
`mRNA transcripts that encode for GLP-1 but not GLP-2 in the
`pancreas and both GLP-1 and GLP-2 in the intestine (17,18).
`Although considerable progress has been made in elucidat-
`ing the factors that control proglucagon gene expression in
`islet cells, much less is known about the regulation of
`proglucagon gene expression and, hence, the control of GLP-
`1 biosynthesis, in enteroendocrine cells. Transgenic experi-
`ments have demonstrated that different
`tissue-specific
`enhancers specify islet versus intestinal proglucagon gene
`expression (19,20); however, the intestine-specific pro-
`glucagon gene enhancer remains poorly defined (21). The pan-
`creatic A-cell and enteroendocrine L-cell both express the
`transcription factor cdx-2/3, which regulates proglucagon
`gene expression in pancreas and intestine (22,23); however,
`transcription factors important for regulation of the
`proglucagon promoter specifically in the enteroendocrine
`
`159
`
`AstraZeneca Exhibit 2069
`Mylan v. AstraZeneca
`IPR2015-01340
`
`Page 1 of 11
`
`
`
`GLUCAGON-LIKE PEPTIDES
`
`Proglucagon
`
`GRPP
`
`30
`
`33
`
`Glucagon
`
`64
`IP-1
`
`69
`
`78
`
`107/8
`
`126
`
`GLP-1
`
`IP-2
`
`GLP-2
`
`61
`
`72
`
`111
`
`123
`
`158
`
`158
`
`Glicentin
`
`Oxyntomodulin
`
`MPGF
`
`Pancreas I ™ g on
`
`Oxyntomodulin
`Intestine G UM
`GLP-2
`LIP-2
`FIG. 1. Structural organization of mammalian proglucagon. The numbers refer to amino acid sequences in proglucagon. The peptides released
`by posttranslational processing in the pancreas and intestine are shown. GRPP, glicentin-related pancreatic polypeptide; IP, intervening pep-
`tide; MPGF, major proglucagon fragment.
`
`cell have not been extensively characterized. Furthermore,
`despite interest in potential new strategies for increasing
`GLP-1 synthesis and secretion in diabetic patients, the factors
`important for the regulation of human intestinal proglucagon
`biosynthesis remain unknown.
`The liberation of GLP-1 in the intestine but not the pancreas
`appears to be due to the tissue-specific expression of pro-
`hormone convertases (PCs) in the enteroendocrine cells of the
`small and large bowel. Whereas both PCI and PC2 cleave
`proglucagon to generate the major proglucagon fragment and
`glicentin and oxyntomodulin (24,25), PCI expressed in
`enteroendocrine cells appears to be the enzyme responsible
`for the liberation of both GLP-1 and GLP-2 (25). Although
`multiple immunoreactive forms of GLP-1 are liberated in vivo,
`including GLP-l(7-36) amide and GLP-l(7-37), the majority of
`circulating GLP-1 in humans appears to be GLP-l(7-36) amide
`(26). Nevertheless, in vivo studies have shown that both
`molecular forms of NH2-terminaUy truncated GLP-1 are
`equipotent with regard to their insulin-stimulating properties;
`in addition, both appear to exhibit similar half-lives in vivo (27).
`
`Whether control of PC activity in the intestine is an important
`regulator of GLP-1 synthesis remains to be determined.
`Because the enteroendocrine cell is exposed to both cir-
`culating humoral factors and luminal intestinal contents,
`intestinal proglucagon-derived peptide (PGDP) biosynthesis
`and secretion are subject to regulation by both hormonal
`and nutritional factors. Nutrient intake stimulates the syn-
`thesis and secretion of PGDPs from the enteroendocrine cell
`in rodents (28). In one study, proglucagon mRNA abundance
`decreased with fasting and increased with refeeding in rat
`jejunum, and the profile of circulating enteroglucagon and
`GLP-1 paralleled changes observed in jejunal proglucagon
`mRNA (29). The rapid rise in plasma GLP-1 levels after nutri-
`ent ingestion and the distal location of the majority of GLP-
`1-containing enteroendocrine cells in the ileum and colon has
`led to the suggestion that one component of the nutrient-
`induced secretory signal may be indirect, perhaps via GIP or
`gastrin-releasing peptide (GRP) release from the proximal
`jejunum. GLP-1 secretion from the distal ileum was abol-
`ished when intervening segments of intestine were resected,
`
`DPP-IVI
`(7-37)
`Human GLP-1 HAEGT FTSDV SSYLE GQAAK EFIAW LVKGR G
`(7-36)Mdde
`Human GLP-1 HAEGT FTSDV SSYLE GQAAK EFIAW LVKGR
`Exendin-4
`HGEGT FTSDL SKQME EEAVR LFIEW LKNGG PSSGA PPPSG
`YAEGT FISDY SIAMD KIHQQ DFVNW LLAQK GKKND WKHNI TQ
`Human GIP
`
`Human GLP2 HADGS FSDEM NTILD NLAAR DFINW LIQTK ITD
`Rat GLP2
`HADGS FSDEM NTILD NLATR DFINW LIQTK ITD
`Mouse GLP2 HADGS FSDEM STILD NLATR DFINW LIQTK ITD
`
`FIG. 2. Amino acid sequences of GLP-1, exendin-4, GIP, and GLP-2. The arrow designates the recognition site for DPP-IV enzymatic cleavage.
`Residues in rat or mouse GLP-2 that differ in sequence from human GLP-2 are underlined.
`
`160
`
`DIABETES, VOL. 47, FEBRUARY 1998
`
`Page 2 of 11
`
`
`
`Biological activities of GLP-1
`
`Adipose
`tissue
`
`and water intake
`
`7 Glycogen synthesis
`Glycogenolysis
`
`JGastric emptying
`Islets of Langerhans
`f Glucagon
`•Insulin
`ASomatostatin
`
`Nutrients
`L
`
`[ Intestine"!
`FIG. 3. Schematic representation of GLP-1 action.
`
`and infusion of a GRP antagonist inhibited the L-cell
`response to nutrient ingestion, observations consistent with
`GRP having an important role in the humoral regulation of
`GLP-1 secretion (30).
`In isolated perfused rat ileum preparations, GLP-1 secretion
`was stimulated by cholinergic agonists, bombesin, calci-
`tonin-gene-related peptide, and GIP (31,32), and in the iso-
`lated perfused rat colon, it was stimulated by pj-adrenergic and
`cholinergic agonists, bombesin, and calcitonin-gene-related
`peptide, and GIP (33). Somatostatin directly inhibits L-cell and
`GLP-1 secretion, and galanin antagonizes the stimulatory
`effect of GIP on GLP-1 release in rats (34). Luminal perfusion
`with glucose, pectin, or the bile acid hodeoxycholate stimu-
`lated GLP-1 secretion from the rat colon, suggesting that
`enteroendocrine cells in the large bowel are also sensitive to
`luminal contents in vivo (35).
`GLP-1 secretion is also stimulated by nutrient ingestion in
`humans. Basal circulating levels of human GLP-1 (7-36)
`amide range from 0.4 to 7.0 pmol/1, depending on the assay,
`and are stimulated after oral but not intravenous glucose
`administration (26,36). Secretion of GLP-1 throughout the
`day increases after meal ingestion, in parallel with meal-
`related increases in insulin secretion (37,38). GLP-1 release
`was also stimulated after oral administration of galactose,
`amino acids, and corn oil (36). GLP-1 is secreted in a pulsatile
`manner in humans; glucose ingestion increases the amplitude,
`but not the frequency, of GLP-1 secretion (39). The integrated
`GLP-1 pulse amplitude was reduced by atropine, consistent
`with the importance of cholinergic mechanisms in the con-
`trol of GLP-1 secretion.
`
`GLP-1 DEGRADATION
`An important determinant of the circulating levels of bioactive
`GLP-1 appears to be the NHg-terminal degradation of the pep-
`tide by the enzyme DPP-TV (40). Cleavage of GLP-1 at the
`penultimate alanine residue to generate GLP-1 (9-36)amide
`occurs rapidly in plasma (40), and the half-life of intact GLP-
`1 in vivo appears to be less than 2 min (41). GLP-l(9-36)amide
`constitutes 53.5% of the concentration of intact GLP-l(7-36)
`amide in the fasted state; however, after nutrient ingestion,
`human GLP-l(9-36)amide is relatively more abundant than the
`intact 7-36(amide) molecule (42). GLP-1 (9-36)amide also
`
`DIABETES, VOL. 47, FEBRUARY 1998
`
`D.J. DRUCKER
`
`binds to the GLP-1 receptor, albeit with lower affinity than the
`(7-36)amide form, and may function as a competitive antago-
`nist of the GLP-1 receptor in vivo (43). Radioimmunoassays
`that do not distinguish between intact GLP-l(7-36) amide and
`the NH2-terminally deleted GLP-1 (9-36)amide may overesti-
`mate the actual concentration of circulating bioactive GLP-1
`(42); experiments that measure total immunoreactive circu-
`lating GLP-1 need to be interpreted with caution in light of this
`new information. DPP-IV activity is inhibited by low temper-
`ature and diprotin A (41,42), hence the importance of col-
`lecting blood samples for measurements of GLP-1 immunore-
`activity on ice in the presence of appropriate protease
`inhibitors. Intact GLP-1 appears to be cleared predominantly
`through renal extraction; the contribution of extrarenal tissues
`to clearance of GLP-1 under normal physiological conditions
`remains to be determined (44).
`
`GLP-1 ACTION
`The GLP-1 receptor. GLP-1 exerts its actions via binding to
`a G-protein-linked receptor expressed on islet (3-cells (45).
`The human GLP-1 receptor (46,47) is 90% homologous to the
`rat receptor, and the gene has been localized to 6p21 (48). No
`GLP-1 receptor mutations have been reported in NIDDM
`patients, and genetic analysis has failed to demonstrate link-
`age between the GLP-1 receptor gene and populations with
`maturity-onset diabetes of the young or NIDDM (49). GLP-1
`receptor mRNA transcripts have been detected by Northern
`blotting in rodent tissues such as the islets, lung, kidney,
`stomach, and brain (45,50). GLP-1 receptor mRNA tran-
`scripts have been more difficult to detect by Northern blot-
`ting in human tissues (46,47), but have been identified in
`human pancreas, lung, kidney, stomach, heart, and brain by
`RNAse protection analyses (51).
`Some controversy remains with regard to the expression
`of GLP-1 receptor mRNA transcripts in peripheral tissues.
`Although low levels of GLP-1 receptor mRNA transcripts
`and GLP-1 binding have been reported in rat muscle and
`liver (50), these findings have not been universally con-
`firmed (51,52). Furthermore, discrepancies among results
`obtained using ligand binding, in situ hybridization, RNAse
`protection, and reverse transcription-polymerase chain
`reaction (RT-PCR) for characterization of GLP-1 receptor
`expression have led to the suggestion that structural variants
`of the GLP-1 receptor, or a second closely related receptor,
`may be expressed in different tissues (52-54); however,
`cDNAs encoding variant GLP-1 receptors have not yet been
`identified. Experiments using primary islet cultures, (J-cell
`lines, and cells transfected with the GLP-1 receptor cDNA
`have shown that GLP-1 signaling is coupled to both activa-
`tion of adenylate cyclase and phospholipase C pathways
`(12,45,50,55,56). GLP-1 binding is associated with an
`increase in cytosolic-free calcium (50,56,57). GLP-1 may
`increase intracellular [Ca2+] via activation of a prolonged
`cAMP-sensitive inward current leading to membrane depo-
`larization and increases in intracellular calcium (58). GLP-1
`receptor responsivity may be desensitized in vitro after
`exposure to agonist or activation of protein kinase C (59,60),
`and receptor desensitization appears to correlate with
`receptor phosphorylation (59,61).
`A number of distinct yet complementary actions con-
`tribute to the glucose-lowering properties of GLP-1 (Fig. 3).
`Binding of GLP-1 to its p-cell receptor stimulates insulin
`
`161
`
`Page 3 of 11
`
`
`
`GLUCAGON-LIKE PEPTIDES
`
`secretion in a glucose-dependent manner, and GLP-1
`increases insulin mRNA (12), likely via induction of insulin
`gene transcription through a cAMP-dependent mechanism
`(62,63). GLP-1 also confers glucose sensitivity to glucose-
`resistant (3-cells (55), thereby enhancing the ability of p-cells
`to secrete insulin in a glucose-dependent manner. Consis-
`tent with this hypothesis, GLP-1 increased the insulinotropic
`effect of glibenclamide in the perfused rat pancreas (64),
`and GLP-1 administration to glucose-intolerant aging Wistar
`rats lowered plasma glucose and increased circulating
`insulin and insulin RNA, in keeping with a role for GLP-1 in
`the restoration of normal islet function and control of insulin
`biosynthesis (65).
`GLP-1 also lowers blood glucose via inhibition of
`glucagon secretion (66,67). GLP-1 likely acts directly on the
`pancreatic A-cell and via indirect mechanisms through stim-
`ulation of somatostatin and insulin secretion. GLP-1 infusion
`in C-peptide-negative diabetic dogs lowered circulating
`plasma glucagon, suggesting that the glucagonostatic effects
`of GLP-1 are at least partially independent of circulating
`insulin (68). Consistent with a direct effect of GLP-1 on
`glucagon and somatostatin secretion, GLP-1 receptors have
`been localized to the a- and 8-cells of the islets (69). Para-
`doxically, GLP-1 stimulated glucagon secretion from iso-
`lated rat a cells, and this stimulatory effect was inhibited by
`somatostatin, raising the possibility that the glucagonostatic
`effects of GLP-1 are partly indirect through a paracrine
`effect on somatostatin secretion (70). In contrast, GLP-1
`directly inhibited glucagon secretion and intracellular cAMP
`in the glucagon-producing InRl-G9 cell line in the absence
`of somatostatin (71); the relative contributions of different
`mechanism(s) underlying the inhibitory effect of GLP-1 on
`the A-cell remain unclear.
`Extrapancreatic effects of GLP-1. GLP-1 attenuates
`meal-associated glucose excursion by directly inhibiting gas-
`tric emptying (72). GLP-1 also inhibits postprandial acid
`secretion, and GLP-1 receptors have been demonstrated in
`the stomach (45,73). The GLP-1 receptor expressed in heart
`is structurally identical to the pancreatic islet receptor (51),
`and GLP-1 increased systolic and diastolic pressure and
`heart rate in rats (74). The highest levels of GLP-1 receptor
`mRNA transcripts are found in lung (45), consistent with
`identification of GLP-1 binding sites in lung membrane
`preparations (75,76). Although GLP-1 stimulated macro-
`molecule secretion from tracheal ring preparations (77), a
`physiological role, if any, for GLP-1 in pulmonary physiology
`in vivo remains to be determined.
`The mechanism of action of GLP-1 in peripheral tissues
`such as liver, muscle, and adipose tissue remains unclear
`(Fig. 3). GLP-1-stimulated glycogen synthesis in isolated
`hepatocytes from normal and diabetic rats (78) and GLP-1
`binding in rat hepatocyte and liver membrane preparations
`(79) has been demonstrated. Both fish and human GLP-1
`stimulated glycogenolysis in fish hepatocyte preparations
`(80). In contrast, other investigators failed to demonstrate an
`effect of GLP-1 on hepatic glycogenolysis or glycogen syn-
`thesis in rat liver (81). Similarly, GLP-1 binding activity has
`been observed in rat skeletal muscle membranes (54), and
`GLP-1 effects on glucose incorporation into glycogen have
`been demonstrated by some investigators (82,83), but not by
`others (84). GLP-1 binding has also been demonstrated in rat
`and human adipose tissue membranes (53,85), and GLP-
`
`162
`
`1-enhanced insulin-stimulated glucose uptake in 3T3-L1
`adipocytes (86) and isolated rat adipocytes (87). Intriguingly,
`GLP-1 decreased intracellular cAMP in 3T3-L1 adipocytes,
`thereby providing indirect evidence for the presence of a
`second receptor with signaling properties distinct from those
`described for the pancreatic GLP-1 receptor (88). Although
`GLP-1 receptor expression has been demonstrated by RT-PCR
`in RNA from rat muscle and fat pad (86), other investigators,
`using a combination of RT-PCR, RNAse-protection, and in situ
`hybridization experiments, failed to detect GLP-1 receptor
`mRNA transcripts in adipose tissue, liver, and muscle (52).
`Given the lack of conclusive evidence for the expression of
`the pancreatic GLP-1 receptor in muscle, liver, and adipose
`tissue, the mechanisms and receptor(s) mediating these
`peripheral effects of GLP-1 remain unclear.
`PGDPs and GLP-1 are synthesized in the central nervous
`system (CNS), and GLP-1 receptors have been localized
`through a combination of in situ autoradiography and
`hybridization studies to different regions of the CNS (89,90).
`Although GLP-1 immunoreactivity is widely distributed in
`many regions of the brain, GLP-1 mRNA transcripts are local-
`ized predominantly to the brain stem and, to a lesser extent,
`the hypothalamus, thereby supporting a role for peptidergic
`transport from brain stem neurons in the regulation of GLP-
`1 CNS distribution (91-93). Consistent with these findings,
`GLP-1 binding sites and GLP-1 receptor RNA transcripts have
`been identified throughout the CNS (89,90,94,95) and in the
`pituitary (96,97). GLP-1 may also play a role in the peripheral
`nervous system, as intraportal GLP-1 activates vagal nerve
`activity in rats (98).
`A potential role for GLP-1 in the central control of feeding
`behavior was suggested by studies demonstrating that intra-
`cerebroventricular (ICV) administration of GLP-1 in rats
`inhibited food intake and induced c-fos immunoreactivity in
`the paraventricular nucleus and amygdala (99). Although
`ICV GLP-1 and leptin both inhibit food intake, leptin acti-
`vated c-/os-like immunoreactivity in regions of the rat brain
`different from those activated by GLP-1 (100). Furthermore,
`the inhibitory effects of leptin were of comparatively longer
`duration, and GLP-1, but not leptin, produced conditioned
`taste aversion, implying distinct roles for these peptides in the
`central regulation of feeding (101). ICV GLP-1 inhibited basal
`water intake (102) and stimulated urinary excretion of water
`and sodium, and both ICV and intraperitoneal GLP-1 inhibited
`basal and ANG II-induced drinking behavior (103) and
`reduced body temperature in rats. Whether the GLP-1 effects
`on water regulation are related to or distinct from the periph-
`eral effects of GLP-1 on heart rate and blood pressure (74)
`remains uncertain.
`Studies using GLP-1 receptor antagonists. The obser-
`vation that a truncated lizard GLP-1-related peptide,
`exendin(9-39), binds to the mammalian GLP-1 receptor and
`functions as a GLP-1 antagonist (46,104) has provided the
`opportunity to carry out studies examining the transient
`reduction or loss of GLP-1 action both in vitro and in vivo.
`Exendin(9-39) administered to rats reduced postprandial
`insulin levels (105), reduced insulin secretory response, and
`increased blood glucose after intraduodenal glucose infu-
`sion (106), thereby providing important evidence that GLP-1
`is a physiologically relevant incretin in vivo. Infusion of
`exendin(9-39) in baboons increased fasting levels of glucose
`and glucagon and increased postprandial glycemic excur-
`
`DIABETES, VOL. 47, FEBRUARY 1998
`
`Page 4 of 11
`
`
`
`D.J. DRUCKER
`
`sion with reduction of postprandial insulin secretion (107).
`The postprandial glycemic excursion was also increased
`after infusion of a GLP-1-specific monoclonal antibody (107).
`Exendin(9-39) blocks the extrapancreatic effects of GLP-1 in
`the cardiovascular system, liver, and muscle (108,109), and
`functions as an antagonist of the brain GLP-1 receptor,
`inhibiting the effects of ICV GLP-1 on food and water intake
`(99,103), and potentiating the stimulatory actions of neu-
`ropeptide Y on food intake (99).
`Despite experimental evidence that exendin(9-39) may be
`a relatively specific GLP-1 receptor antagonist, experiments
`with the cloned rat and human GIP receptors have demon-
`strated that relatively high concentrations of exendin(9-39)
`may also function as a GIP receptor antagonist (110,111).
`The results of recent experiments suggest that truncation of
`the first 3-7 amino acids of exendin-4 may produce exendin
`analogs that are up to 10-fold more potent as GLP-1 antago-
`nists than exendin(9-39) (112); however, the relative specificity
`of these analogs for the GLP-1 versus the GIP receptor has not
`yet been reported.
`The GLP-1 receptor - /- mouse. Targeted disruption of the
`gene encoding the pancreatic islet GLP-1 receptor (GLP-1R) in
`embryonic stem cells followed by the derivation oftransgenic
`GLP-1R -/- mice has permitted the analysis of the role of GLP-
`1 in both glucose control and appetite regulation in vivo. GLP-
`1R -/- mice exhibit mild fasting hyperglycemia and glucose
`intolerance after oral glucose challenge (113). The abnormal
`glycemic excursion after oral glucose loading was associated
`with a reduction in glucose-stimulated insulin secretion, con-
`sistent with an essential role for GLP-1 signaling in the regu-
`lation of glucose-dependent insulin secretion (113). Remark-
`ably, GLP-1R -/- mice also exhibit abnormal glycemic excur-
`sion after intraperitoneal glucose challenge, suggesting that
`intact GLP-1 signaling is important for the handling of a glucose
`load, independent of the site of glucose entry.
`Despite evidence that ICV GLP-1 is a potent inhibitor of
`food intake, analysis of body weight in GLP-1R -/- mice (up
`to age 18 months; D.J.D., unpublished observations) did not
`demonstrate any significant changes in body mass com-
`pared with age- and sex-matched control mice (113). Dis-
`ruption of GLP-1 signaling in the brain does not appear to
`be associated with chronically increased food intake, and
`GLP-1 receptor -/- mice do not eat more than control mice
`in short-term feeding studies (113). Furthermore, despite
`evidence for significantly increased leptin sensitivity in the
`GLP-1R -/- islet, the inhibition of food intake after ICV
`leptin appears relatively normal in the GLP-1 R -/- mouse
`(114). No GLP-1 binding sites are detectable in the CNS of
`GLP-1R -/- mice, consistent with the presence of a single
`brain GLP-1 receptor. Taken together, studies with the
`GLP-1R -/- mouse support an essential role for GLP-1 in the
`regulation of glycemia and glucose-stimulated insulin
`secretion; however, the available data suggest that GLP-1
`signaling may not be essential for regulation of satiety or
`body weight. The CNS results may be explained by genetic
`redundancy, in that multiple compensatory mechanisms
`likely exist for central regulation of food intake and body
`weight (115). Furthermore, the possibility that disruption
`of GLP-1 signaling from birth may be associated with sub-
`tle developmental abnormalities in the CNS that may influ-
`ence the regulation of feeding and body weight cannot be
`excluded.
`
`DIABETES, VOL. 47, FEBRUAEY 1998
`
`GLP-1 IN HUMAN STUDIES
`Normal human subjects. GLP-1, either as the (7-36)amide
`or in the 7-37 forms (27), stimulates insulin secretion,
`inhibits glucagon secretion, and lowers blood glucose in
`humans in the fasting or postprandial state (116,117). Infu-
`sion of GLP-1 in normal human volunteers delays gastric
`emptying (72), although nausea and vomiting, likely due to
`inhibition of gastric emptying, have been observed with
`higher dosages of GLP-1 (117). GLP-1 may also regulate gly-
`cemia by modulating hepatic glucose production, predomi-
`nantly through its effects on levels of circulating insulin and
`glucagon (118,119). The mechanisms underlying the insulin-
`independent effects of GLP-1 facilitating glucose disposal
`remain unclear. GLP-1 infusion increased glucose disposal
`and glucose effectiveness in short-term (4-h) studies in nor-
`mal subjects (120,121), but had no effect on glucose dis-
`posal (independent from the effect of insulin) after intra-
`venous glucose loading (119). Furthermore, no effect of
`GLP-1 on insulin sensitivity was observed during a 3-h
`hyperinsulinemic, euglycemic clamp (122) or after oral fat
`ingestion or intravenous glucose loading (121). Although
`the majority of human studies deliver GLP-1 by intravenous
`or subcutaneous injection, a recent promising study demon-
`strated that formulation of GLP-1 as a buccal tablet pro-
`motes transmucosal absorption, resulting in increased lev-
`els of insulin and decreased glucagon and glucose in healthy
`human volunteers (123).
`NIDDM patients. The demonstration that GLP-1 exhibited
`considerably greater potency compared with that of GIP as
`a glucose-dependent stimulator of insulin secretion in diabetic
`subjects has stimulated considerable interest in the use of
`GLP-1 for the treatment of NIDDM (124). Although both GLP-
`1 and GIP stimulate insulin secretion, GLP-1, but not GIP,
`inhibits gastric emptying and lowers circulating glucagon in
`NIDDM patients (124,125). GLP-1 infusion normalized fasting
`hyperglycemia in NIDDM patients with poor glycemic control
`(126) and improved basal and glucose- and arginine-stimu-
`lated insulin secretion in NIDDM subjects (127). Several
`short-term studies in NIDDM patients have demonstrated
`that GLP-1, whether administered by intravenous infusion or
`subcutaneous injection, normalizes both fasting and post-
`prandial glycemia (128), predominantly by enhancing fJ-cell
`function and inhibiting both gastric emptying and glucagon
`secretion (129-131). Additional evidence for a beneficial
`effect of GLP-1 on islet function in NIDDM patients derives
`from studies demonstrating that the glucose-lowering effect
`of GLP-1 is enhanced by the sulfonylurea glibenclamide in
`patients previously resistant to glibenclamide alone (64).
`A recent study has examined the effect of more prolonged
`GLP-1 treatment on glucose control in NIDDM patients
`receiving intensive insulin therapy for 1 week followed by
`either an additional 7 days on insulin alone or insulin plus
`GLP-1 at meals. The GLP-1-treated group required less
`exogenous insulin and exhibited a reduction in postprandial
`hyperglycemia but increased preprandial glycemia, possibly
`due to the short duration of action of GLP-1 (132). GLP-1 treat-
`ment also increased LDL particle diameter and reduced both
`lipoprotein lipase and hepatic lipase activity. A second study
`reported the results of a 3-week, double-blind crossover trial
`of GLP-1 or saline three times a day before meals in five
`NIDDM patients with poor glycemic control. GLP-1 treat-
`ment lowered postprandial glucagon levels and improved
`
`163
`
`Page 5 of 11
`
`
`
`GLUCAGON-LIKE PEPTIDES
`
`postprandial glycemic control, despite no significant increase
`in postprandial insulin levels in these patients (133).
`IDDM patients. GLP-1 lowered postprandial blood glucose
`and the meal-related insulin requirement in IDDM patients in
`association with a reduction in circulating glucagon and
`somatostatin (116). These observations emphasized the
`potential importance of GLP-1 for lowering blood glucose
`independent of its actions on the pancreatic (3-cell. The glu-
`cose-lowering properties of GLP-1 in IDDM patients after
`meal ingestion are likely due in large part to a delay of gas-
`tric emptying and inhibition of glucagon secretion (134).
`Administration of lower dosages of GLP-1 to IDDM patients
`decreased postprandial glycemic excursion but not glucagon
`levels, suggesting that delayed gastric emptying is a major con-
`tributor to the decreased blood glucose observed in these
`studies (135). In contrast, inhibition of glucagon secretion is
`the primary determinant of the reduction in fasting glycemia
`observed in IDDM patients infused with GLP-1 (136).
`Novel glucagon-like peptides. Peptides originally isolated
`from the venom of the Gila monster lizard H. suspectum or
`H. horridum increased cyclic AMP and amylase secretion
`from dispersed pancreatic acini in vitro. Screening of lizard
`venom for the presence of peptides with amino terminal his-
`tidine residues culminated in the isolation of two peptides,
`designated exendin-3 (137) and exendin-4 (138). Both peptides
`exhibit approximately 50% amino acid identity to mammalian
`GLP-1 (Fig. 2), but are encoded by unique exendin genes
`with different patterns of tissue-specific expression in the
`lizard (18). Both exendin peptides increase cAMP in dis-
`persed pancreatic acini, but only exendin-3 increases amylase
`secretion; exendin-4 does not most likely because exendin-4
`does not bind to the pancreatic vasoactive intestinal peptide
`(VIP) receptor (138). Exendin-4 binds to the GLP-1 receptor
`and stimulates glucose-dependent insulin secretion in islet
`cells in vitro (46,104) and in animal studies in vivo (106).
`Exendin-4 also mimics the majority of peripheral actions of
`GLP-1 in the cardiovascular system, stomach, and brain
`(139). Intriguingly, GLP-1 also binds to the putative exendin
`receptor and increases acinar cAMP; however, the identity of
`the exendin receptor expressed on pancreatic acinar cells
`remains unclear (140).
`The first three amino acids at the NH2-terminus of