Current Pharmaceutical Design, 2001, 7, 1399-1412
`
`1399
`
`Development of Glucagon-Like Peptide-1-Based Pharmaceuticals as
`Therapeutic Agents for the Treatment of Diabetes
`
`Daniel J. Drucker*
`
`Department of Medicine, Banting and Best Diabetes Centre, Toronto General Hospital, University of Toronto,
`Toronto, Canada
`
`Abstract: Glucagon-like peptide-1 (GLP-1) is released from gut endocrine cells following nutrient
`ingestion and acts to regulate nutrient assimilation via effects on gastrointestinal motility, islet hormone
`secretion, and islet cell proliferation. Exogenous administration of GLP-1 lowers blood glucose in normal
`rodents and in multiple experimental models of diabetes mellitus. Similarly, GLP-1 lowers blood glucose
`in normal subjects and in patients with type 2 diabetes. The therapeutic utility of the native GLP-1
`molecule is limited by its rapid enzymatic degradation by the serine protease dipeptidyl peptidase IV.
`This review highlights recent advances in our understanding of GLP-1 physiology and GLP-1 receptor
`signaling, and summarizes current pharmaceutical strategies directed at sustained activation of GLP-1
`receptor-dependent actions for glucoregulation in vivo. Given the nutrient-dependent control of GLP-1
`release, neutraceuticals or modified diets that enhance GLP-1 release from the enteroendocrine cell may
`exhibit glucose-lowering properties in human subjects. The utility of GLP-1 derivatives engineered for
`sustained action and/or DP IV-resistance, and the biological activity of naturally occurring GLP-1-related
`molecules such as exendin-4 is reviewed. Circumventing DP IV-mediated incretin degradation via
`inhibitors that target the DP IV enzyme represents a complementary strategy for enhancing GLP-1-
`mediated actions in vivo. Finally, the current status of alternative GLP-1-delivery systems via the buccal
`and enteral mucosa is briefly summarized. The findings that the potent glucose-lowering properties of
`GLP-1 are preserved in diabetic subjects, taken together with the potential for GLP-1 therapy to preserve
`or augment b cell mass, provides a powerful impetus for development of GLP-1-based human
`pharmaceuticals.
`
`INTRODUCTION
`
`Glucagon-like peptide-1 is a posttranslational
`product of the proglucagon gene liberated from gut
`endocrine cells in response to nutrient ingestion.
`GLP-1 exerts multiple actions that converge on the
`lowering of blood glucose in rodents and human
`subjects. The pleiotropic actions of GLP-1 (Figure
`1) are preserved in human subjects and GLP-1
`administration lowers blood glucose in patients
`with both type 1 and type 2 diabetes [1-8]. These
`findings suggest that strategies for enhancing GLP-
`1 action (Figure 2), either via stimulating GLP-1
`release, reducing GLP-1 degradation, delivery of
`more potent GLP-1 peptide analogues, or
`
`*Address correspondence to this author at the Toronto General Hospital,
`101 College Street CCRW3-845, Toronto Ontario Canada, M5G 2C4;
`Ph.:+416-340-4125; Fax: 416-978-4108; e-mail: d.drucker@utoronto.ca
`
`derivation of small molecules that activate GLP-1
`receptor signaling, warrant vigorous and rigorous
`scientific assessment. The aim of this review is to
`highlight our current understanding of GLP-1
`action with a focus on reviewing the efficacy, and
`theoretical advantages and pitfalls of different
`pharmaceutical approaches
`that converge on
`increasing signaling through the GLP-1 receptor.
`The reader is referred to several comprehensive
`recent reviews on GLP-1 action for an introduction
`to GLP-1 physiology [9-12].
`
`GLP-1 SYNTHESIS AND SECRETION
`
`the primary physiological
`Nutrients are
`regulators of GLP-1 secretion from gut endocrine
`cells. Both fats and carbohydrates stimulate GLP-
`1 secretion in rodent and human studies. The
`precise mechanisms underlying the detection and
`
`1381-6128/01 $28.00+.00
`
`© 2001 Bentham Science Publishers Ltd.
`
`MPI EXHIBIT 1017 PAGE 1
`
`MPI EXHIBIT 1017 PAGE 1
`
`MPI EXHIBIT 1017 PAGE 1
`
`

`

`1400 Current Pharmaceutical Design, 2001, Vol. 7, No. 14
`
`Daniel J. Drucker
`
`Fig. (1). The biological actions of GLP-1.
`
`Glucagon-like peptide-1 is derived from proglucagon by tissue-specific posttranslational processing. The actions of GLP-1 are
`shown below the proglucagon molecule.
`
`transmission of the nutrient signal to the secretory
`apparatus of the gut L cell remain unclear [13].
`Although the majority of gut endocrine cells are
`located in the distal ileum and colon, GLP-1
`secretion occurs within minutes of nutrient
`ingestion, implying the existence of a proximal-
`distal intestinal loop for rapid transmission of
`nutrient-induced secretory
`signals
`from
`the
`duodenum and jejunum to the distal ileum and
`colon. GLP-1 release appears biphasic, with a
`rapid early response (mediated by humoral or
`neural mechanisms) followed by a more delayed
`response (direct nutrient contact with the distal L
`cells). Various mediators of this proximal-distal
`axis have been proposed based on rodent studies,
`including
`gastrin-releasing
`peptide
`and
`gastrointestinal inhibitory peptide (GIP) [14-16].
`Taken
`together,
`these findings suggest
`that
`identification of nutrient components that function
`as potent GLP-1 secretagogues represents a useful
`strategy for enhancing GLP-1 activity in vivo.
`
`Among various nutrients examined, fatty acids
`and dietary fiber up regulate both proglucagon
`mRNA transcripts and GLP-1 secretion in the
`rodent gastrointestinal
`tract [17-22]. Luminal
`glucose, peptones, and fatty acids increase GLP-1
`secretion from the isolated rat ileum [23]. The
`rapid rise of circulating levels of GLP-1 within
`minutes of food ingestion has stimulated inquiry
`into the endocrine and neural mediators, activated
`in the proximal gut, that signal the distal ileum and
`colon to release GLP-1.
`
`Neuromedin C [24], calcitonin gene-related
`peptide [15, 25] acetylcholine and muscarinic
`agonists [26, 27], GIP [14, 28-30] and gastrin-
`releasing peptide [16, 31] stimulate GLP-1
`secretion; the latter two peptides have been
`identified as putative peptide mediators of the
`proximal to distal signal in rodents [14, 16, 24].
`The stimulatory effects of GIP on GLP-1 release
`from canine
`ileal cells was
`inhibited by
`somatostatin and the protein kinase A inhibitor H-
`89 [28]. Consistent with a physiological role for
`gastrin-releasing peptide in the regulation of GLP-
`1 release, mice with inactivation of the GRP gene
`exhibit defective glucose-stimulated insulin release
`in association with reduced glucose-stimulated
`GLP-1 and insulin secretion [32].
`Adrenaline, acting through the b 2-adrenergic
`receptor stimulates GLP-1 release from
`the
`perfused rat ileum [33]. The importance of the
`vagus nerve for transmission of the proximal-distal
`secretory signal has been demonstrated in studies
`examining the effect of ganglionic blockade or vagal
`transection, maneuvers which
`significantly
`diminish GLP-1 secretion
`in
`rodents
`[30].
`Pharmacological or surgical selective hepatic
`branch vagotomy significantly attenuates GIP-
`stimulated increases in GLP-1 secretion [30].
`Furthermore, bilateral subdiaphragmatic vagotomy
`abolishes fat-stimulated intestinal PGDP secretion
`in the rat [30]. Somatostatin-28 exerts inhibitory
`effects on GLP-1 secretion, as somatostatin
`immunoneutralization increases GLP-1 release in
`
`MPI EXHIBIT 1017 PAGE 2
`
`MPI EXHIBIT 1017 PAGE 2
`
`MPI EXHIBIT 1017 PAGE 2
`
`

`

`Therapeutic Agents for the Treatment of Diabetes
`
`Current Pharmaceutical Design, 2001, Vol. 7, No. 14 1401
`
`the perfused porcine ileum preparation [24].
`Similarly, galanin also inhibits intestinal GLP-1
`release [34, 35]. In contrast, although insulin
`inhibits pancreatic glucagon
`secretion
`and
`biosynthesis, a direct role for insulin in the
`regulation of gut GLP-1 secretion remains unclear.
`
`PHYSIOLOGY OF GLP-1 ACTION
`
`Original concepts of GLP-1 action focused
`primarily on its role as a meal-stimulated incretin
`that
`functioned by potentiation of glucose-
`stimulated insulin release from the islet b
` cell
`following nutrient ingestion [36-39]. Accordingly,
`administration of exogenous GLP-1 immediately
`prior to a meal would be predicted to mimic the
`incretin-like actions of endogenous GLP-1 and
`control postprandial glycemic excursion. A large
`body of evidence from animal and human studies
`has shown that GLP-1 exerts multiple effects that
`serve to lower blood glucose independent of its
`actions on the islet b cell. It is now clear that GLP-
`1 potently inhibits gastric emptying [40-45] and
`glucagon secretion [46-50], additional actions that
`lower glucose in rodent and human studies. Indeed,
`the potent inhibition of gastric emptying might be
`predicted to reduce the rate of nutrient absorption
`and decrease the requirement for insulin secretion
`from the islet b
` cell [51].
`
`Continuous intravenous or subcutaneous GLP-
`1 infusion is effective in controlling blood glucose
`around the clock, and not just following meal
`ingestion [4, 6, 52]. Continuous subcutaneous
`infusion of GLP-1 for 48 hours in human subjects
`with type 2 diabetes lowered fasting and meal-
`related plasma glucose and reduced appetite [53].
`Furthermore, injection of subcutaneous GLP-1
`three
`times daily
`immediately before meals
`increased insulin, lowered glucagon, and decreased
`blood glucose in patients with early
`type 2
`diabetes over a 3 week study period [54].
`Encouragingly, GLP-1 also improved postprandial
`glycemic control in a similar experimental design
`over a 3 week period in 5 patients with poorly
`controlled diabetes [1].
`
`The results of short-term studies have shown
`that GLP-1 retains its glucose-lowering potency in
`human subjects after 7 days of continuous
`infusion. Nevertheless, infusion studies with the
`native molecule have shown a reduction of glucose-
`
`lowering effectiveness with increasing duration of
`GLP-1 infusion, suggesting that degradation of the
`intact peptide to GLP-19-36amide may potentially
`limit its sustained activity in this setting [4, 55,
`56]. Hence it seems clear that long acting GLP-1
`analogues or more stable formulations would
`exhibit considerable advantages over native GLP-1
`for achieving prolonged reduction of blood glucose
`over long periods of time.
`
`The physiological importance of GLP-1 for
`glucoregulation has been defined in experiments
`employing
`receptor
`antagonists,
`immunoneutralizing antisera, and knockout mice.
`In human subjects, GLP-1 administration reduces
`gastric emptying which may paradoxically reduce
`meal-stimulated insulin secretion [51]. Both rodent
`and human studies employing GLP-1 antagonists
`reveal an essential role for GLP-1 in the control of
`postprandial
`nutrient
`disposal
`and
`insulin
`secretion. Infusion of GLP-1 immunoneutralizing
`antisera or the GLP-1 receptor antagonist exendin
`(9-39) increased glycemic excursion and decreased
`insulin secretion in baboons, rats and human
`subjects [57-60]. Surprisingly, GLP-1 action is
`also essential for control of fasting glycemia and
`glucose clearance following non-enteral glucose
`challenge
`in mice
`[61, 62]. These
`latter
`observations are
`likely attributable
`to
`the
`importance of GLP-1 for both basal b
` cell function
`and for glucagon secretion. The comparatively
`modest degree of glucose intolerance observed in
`GLP-1R-/- mice is accounted for
`in part by
`compensatory up-regulation of GIP secretion and
`enhanced sensitivity to GIP action [63]. In
`contrast
`to
`the role of GLP-1 for glucose
`homeostasis following both enteral and non-enteral
`glucose challenge, the role of GIP appears more
`restricted as GIP regulates glucose absorption and
`glycemic excursion only following enteral glucose
`challenge [62, 64].
`
`DIPEPTIDYL PEPTIDASE IV
`
`As GLP-1 degradation represents a significant
`obstacle to the use of the native peptide for the
`chronic treatment of diabetic patients, the rapid
`enzymatic
`inactivation of
`the
`two naturally
`ocurring forms of GLP-1 has been carefully
`studied. Both full length bioactive GLP-17-36amide
`and GLP-17-37 (Fig. 2) are degraded within seconds
`of their release by the gut endocrine cell [65]. As
`
`MPI EXHIBIT 1017 PAGE 3
`
`MPI EXHIBIT 1017 PAGE 3
`
`MPI EXHIBIT 1017 PAGE 3
`
`

`

`1402 Current Pharmaceutical Design, 2001, Vol. 7, No. 14
`
`Daniel J. Drucker
`
`GLP-1 contains an alanine at position 2, it is an
`excellent substrate for the enzyme dipeptidyl
`peptidase IV (also known as the transmembrane
`protein CD26), leading to the generation of GLP-
`19-36amide and GLP-19-37 [56, 66-68]. Although
`these N-terminally truncated peptides have been
`shown to be weak antagonists of GLP-1 action in
`vitro [69],
`the physiological significance of
`circulating GLP-19-36amide and GLP-19-37 remains
`unclear.
`
`Fig. (2). Strategies for enhancing GLP-1 action in diabetic
`patients.
`
`GLP-1 is synthesized in and secreted from gut endocrine
`cells and acts on distant target tissues, including islet b
`cells. DP IV=dipeptidyl peptidase IV.
`
`In the absence of plasma, GLP-1 is a fairly
`stable peptide, as incubations at up to 55oC for 72
`h did not result in significant peptide degradation
`[70]. Furthermore, storage of the peptide at 4oC
`for 11 months did not result in significant peptide
`degradation as assessed by high pressure liquid
`chromatography [70]. In contrast, intravenous or
`subcutaneous infusion of GLP-1 is associated with
`rapid degradation of the full-length bioactive
`peptide in both normal and diabetic subjects, and
`the in vivo elimination half life of GLP-1 in human
`subjects is estimated to be approximately 90
`seconds [56].
`
`Inhibition of the serine protease dipeptidyl
`peptidase IV (DP IV) appears to represent a
`useful strategy for enhancing the bioactivity of
`GLP-1 in vivo [71]. DP IV is a widely expressed
`soluble and membrane-associated enzyme present
`in many tissues including the kidney, lung, liver,
`pancreas and intestine, and is highly expressed on
`
`both lymphocytes and endothelial cells [72]. The
`expression of DP IV on vascular endothelial cells
`that surround the GLP-1-producing gut endocrine
`cell, taken together with the expression of DP IV
`in gut epithelium [73], provides an explanation for
`the finding that over 50% of GLP-1 leaving the
`intestinal venous circulation has already been
`degraded at the N-terminus [65]. Unlike other
`endocrine systems such as the parathyroid cell
`that degrades intracellular parathyroid hormone,
`GLP-1 is stored within gut endocrine cells as
`primarily intact biologically active GLP-17-36amide
`and GLP-17-37.
`
`DP IV inhibitors represent effective glucose-
`lowering compounds in vivo. The DP IV inhibitor
`valine-pyrrolidide prevented degradation of intact
`GLP-1 and potentiated the action of exogenously
`administered GLP-1, leading to enhanced glucose
`clearance and increased insulin secretion in the
`non-diabetic pig [74]. Valine-pyrrolidide also
`increased levels of glucose-stimulated GLP-1 and
`improved insulin secretion and glucose tolerance in
`glucose intolerant C57BL/6J mice [75]. Similar
`results were obtained in Zucker fatty rats using the
`inhibitors Ile-thiazolidide [76] and NVP-DPP728
`[77]. Hence, the available data from short term
`studies clearly demonstrates that inhibition of DP
`IV activity is an effective method for improving
`glucose tolerance via potentiation of
`incretin
`action. Whether long-term inhibition of DP IV
`activity will result in sustained improvement in
`glycemic control is a subject of ongoing current
`investigation.
`
`The glucose lowering properties of DP IV
`enzyme inhibitors are clearly not attributable
`solely to reduced degradation of GLP-1. Gastric
`inhibitory polypeptide, secreted from duodenal K
`cells in a nutrient-dependent manner, is a potent
`stimulator of glucose-dependent insulin release
`that is also inactivated by DP IV cleavage [66, 67].
`Furthermore, DP IV inhibitors potently lower
`blood glucose in mice with complete inactivation
`of GLP-1 receptor signaling,
`likely due
`to
`potentiation of the bioactivity of insulinotropic
`DP IV substrates such as GIP [78].
`
`The consequences of inactivating mutations of
`the DP IV gene has been reported in two different
`animal models, the Fischer 344 DP IV mutant rat
`[79] and the CD26-/- mouse [78]. The Fischer 344
`DP IV deficient rat expresses a DP IV mRNA
`
`MPI EXHIBIT 1017 PAGE 4
`
`MPI EXHIBIT 1017 PAGE 4
`
`MPI EXHIBIT 1017 PAGE 4
`
`

`

`Therapeutic Agents for the Treatment of Diabetes
`
`transcript that encodes a mutation at position 633
`of the enzyme in the catalytic site, leading to
`defective processing and activity of the enzyme
`[79-81]. Remarkably, the levels of circulating
`GLP-1 and GLP-1 action are not perturbed in the
`Fischer 344 mutant
`rat, whereas glucose-
`stimulated GIP and GIP action are diminished, for
`reasons that remain unclear [82]. In contrast,
`inactivation of
`the murine DP IV gene by
`homologous recombination results in apparently
`normal mice with normal fasting glucose but
`enhanced glucose clearance following oral glucose
`challenge [78]. Consistent with
`the central
`importance of DP IV for incretin action, the levels
`of bioactive GLP-1 and GIP are
`increased
`following glucose administration in DP IV-/- mice
`[78].
`
`DP IV is also known as the lymphocyte cell
`surface membrane-associated peptidase CD 26, a
`molecule that regulates chemokine cleavage and T
`cell responses to antigen stimulation. CD26 was
`originally identified as an adenosine deaminase
`binding protein [83]. CD26, herein referred to as
`DP IV, cleaves peptides with an alanine or proline
`at position 2. Numerous chemokines are substrates
`for DP IV. In some instances, DP IV cleavage
`appears to have no effect on chemokine activity
`[84]. In other studies, DP IV may act as a
`costimulatory molecule for T cell activation, and
`DP
`IV processing may yield N-terminally
`modified chemokines with novel biological
`activities [85-87]. Inhibition of DP IV activity
`with relatively specific inhibitors may reduce T
`cell activation and hence DP IV is thought to exert
`immunomodulatory properties in vitro and in vivo
`[84].
`
`A large number of CNS and gut regulatory
`peptides including NPY, GHRH, GIP, PYY,
`PACAP and GLP-2 are also substrates for DP IV
`activity [84]. Accordingly, the use of DP IV
`inhibitors for potentiation of GLP-1 activity is
`likely to be associated with reduced degradation of
`numerous bioactive peptides and chemokines.
`These considerations suggest that inhibition of DP
`IV activity for the treatment of diabetes may be
`associated with additional biological consequences
`beyond simple potentiation of incretin (GLP-1 and
`GIP) action. Hence it seems prudent to assess
`immune function and additional physiological
`endpoints in both short and long term studies of
`DP IV inhibitors in vivo.
`
`Current Pharmaceutical Design, 2001, Vol. 7, No. 14 1403
`
`Several lines of evidence support a role for one
`or more DP IV-related enzymes in the cleavage of
`substrates exhibiting alanine or proline at position
`2. For example, the levels of incretin hormones are
`actually lower than normal in the DP IV mutant rat
`[82]. Furthermore, residual DP IV-like activity has
`been detected in plasma from DP IV-mutant rats
`[88], and on DP
`IV-negative human T
`lymphoblastoid cells [89]. These findings may be
`explained in part by evidence that DP IV appears
`to be a member of an expanding enzyme family,
`whose related members that share overlapping
`substrate
`specificity
`include attractin
`[90],
`fibroblast activation protein
`[91], quiescent
`peptidyl peptidase [92], N-acetylated alpha-linked
`acidic dipeptidase II [93], DPP6 [94], and DPP8
`[95]. The contribution if any of these related
`enzymes
`to
`incretin degradation, and
`their
`comprehensive
`substrate
`specificity
`profiles
`remains unclear and requires further investigation.
`
`GLP-1 DEGRADATION AND CLEARANCE
`
`store GLP-1
`cells
`gut L
`Although
`as
`intact GLP-17-36amide,
`a
`predominantly
`substantial proportion of degraded GLP-19-36amide
`is detected even in the intestinal venous circulation
`[65]. GLP-1 also undergoes endoproteolysis and is
`a substrate
`for membrane-associated neutral
`endopeptidase (NEP) 24.11 [96, 97]. Despite the
`importance
`of
`enzymatic
`degradation
`for
`termination
`of GLP-1
`activity,
`additional
`mechanisms such as renal metabolism of GLP-1
`account for substantial clearance of the peptide
`from the circulation [98-100]. The liver and lung
`also contribute to removal of GLP-1 from the
`circulation [99]. A small amount (14%) of glycated
`GLP-1 has been detected in mouse intestinal
`extracts and glycation of GLP-1 in vitro at
`position 7 impaired GLP-1-stimulated insulin
`release from BRIN-BD11 rat insulinoma cells in
`vitro
`[101]. Nevertheless,
`the
`biological
`significance of GLP-1 glycation for development
`of a GLP-1-based therapeutic remains uncertain.
`
`THE GLUCAGON-LIKE
`RECEPTOR
`
`PEPTIDE
`
`1
`
`The rat GLP-1 receptor was identified through
`expression cloning in 1992 and is a 463 amino acid
`member of
`the G protein coupled receptor
`
`MPI EXHIBIT 1017 PAGE 5
`
`MPI EXHIBIT 1017 PAGE 5
`
`MPI EXHIBIT 1017 PAGE 5
`
`

`

`1404 Current Pharmaceutical Design, 2001, Vol. 7, No. 14
`
`Daniel J. Drucker
`
`superfamily [102]. Transfection of the receptor
`cDNA into COS cells yields high affinity binding
`sites for GLP-1 with a Kd of 0.6 nM, comparable
`to the Kd observed for GLP-1 binding to rat
`insulinoma INS-1 cells [103]. Glucagon displaces
`GLP-1 binding only at concentrations of ~ 1 uM.
`The GLP-1 receptor exhibits amino acid identity
`not only with the glucagon and GLP-2 receptors
`[104, 105], but also with receptors for secretin
`(40% identity), parathyroid hormone (32.4%), and
`calcitonin (27.5%) [102]. The human GLP-1
`receptor exhibits 90%
`identity with
`the rat
`receptor, and binds GLP-1 with comparable
`affinity, KD=0.5 nM [106, 107]. Although the
`ligand binding pocket of the GLP-1R has not been
`defined,
`the purified N-terminal extracellular
`domain competes for GLP-1 receptor binding in
`vitro [108]. The human GLP-1 receptor has been
`localized to chromosome 6p21 [109], however
`linkage studies have not demonstrated significant
`association between type 2 diabetes or obesity and
`inheritance
`of
`specific GLP-1
`receptor
`polymorphisms [110-112].
`
`The GLP-1 receptor is expressed in the lung,
`stomach, pancreas, heart, and brain [102, 113,
`114]. In situ hybridization studies localized GLP-
`1R transcripts to the gastric pits, duodenal crypts
`and pancreatic
`islets [113]. GLP-1R mRNA
`transcripts have also been detected in the jejunum,
`ileum and colon of rats and mice [114, 115].
`Despite
`suggestions
`that additional GLP-1
`receptors may exist that mediate one or more ‘non-
`classical’ actions of GLP-1, RNA hybridization
`and RT-PCR studies have not detected evidence
`for a
`second GLP-1
`receptor
`[113-115].
`Furthermore, cloning of human GLP-1 receptors
`from lung, brain, heart and pancreas has revealed
`that all GLP-1 receptor cDNAs isolated to date
`exhibit identical amino acid sequences [102, 107,
`116, 117]. Similarly, GLP-1 binding and GLP-1
`actions are completely eliminated in GLP-1R-/-
`mice [61].
`
`Although the majority of islet GLP-1 receptor
`expression is localized to b
` cells, GLP-1 receptor
`expression was detected in isolated rat islet a
` cells
`by RT-PCR, and 20% of dispersed islet alpha cells
`and 76% of islet delta cells exhibited GLP-1
`receptor
`immunopositivity [118]. In contrast,
`other studies failed to detect evidence for GLP-1R
`expression in isolated rat islet a cells using
`Western blot analysis [119]. It is not clear whether
`
`transcriptional regulation represents a major locus
`for control of GLP-1R expression. A small but
`significant increases in the levels of GLP-1R
`mRNA transcripts was observed when rat islets
`were cultured under high glucose (20 mM)
`conditions. In contrast to the glucagon receptor,
`GLP-1R
`transcripts were not
`regulated by
`activators of
`the cAMP-dependent pathway,
`however both glucagon and GLP-1R mRNA
`transcripts were negatively regulated by 10 nM
`dexamethasone [120].
`
`The rat and human GLP-1 receptors are
`coupled
`to adenylate cyclase and
`cAMP
`formation, with the human receptor exhibiting an
`EC50 for cAMP formation of 93 pM [107]. The
`GLP-1 receptor also mediates a cAMP-dependent
`increase in free cytosolic calcium in studies of islet
`cells and transfected COS-7 cells [106, 121-123].
`Although increased cAMP formation is a common
`feature of GLP-1 receptor activation, several
`studies demonstrate downstream GLP-1 actions
`are mediated by PKA-independent signaling
`pathways [124]. For example, in some cell lines
`and in Xenopus oocytes, GLP-1R activation
`increases
`inositol
`triphosphate-dependent
`intracellular Ca2+ mobilization
`in a PKA-
`independent manner [125, 126]. Similarly, GLP-1
`effects on immediate early gene expression were
`markedly attenuated by
`the calcium channel
`blocker nifedipine
`in
`islet cell
`lines [127].
`Furthermore, the effects of GLP-1 on DNA
`synthesis, and induction of (PDX-1) DNA binding
`activity in beta (INS-1)-cells appears
`to be
`mediated through
`the phosphatidylinositol 3-
`kinase-dependent pathway
`[128]. Moreover,
`disruption of GLP-1R signaling in mouse b
` cells
`results in abnormal glucose-stimulated calcium
`oscillations that are not corrected by addition of
`cAMP agonists in vitro [129].
`
`Understanding the conditions and mechanisms
`underlying GLP-1
`receptor
`desensitization
`represents an important question with clinical
`relevance. Exposure of cells expressing a
`transfected or endogenous GLP-1R to GLP-1
`results
`in decreased expression of plasma
`membrane associated GLP-1R and
`receptor
`internalization [130, 131]. In contrast, the GLP-1
`receptor antagonist exendin (9-39) did not affect
`cell surface expression of the GLP-1R or receptor
`endocytosis [130]. In normal islet cells, rodents,
`and human subjects, GLP-1 and GIP exhibit
`
`MPI EXHIBIT 1017 PAGE 6
`
`MPI EXHIBIT 1017 PAGE 6
`
`MPI EXHIBIT 1017 PAGE 6
`
`

`

`Therapeutic Agents for the Treatment of Diabetes
`
`modest but detectable additive effects on insulin
`secretion
`[132-134].
`Supraphysiological
`concentrations of GLP-1
`(100 nM)
`induce
`reversible homologous desensitization on hamster
`insulinoma cells. In contrast, pretreatment with
`comparable concentrations of GIP or glucagon did
`not affect the insulin secretory response to GLP-1
`[135]. Similarly, pre-exposure of rat and mouse
`islet cells to 10 nM GLP-1 or the phorbol ester
`PMA induced desensitization of subsequent GLP-
`1 stimulated cAMP accumulation [136, 137]. The
`protein kinase C inhibitor RO-318220 inhibited
`desensitization induced by PMA but not by GLP-
`1 [136]. The importance of phosphorylation of
`specific amino acids in the carboxyterminal region
`of
`the GLP-1R was
`illustrated by studies
`demonstrating marked
`attenuation of both
`homologous
`and
`PMA-induced GLP-1R
`desensitization following mutation of specific
`serine doublets in the cytoplasmic tail [138, 139].
`
`STRUCTURAL DETERMINANTS OF GLP-1
`ACTION
`
`The amino acid sequences of glucagon, GLP-1,
`exendin-4, and GLP-2 are 100% conserved at
`positions 1, 4, 6, 22, 26, and 27, and GLP-1 and
`exendin-4 exhibit amino acid identity at 16 of 30
`positions in the native GLP-17-36amide molecule.
`The structure of GLP-1, as characterized in studies
`using 2D 1H NMR, is similar to glucagon, as GLP-
`1 exhibits an N-terminal random coil (residues 1-
`7), and two helical segments (7-14 and 18-29)
`separated by a linker region (15-17) [140]. Despite
`only 53% identity with GLP-1, the lizard peptide
`exendin-4 is a highly potent agonist at the GLP-1
`receptor both in vitro and in vivo [107, 141, 142].
`In contrast, although glucagon and GLP-1 exhibit
`identity at 14 amino acids, glucagon and GLP-1 do
`not exhibit cross-reactivity at their respective
`receptors
`at
`physiologically
`relevant
`at
`concentrations
`[102]. However
`higher
`concentrations, 3 x 10-7 M glucagon is a weak
`agonist at the GLP-1 receptor [70]. Furthermore,
`the GLP-1 receptor antagonist exendin (9-39)
`inhibited glucagon-stimulated cAMP formation in
`purified rat b
` cells at glucagon concentrations of
`10-8 M [143]. The N-terminus of GLP-1 is
`unlikely
`to be
`responsible
`for differential
`recognition of the GLP-1 compared to the glucagon
`receptor given the strong sequence conservation in
`this region, with identity at 10/14 N-terminal
`
`Current Pharmaceutical Design, 2001, Vol. 7, No. 14 1405
`
`amino acids [144]. The carboxy-terminus of GLP-
`1 is essential for ligand binding, as deletion of these
`sequences results in peptides that do not recognize
`the GLP-1 receptor [70, 145] and are partially or
`completely
`inactive
`in
`assays measuring
`stimulation of cAMP formation and
`insulin
`secretion [146]. Similarly, removal of the N-
`terminal histidine at position 1 significantly
`attenuates the insulinotropic action of GLP-1 [70,
`146].
`
`Initial studies of the structural determinants of
`GLP-1 action included analyses of specific amino
`acids substitutions via site-directed mutagenesis.
`The results of studies using alanine-substitutions
`demonstrated that side chains in positions 7, 10,
`12, 13, 15, 28, and 29 are critically important for
`ligand-receptor interaction as alanine substitutions
`lead to a significant loss in receptor affinity and in
`some instances changes in peptide conformation as
`assessed by circular dichroism spectroscopy [144,
`147]. The positive charge of the imidazole side
`chain at position 7 accounts for the importance of
`the position histidine for receptor binding [148].
`Swapping selective residues from growth hormone
`releasing hormone into the GLP-1 backbone
`identified amino acid positions 1, 10, 15, and 17 as
`essential for receptor binding and cAMP activation
`in a RINm5F cell assay [149].
`
`The functional domains of GLP-1 have also
`been studied using domain swap experiments and
`chimeric peptides. Substitutions of glucagon for
`GLP-1
`amino
`acid
`sequences
`at
`the
`carboxyterminus produced significant decreases in
`affinity for the GLP-1 receptor [150]. Conversely,
`transfer of GLP-1 carboxyterminal sequences to
`the glucagon molecule improved affinity for the
`GLP-1 receptor. Substitutions at the N-terminus
`(positions 2, 3, 10, and 12) were generally well
`tolerated, and produced only modest decreases in
`receptor affinity [150]. Substitution of GIP
`sequences for GLP-1 sequences in the N-terminus
`retained binding specificity for the GLP-1 receptor
`but led to a significant diminution of GLP-1
`receptor
`binding
`[151].
`Carboxyterminal
`substitutions also diminished GLP-1 receptor
`binding, but were better tolerated compared to
`substitutions at the N-terminus. Consistent with
`the results described above, substitutions at
`positions 13 (tyrosine) and 15 (glutamic acid) were
`poorly tolerated, leading to decreased GLP-1
`receptor affinity [151].
`
`MPI EXHIBIT 1017 PAGE 7
`
`MPI EXHIBIT 1017 PAGE 7
`
`MPI EXHIBIT 1017 PAGE 7
`
`

`

`1406 Current Pharmaceutical Design, 2001, Vol. 7, No. 14
`
`Daniel J. Drucker
`
`GLP-1 ANALOGUES
`
`EXENDIN-4
`
`IV
`for DP
`GLP-1 analogues engineered
`resistance in vitro generally exhibit longer plasma
`half lives and enhanced bioactivity in vivo. Clearly
`preferred analogues are
`those which exhibit
`preserved
`to
`enhanced
`receptor
`binding
`concomitant with increased resistance to DP IV-
`mediated degradation. GLP-1 derivatives with a
`glycine or a
`-aminoisobutyric acid at position 8
`exhibit both DP IV-resistance and retained binding
`affinity
`for
`the GLP-1
`receptor
`[152]. A
`considerable number of GLP-1 derivatives have
`been developed that exhibit enhanced stability and
`sustained bioactivity in vivo. Derivativization of
`GLP-1 by addition of fatty acid moieties at the
`carboxyterminus promotes albumin binding leading
`to molecules with prolonged duration of action. In
`contrast, fatty acid derivatization at
`the N-
`terminus is less well tolerated, leading to reduced
`receptor binding and loss of potency [153]. GLP-1
`peptides acylated with simple fatty acids alone, or
`with a L-glutamoyl spacer or diacids exhibit an
`enhanced negative charge, leading to greater
`albumin binding and
`improved solubility at
`physiologically relevant pH [153]. In contrast to
`native GLP-1, such derivatives exhibit half-lives of
`at least 9 hours and potently lower blood glucose
`in vivo.
`
`GLP-1 analogues with position 1 substitutions
`may also exhibit enhanced resistance to DP IV-
`mediated degradation. N-methylated, desamidated,
`and
`imidazole-lactic acid-substituted GLP-1
`molecules were resistant to DP IV degradation and
`both N-methylated and and imidazole-lactic acid-
`substituted GLP-1 exhibited preserved affinity and
`cAMP stimulatory activity
`in studies using
`RINm5F cells [154]. Similarly, GLP-1 analogues
`engineered
`for
`resistance
`to DP
`IV with
`substitutions at position 2 alone, or positions 2
`and 8 exhibited DP IV resistance but preserved
`insulinotropic properties in isolated rat islets [155,
`156]. Similarly, GLP-1 derivatives with N-terminal
`substitutions at positions 8 with either alpha-
`aminoisobutyric acid, threonine, glycine or serine
`exhibited DP IV resistance, with
`the alpha-
`