The Multiple Actions of GLP-1 on the Process of
`Glucose-Stimulated Insulin Secretion
`
`Patrick E. MacDonald,1 Wasim El-kholy,1 Michael J. Riedel,2 Anne Marie F. Salapatek,1
`Peter E. Light,2 and Michael B. Wheeler1
`
`The physiological effects of glucagon-like peptide-1
`(GLP-1) are of immense interest because of the poten-
`tial clinical relevance of this peptide. Produced in intes-
`tinal L-cells through posttranslational processing of the
`proglucagon gene, GLP-1 is released from the gut in
`response to nutrient ingestion. Peripherally, GLP-1 is
`known to affect gut motility, inhibit gastric acid secre-
`tion, and inhibit glucagon secretion. In the central
`nervous system, GLP-1 induces satiety, leading to re-
`duced weight gain. In the pancreas, GLP-1 is now known
`to induce expansion of insulin-secreting ␤-cell mass, in
`addition to its most well-characterized effect: the aug-
`mentation of glucose-stimulated insulin secretion.
`GLP-1 is believed to enhance insulin secretion through
`mechanisms involving the regulation of ion channels
`(including ATP-sensitive Kⴙ channels, voltage-depen-
`dent Ca2ⴙ channels, voltage-dependent Kⴙ channels,
`and nonselective cation channels) and by the regulation
`of intracellular energy homeostasis and exocytosis.
`The present article will focus principally on the mecha-
`nisms proposed to underlie the glucose dependence of
`GLP-1’s insulinotropic effect. Diabetes 51 (Suppl. 3):
`S434 –S442, 2002
`
`Glucagon-like peptide 1 (GLP-1) is a potent in-
`
`cretin hormone produced in the L-cells of the
`distal ileum and colon. In the L-cells, GLP-1 is
`generated by tissue-specific posttranslational
`processing of the proglucagon gene (1). Nutrients, includ-
`ing glucose, fatty acids, and dietary fiber, are all known to
`upregulate the transcription of the gene encoding GLP-1,
`
`From the 1Departments of Medicine and Physiology, University of Toronto,
`Toronto, Ontario, Canada; and the 2Department of Pharmacology, University
`of Alberta, Edmonton, Alberta, Canada.
`Address correspondence and reprint requests to Michael B. Wheeler,
`Department of Physiology, University of Toronto, 1 Kings College Circle,
`Toronto, ON, Canada, M5S 1A8. E-mail: michael.wheeler@utoronto.ca.
`Received for publication 18 March 2002 and accepted in revised form 17
`May 2002.
`AKAP, A-kinase anchoring protein; [Ca2⫹]i, intracellular concentration of
`Ca2⫹; CICR, Ca2⫹-induced Ca2⫹ release; CNS, central nervous system; cPKA,
`catalytic subunit of protein kinase A; DP-IV, dipeptidyl-peptidase IV; ERK,
`extracellular signal-related kinase; GEF-II, guanine nucleotide exchange fac-
`tor II; GI, gastrointestinal; GIP, glucose-dependent insulinotropic peptide;
`GLP-1, glucagon-like peptide-1; GSIS, glucose-stimulated insulin secretion;
`HSL, hormone-sensitive lipase; IP3, inositol triphosphate; KATP, ATP-sensitive
`K⫹ channel; KCa, Ca2⫹-sensitive voltage-dependent K⫹ channel; Kv, voltage-
`dependent K⫹ channel; MAPK, mitogen-activated protein kinase; NSCC,
`nonspecific cation channel; PDX-1, pancreatic duodenal homeobox-1; PI3-K,
`phophatidylinositol 3-kinase; PKA, protein kinase A; PKC␨, protein kinase C␨;
`Po, open probability; SU, sulfonylurea; VDCC, voltage-dependent Ca2⫹ chan-
`nel.
`The symposium and the publication of this article have been made possible
`by an unrestricted educational grant from Servier, Paris.
`
`and they can stimulate the release of this hormone (2).
`Although the majority of L-cells are located in the distal
`ileum and colon, the levels of GLP-1 rise rapidly upon food
`ingestion. It is now well accepted that nutrients, princi-
`pally sugars and fats, liberate GLP-1 and GLP-1–releasing
`factors, including glucose-dependent insulinotropic pep-
`tide (GIP), gastrin-releasing peptide, and selective neural
`regulators that also stimulate GLP-1 secretion (rev. in 1–3).
`Upon its release, GLP-1 affects multiple target tissues
`throughout the body, actions thought to be mediated by a
`single G-protein– coupled receptor isoform. GLP-1 recep-
`tor transcripts and/or protein have been identified in
`several tissues, including pancreatic islets, lung, gastroin-
`testinal (GI) tract, and the central nervous system (CNS)
`(2,3). More questionable is the expression of functional
`GLP-1 receptors in liver and skeletal muscle tissues, where
`gene expression has been detected (4). GLP-1’s ability to
`augment insulin release in a glucose-dependent manner is
`its most well-characterized physiological effect and one of
`its most promising characteristics from a clinical perspec-
`tive.
`
`THE ACTIONS OF GLP-1
`GLP-1 in the pancreas: insulin secretion and ␤-cell
`mass. There are several known and speculated pancreatic
`functions for GLP-1. The GLP-1 receptor is expressed in
`␤-cells, where its activation is proposed to have multiple
`acute and long-term actions (3). With respect to ␤-cell
`function, GLP-1 rapidly and potently stimulates insulin
`secretion, a well-known action that will be discussed
`below. However, GLP-1 also stimulates insulin gene tran-
`scription,
`islet cell growth, and neogenesis, additional
`potentially important functions that may be clinically
`relevant for the treatment of diabetes. In mice, GLP-1
`stimulates ␤-cell proliferation (5), and in rat ␤-cell lines, it
`stimulates DNA synthesis through a phosphatidylinositol
`3-kinase (PI3-K)-dependent pathway (6). Effects on ␤-cell
`insulin gene transcription and proliferation are proposed
`to occur via the upregulation of the transcription factor
`pancreatic duodenal homeobox-1 (PDX-1) (6 – 8). PDX-1
`translocation to the nucleus of RIN 1046-38 cells was also
`shown to be dependent on cAMP/protein kinase A (PKA)
`(9). The actions of GLP-1 on proliferation may involve the
`PI3-K downstream targets extracellular signal-related ki-
`nase (ERK) 1/2 and p38 mitogen-activated protein kinase
`(MAPK) (10). These studies have also demonstrated that a
`GLP-1–induced activation of protein kinase C␨ (PKC␨) is
`implicated in ␤-cell proliferation (10). Thus, at least two
`signaling pathways (PI3-K activation of ERK, MAPK, or
`
`S434
`
`DIABETES, VOL. 51, SUPPLEMENT 3, DECEMBER 2002
`
`MPI EXHIBIT 1090 PAGE 1
`
`

`

`PKC and cAMP/PKA-mediated translocation of PDX-1) can
`be implicated in GLP-1’s effects on ␤-cell proliferation.
`GLP-1 can also increase islet and ␤-cell mass by promoting
`cellular differentiation. Studies have shown that GLP-1 can
`induce the differentiation of rat (ARIP), pancreatic exo-
`crine (AR4J), and human islet progenitor cells and impli-
`cate PDX-1 activation as a potential common mediator
`(3,11). A hint of the clinical relevance of GLP-1 treatment
`to circumvent the loss of ␤-cell capacity can be seen in
`vivo. Prolonged administration of GLP-1 to obese diabetic
`(db/db) mice stimulated insulin secretion, improved blood
`glucose/glycemic excursions, and also enhanced ␤-cell
`neogenesis and islet mass (7). Similar improvements in
`glycemic control and ␤-cell mass could be seen in mice
`with a partial pancreatectomy (12) and in neonatal mice
`treated with streptozotocin (13). Studies examining GLP-1
`loss of function in GLP-1 receptor⫺/⫺ mice demonstrated
`alterations in pancreatic insulin content and insulin gene
`transcription in one study (14) and islet size and compo-
`sition in another (15). It is now thought that the phenotype
`of the GLP-1 receptor⫺/⫺ mouse is likely diminished by
`compensatory upregulation of other signaling peptides,
`including GIP. Together, this work suggests that future
`therapies targeting the GLP-1 signal transduction path-
`ways may be used to improve ␤-cell capacity by improving
`␤-cell mass and secretion.
`GLP-1 in the periphery: gut motility and insulin
`sensitivity. GLP-1 appears to exert a wide range of
`extrapancreatic effects. It participates in an ileal break
`phenomenon by playing an inhibitory role in gastric emp-
`tying and small intestinal transit (16,17). GLP-1 decreases
`gastric motility via direct effects on gastric smooth muscle
`and also inhibits postprandial acid secretion (4). It also
`decreases small intestine movement through inhibition of
`smooth muscle activity, resulting in an overall reduction in
`the absorption of nutrients from the GI tract (4). Reduced
`motility likely causes less severe postprandial glucose
`fluctuations and reduces the need for a large and rapid
`postprandial insulin response. GLP-1 also appears to im-
`prove insulin sensitivity and glucose uptake of both human
`and rat adipose tissue and skeletal muscle (4). Several
`studies suggest that GLP-1 may directly enhance glucose
`disposal in an insulin-independent fashion, although this
`may also result from the overall inhibition of glucagon
`secretion (3). GLP-1 also appears to have some cardiovas-
`cular effects, in part through actions on the CNS, as the
`administration of GLP-1 by intravenous or intracerebro-
`ventricular injection increases heart rate and blood pres-
`sure in rats, an effect that can be reversed by application of
`the antagonist exendin 9-39 or bilateral vagotomy, but not
`when GLP-1 is given by peripheral
`injection (3). The
`physiological relevance of this latter effect is unknown.
`GLP-1 in the CNS: control of appetite and weight.
`GLP-1 has profound effects on feeding behavior. Although
`these actions of GLP-1 could be in part related to its effects
`on intestinal motility, they also appear to involve direct
`effects on hypothalamic feeding centers as GLP-1 recep-
`tors are found in specific nuclei within the hypothalamus
`(18). Acute administration of GLP-1 in humans and ro-
`dents induces satiety and decreases caloric intake (19 –
`21). Administration of the GLP-1 antagonist exendin 9-39
`abrogates the effect of GLP-1 and can itself promote
`
`P.E. MACDONALD AND ASSOCIATES
`
`weight gain (22). Long-term exendin 4 treatment of Zucker
`rats reduces food intake and decelerates weight gain (23).
`Interestingly, GLP-1 receptor⫺/⫺ mice do not become
`obese, possibly because this peptide is not essential for
`weight regulation or because compensatory mechanisms
`are upregulated (24). In humans with type 2 diabetes,
`short-term GLP-1 or exendin 4 administration curbs appe-
`tite and food intake in addition to its insulinotropic actions
`(19,25), suggesting long-term delivery would promote
`weight loss in these patients. The ability of GLP-1 analogs
`to promote weight loss and improve ␤-cell function could
`prove to be ideal for the treatment of type 2 diabetes.
`Type 2 diabetes and GLP-1 receptor agonists. It is well
`known that type 2 diabetes is characterized by defects in
`both insulin secretion and in peripheral insulin sensitivity
`(26). Current treatments for the ␤-cell defect include the
`sulfonylurea (SU) drugs, which were the first therapy
`targeted against insufficient insulin secretion. However,
`these compounds promote insulin secretion independent
`of blood glucose and can therefore cause hypoglycemia
`(27). In addition, as clearly demonstrated in the U.K.
`Prospective Diabetes Study, the ability of SUs to stimulate
`insulin secretion decreases over time, likely reflecting a
`deterioration in ␤-cell function (28). For these reasons,
`there is a need for the development of new drugs to treat
`the ␤-cell defect component of type 2 diabetes. Meglitinide
`analogs belong to a newer family of insulin secretagogues
`with actions similar to the SUs, e.g. ATP-sensitive K⫹
`(KATP) channel inhibition, but have distinct pharmacoki-
`netic and pharmacodynamic properties (rapid onset/short
`duration) (27). The properties of GLP-1 described above,
`including glucose-dependent stimulation of insulin secre-
`tion and the expansion of ␤-cell mass, coupled with the
`inhibition of glucagon secretion and food intake, suggest
`that it would greatly complement current ␤-cell therapies.
`Trials with GLP-1 in diabetic patients have shown it to
`stimulate insulin secretion, inhibit gastric emptying, lower
`circulating glucagon, and improve overall glycemic control
`through both intravenous and subcutaneous injection (1).
`Two major drawbacks in the use of GLP-1 therapy are
`related to its rapid inactivation in circulation and its
`delivery via injection. The issue regarding rapid inactiva-
`tion has been addressed using two promising strategies.
`The first involves the development and use of GLP-1
`analogs resistant to proteolytic cleavage by the enzyme
`dipeptidyl-peptidase IV (DP-IV) (1). Currently, several
`pharmaceutical companies are testing their GLP-1 analogs
`in a variety of models, including humans. The second
`strategy involves the neutralization of the DP-IV enzyme
`itself. Thus far, DP-IV inhibition has been shown to pro-
`long the action of GLP-1, improve glycemia in diabetic
`models, and more recently to delay the onset of diabetes in
`Zucker diabetic fatty rats (29 –32). Although both strate-
`gies are promising, several obstacles still remain. With
`respect to DP-IV inhibition, it is not clear what long-term
`effects such a therapy would have, given that this enzyme
`cleaves a number of biologically active peptides with
`various biological activities, including glucagon, vasoac-
`tive intestinal peptide, glucose-dependent insulinotropic
`polypeptide, neuropeptide Y, and substance P, for exam-
`ple (33). With respect to GLP-1 therapy, it is unlikely that
`it will gain widespread acceptance in the diabetes commu-
`
`DIABETES, VOL. 51, SUPPLEMENT 3, DECEMBER 2002
`
`S435
`
`MPI EXHIBIT 1090 PAGE 2
`
`

`

`GLP-1 AND THE ␤-CELL
`
`FIG. 1. The ionic mechanism (KATP channel– dependent) of
`stimulus-secretion coupling in the pancreatic ␤-cell. 1: Glucose
`enters the ␤-cell via GLUT2 transporters. 2: It is metabolized to
`produce an increase in the intracellular ATP-to-ADP ratio. 3:
`This results in closure of KATP channels. 4: This also results in
`membrane depolarization. 5: Finally, it results in the opening of
`VDCCs. 6: Entry of Ca2ⴙ is the main trigger for exocytosis of
`insulin-containing granules. 7: Also triggered by membrane
`depolarization, Kv and KCa channels open. 8: This repolarizes
`the membrane. 9: Repolarization leads to closure of VDCCs,
`limiting Ca2ⴙ entry and insulin secretion.
`
`coexpressed with the GLP-1 receptor in a mammalian cell
`line. The physiological consequences of GLP-1–facilitated
`KATP channel closure would be to 1) augment the excit-
`ability of cells already above the threshold for insulin
`
`FIG. 2. Effects of GLP-1 on ␤-cell excitability and native and recombi-
`nant KATP channels. A: Membrane potential was recorded from an
`INS-1 cell using the patch-clamp technique in current-clamp mode.
`Application of GLP-1 (20 nmol/l) causes membrane depolarization and
`an increase in the action potential firing rate. A family of ionic currents
`elicited from an individual INS-1 cell using the patch-clamp technique
`in voltage-clamp mode is shown in B, upper panel. Cells were held at
`ⴚ80 mV and stepped at 10-mV intervals from ⴚ120 to 40 mV for 200 ms.
`Recordings were made with 5 mmol/l Kⴙ in the bath solution. In the
`lower panel of B, GLP-1 is shown to inhibit KATP channels in INS-1 cells.
`To measure inward KATP current, cells were held at 0 mV and stepped
`to ⴚ100 mV every 10 s. GLP-1 (20 nmol/l) was superfused over cells for
`3–5 min until a steady-state current was obtained. Tolbutamide sensi-
`tivity was confirmed as a marker of KATP current (traces not shown). C:
`GLP-1 inhibits KATP channels in a recombinant system. tsA201 cells
`were transiently transfected with GLP-1R, Kir6.2, and SUR1 clones
`48 –72 h before recording. Current recordings were performed under
`symmetrical Kⴙ conditions (140 mmol/l Kⴙ in the bath solution). GLP-1
`(20 nmol/l) was applied to the cells in a similar manner to B, upper
`panel. Barium chloride (BaCl2, 2 mmol/l) was used in the recombinant
`system as a fully washable potassium channel blocker to assess the
`amount of recombinant KATP current present. The amphotericin-perfo-
`rated patch-clamp technique mode was used to measure whole-cell
`KATP currents and membrane potentials while maintaining the integrity
`of the intracellular environment.
`
`nity when effective oral agents are available. However, one
`can anticipate that a small molecule that holds some or all
`of the properties of native GLP-1 will eventually be devel-
`oped. To facilitate this process, it is important to under-
`stand how GLP-1 controls insulin secretion.
`
`GLP-1 AND INSULIN SECRETION
`Overview of the ATP-sensitive pathway. Glucose-stim-
`ulated insulin secretion (GSIS) is regulated by a number of
`ionic and nonionic signaling pathways, also known as the
`KATP-dependent and -independent pathways (34,35). The
`KATP-dependent mechanism of stimulus-secretion cou-
`pling is reviewed in Fig. 1. In general, the ␤-cell adapts
`insulin secretion to prevailing blood glucose levels
`through glucose metabolism. When glucose levels rise, the
`rate of glycolysis increases, which generates substrates
`(mainly pyruvate) for mitochondrial oxidative metabo-
`lism, the result of which is the generation of ATP, or more
`correctly an increase in the ATP-to-ADP ratio (36). This
`event provides the functional
`link between a glucose
`stimulus and insulin secretion. The increase in this ratio
`causes the closure of ␤-cell KATP channels, leading to
`plasma membrane depolarization, activation of voltage-
`dependent Ca2⫹ channels (VDCCs), and an increase in the
`intracellular concentration of Ca2⫹ ([Ca2⫹]i), the main
`trigger for insulin secretion. Repolarization of ␤-cells is
`likely mediated by voltage-dependent K⫹ (Kv) channels
`and Ca2⫹-sensitive voltage-dependent K⫹ (KCa) channels
`(37), which open in response to glucose-induced mem-
`brane depolarization to restore the outward flux of K⫹.
`GLP-1 is proposed to modulate GSIS by regulating the
`activity of several ion channels involved in KATP-depen-
`dent insulin secretion as well as steps distal to channel
`modulation.
`GLP-1 and ␤-cell KATP channels. One of the many
`observed cellular effects of GLP-1 is the inhibition of ␤-cell
`KATP channels (38 – 40). The resulting membrane depolar-
`ization induced by KATP channel closure initiates Ca2⫹
`influx through VDCCs and triggers the exocytotic release
`of insulin. Figure 2 shows the excitatory effect of GLP-1 on
`membrane potential and its inhibitory effect on both native
`KATP channel currents from INS-1 cells and currents
`mediated by recombinant KATP channels (SUR1/Kir6.2)
`
`S436
`
`DIABETES, VOL. 51, SUPPLEMENT 3, DECEMBER 2002
`
`MPI EXHIBIT 1090 PAGE 3
`
`

`

`release and 2) increase the percentage of ␤-cells actively
`secreting insulin at glucose concentrations normally sub-
`threshold for the release of insulin.
`The consensus view is that the inhibitory effect of GLP-1
`on KATP channels is cAMP/PKA-dependent (38 – 41), al-
`though one study using rat ␤-cells disagrees (42). This
`assertion by Suga et al. (42) is based on their finding that
`the specific PKA inhibitor Rp-cAMPS (100 ␮mol/l) was
`unable to prevent the cellular depolarization and reduction
`in the whole-cell KATP current elicited by GLP-1. The
`completeness of PKA inhibition by Rp-cAMPS should be
`questioned because in the same study, forskolin induced
`significant insulin secretion even in the presence of Rp-
`cAMPS. Secondly, Suga et al. suggest that GLP-1 causes a
`slight increase in the ATP sensitivity of the KATP channel
`such that at low micromolar ATP concentrations, the KATP
`channel will be more susceptible to closure. However, in
`the normal rat pancreatic ␤-cell, millimolar levels of ATP
`are present (43), and at these physiological ATP levels,
`very similar KATP channel open probability (Po) in the
`absence and presence of GLP-1 are predicted as follows.
`Our calculations indicate that with an intracellular [ATP]
`of 2 mmol/l, the KATP channel Po is reduced from 0.005 to
`0.003 in the presence of GLP-1. It is plausible that a
`leftward shift of ATP sensitivity does occur in the pres-
`ence of GLP-1. However, the observed magnitude of
`GLP-1–induced KATP current reduction seen in whole-cell
`patch-clamp recordings (Fig. 2) is likely too large to be
`accounted for solely by the small decreases in Po calcu-
`lated from the data of Suga et al. (42). Recent work from
`our laboratory has shown that the membrane-permeant
`specific PKA inhibitor H-89 (44) is capable of completely
`inhibiting KATP current reduction by GLP-1 (45). Moreover,
`others have shown similar results using Rp-8-Br-cAMPS, a
`more membrane-permeable analog of Rp-cAMPS (38,41).
`The actions of GLP-1 on KATP channels may also involve
`other signaling pathways because KATP channel inhibition
`by GLP-1 in mouse ␤-cells was demonstrated to be cal-
`modulin dependent, using the calmodulin inhibitors W-7
`and calmidazolium (46).
`The glucose dependency of GLP-1 actions has been well
`established, although the precise mechanisms for this
`dependence are unclear (41,47). However, the cellular
`actions of PKA on the KATP channel may provide a link
`between this kinase and the glucose sensitivity of GLP-1.
`Other groups have shown that addition of the catalytic
`subunit of PKA (cPKA) to excised patches containing KATP
`channels results in an augmentation of KATP current
`(39,48). Our laboratory has recently shown that the effect
`of cPKA on KATP current is dependent on ADP (45). When
`ADP levels are elevated, cPKA increases KATP channel
`current in a recombinant system, consistent with the
`results of Lin et al. (39). Conversely, as ADP levels are
`decreased, cPKA reduces KATP current (45). Physiologi-
`cally, this may result in a negligible enhancement of ␤-cell
`excitability when glucose levels are low (high [ADP]),
`whereas when glucose levels rise (low [ADP]), GLP-1–
`mediated closure of KATP channels, via a PKA-dependent
`pathway, leads to membrane depolarization and subse-
`quent increases in ␤-cell excitability. Special attention
`must be given to the cellular ATP-to-ADP ratio when
`considering KATP channel activity because it is changes in
`
`DIABETES, VOL. 51, SUPPLEMENT 3, DECEMBER 2002
`
`P.E. MACDONALD AND ASSOCIATES
`
`this ratio, more than simply changes in intracellular [ATP]
`per se, that govern the activity of KATP channels in the
`intact ␤-cell. Free ATP is highly buffered within the cell by
`membrane and cytosolic ATPases (43) and is predicted not
`to change significantly with increased glucose metabolism
`(36). In contrast, the reciprocal change in ADP, which is
`not buffered to the same extent, is more significant and
`results in a change in the ATP-to-ADP ratio (36). Indeed,
`the importance of ADP in controlling ␤-cell KATP channel
`activity has been demonstrated because mutations in the
`ADP-sensing region of the human KATP channel lead to
`uncontrolled insulin secretion and hypoglycemia (49). The
`molecular identity of the PKA phosphorylation site(s) is of
`significant importance and is still under investigation.
`Both the Kir6.2 and SUR1 subunits of the KATP channel
`contain putative target sequences for PKA-mediated phos-
`phorylation (39,48), and systematic mutation of these
`residues should clarify the relative contributions of these
`sites to the action of PKA on the KATP channel.
`GLP-1, VDCCs, and intracellular Ca2ⴙ stores. GLP-1
`has been shown to enhance currents through VDCCs in
`mouse, rat, and human ␤-cells (38,50 –52), although the
`magnitude of this effect varies and often does not reach
`statistical significance. In single human ␤-cells, GLP-1
`was shown to increase L-type VDCC activity and the
`amplitude of depolarization-evoked intracellular calcium
`transients, an effect which accounted for 40% of the
`increase in GLP-1–potentiated exocytosis (52). Although
`L-type VDCCs are classically regarded as the major regu-
`lators of Ca2⫹ influx leading to insulin secretion, ␤-cells
`are known to express multiple Ca2⫹ channel isoforms
`(53). Pereverzev et al. (54) reported recently that mice
`lacking the ␣1E-isoform of Cav2.3 have reduced glucose
`tolerance and diminished insulin responses to glucose.
`They speculate that G-protein regulation of this channel
`may modulate insulin secretion based on muscarinic ace-
`tylcholine receptor regulation of VDCCs in vitro (55).
`We have found in HIT-T15 insulinoma cells transfected
`with the GLP-1 receptor that GLP-1 causes an increase in
`voltage-dependent Ca2⫹ currents (see 56 and Fig. 3A). This
`is due, at least partly, to a leftward shift in the voltage
`dependence of activation reminiscent of the effect of
`VDCC phosphorylation by PKA (57,58). We also observed
`a rightward shift in the voltage dependence of steady-state
`inactivation such that in the presence of GLP-1,
`less
`channels were effectively inactivated at a given holding
`potential (50). This is supported by Britsch et al. (50), who
`suggest that GLP-1 treatment of mouse ␤-cells slows the
`inactivation of voltage-dependent Ca2⫹ currents. Addition-
`ally, GLP-1 led to an increase in intracellular calcium only
`after glucose addition, an effect that was blocked in part
`by VDCC antagonists (Fig. 3C). The ability of GLP-1 to
`enhance Ca2⫹ currents is, like the effect on KATP channels,
`cAMP dependent (51,52), based on the ability of Rp-
`cAMPS to prevent an increase in currents. Indeed, treat-
`rat ␤-cells with dibutyryl cyclic-AMP, a
`ment of
`membrane-permeable cAMP analog, replicated the effect
`of GLP-1 on Ca2⫹ currents (51). Additionally,
`in our
`studies (56), the VDCC response to GLP-1 was lost in
`HIT-T15 cells expressing a mutant GLP-1 receptor lacking
`critical residues required for coupling to adenylyl cyclase,
`whereas VDCC activity could still be enhanced by the
`
`S437
`
`MPI EXHIBIT 1090 PAGE 4
`
`

`

`GLP-1 AND THE ␤-CELL
`
`FIG. 3. GLP-1 enhances VDCC activity and [Ca2ⴙ]i responses to glucose
`in HIT-T15 cells overexpressing the wild-type GLP-1 receptor. HIT-T15
`cells overexpressing the wild-type GLP-1 receptor were voltage-
`clamped in the whole-cell configuration under conditions to measure
`voltage-dependent Ca2ⴙ currents, held at ⴚ70 mV and stepped from
`ⴚ60 to 60 mV in 10-mV steps for 250 ms. A: Representative traces for
`basal control and after sequential cumulative addition of GLP-1 (10ⴚ8
`mol/l) and nifedipine (10 ␮mol/l). Current-voltage relationships are
`shown in B. C: A representative trace of [Ca2ⴙ]i in wild-type GLP-1
`receptor–transfected HIT-T15 cell measured with Fura 2-AM after
`sequential cumulative addition of glucose (10 mmol/l), GLP-1, and
`nifedipine as in A and B.
`
`cAMP independent agonist BAYK8644. Recent evidence
`suggests that an A-kinase anchoring protein (AKAP),
`AKAP18, targets PKA to VDCCs and that this kinase may
`be involved in GLP-1 modulation of these channels (59).
`In addition to effects on VDCCs, GLP-1 can mobilize
`intracellular calcium stores in a cAMP-dependent manner
`(60,61), possibly contributing to the oscillatory [Ca2⫹]i
`response to GLP-1 seen in HIT-T15 cells (Fig. 3C) (62). Our
`studies suggest that GLP-1 application results in oscilla-
`tions in [Ca2⫹]i in HIT-T15 cells and that these oscillations
`are not abolished (although they are diminished in ampli-
`tude) by removal of extracellular Ca2⫹ or by VDCC block-
`ade (Fig. 3C). Indeed, in a number of cell types, Ca2⫹
`oscillations induced by an agonist are primarily caused by
`Ca2⫹ release through inositol trisphosphate (IP3) and/or
`ryanodine-sensitive stores (63– 65). In ␤-cells, GLP-1 mo-
`bilizes Ca2⫹ stores in large part by sensitizing the ryano-
`dine receptor (likely the type 2 isoform, RYR-2) to the
`process of Ca2⫹-induced Ca2⫹ release (CICR) (61,66).
`Several studies have demonstrated that GLP-1 can in-
`crease [Ca2⫹]i in a PKA-independent manner (67– 69). This
`mechanism has recently been ascribed to CICR from
`ryanodine-sensitive stores via cAMP-regulated guanine
`nucleotide exchange factor II (GEF-II or Epac2) and its
`interaction with either the Ras-related small G-protein
`Rap1 or with Rab3 small G-protein effector Rim2 (68). The
`importance of cAMP-GEF-II–Rim2 has been demon-
`strated, because inactivation of this complex (by antisense
`
`oligonucleotides or mutant constructs) attenuated the
`secretory response of mouse islets or MIN6 insulinoma
`cells to GLP-1 (67). Because the affinity of cAMP for PKA
`is much higher (⬃100 nmol/l) compared with cAMP–
`GEF-II (⬃10 mmol/l), it is interesting to speculate that the
`GLP-1–stimulated cAMP–GEF-II pathway might operate
`on a rise in local cAMP rather than global changes.
`Although GLP-1 receptor signaling stimulates IP3 produc-
`tion in GLP-1 receptor– expressing COS cells (70), a role
`for IP3-sensitive Ca2⫹ stores in global [Ca2⫹]i is in doubt
`because the GLP-1–stimulated IP3 production in primary
`␤-cells is reportedly minimal (71,72) and the IP3 receptor
`antagonist xestospongin C failed to block release of intra-
`cellular Ca2⫹ stores by forskolin treatment (68). However,
`IP3-regulated Ca2⫹ release from insulin granules has been
`suggested by the studies of Nakagaki et al. (62), who
`suggest that GLP-1 may uniquely regulate temporal and
`spatial release of intracellular calcium through local IP3
`signaling. Thus, GLP-1–mediated release of intracellular
`stores, along with potentiation of Ca2⫹ entry through
`VDCCs, likely contribute to the insulinotropic effect of
`GLP-1.
`GLP-1 and ␤-cell Kv channels. Voltage-dependent K⫹
`currents, such as those mediated by Kv or KCa channels,
`mediate repolarization of ␤-cells after a depolarizing stim-
`ulus, such as glucose (37). Recently, we reported that Kv1
`and Kv2 family channels regulate insulin secretion, be-
`cause dominant-negative functional knockout of either of
`these channel families enhanced GSIS (73). Kv2.1 channels
`mediate the majority of this effect (⬎60%), the mechanism
`of which involves enhanced glucose-stimulated membrane
`depolarization and Ca2⫹ entry (unpublished observations).
`Because ␤-cell Kv currents are potent glucose-dependent
`regulators of insulin secretion, we hypothesized that phys-
`iological secretagogues, such as GLP-1, may regulate Kv
`channel function. Indeed, we report elsewhere in this
`supplement that the GLP-1 receptor agonist exendin 4
`inhibits voltage-dependent outward K⫹ currents in rat
`␤-cells voltage-clamped in the whole-cell configuration by
`40% and significantly prolongs the time course of ␤-cell
`repolarization after transient depolarization by current
`injection. This compares to an 86% reduction in outward
`K⫹ currents achieved with the general Kv channel antag-
`onist tetraethylammonium. GLP-1 antagonized voltage-
`dependent outward K⫹ currents in rat ␤-cells in the
`absence of glucose. However, this effect may still contrib-
`ute to the glucose dependence of GLP-1’s insulinotropic
`effect, because Kv channels are not normally expected to
`be active until after a glucose-induced depolarization of
`the cell membrane (37). Additionally, and similar to the
`effect of GLP-1 on the other ion channels mentioned
`above, exendin 4 –mediated inhibition of ␤-cell Kv chan-
`nels is dependent on cAMP signaling. One recent study,
`however, suggested that cAMP signaling was not sufficient
`in itself to antagonize voltage-dependent K⫹ currents in an
`insulin-secreting cell line (INS-1) (74).
`Numerous studies have described effects of hormone-
`mediated alterations in voltage-dependent K⫹ currents,
`both excitatory and inhibitory. The best-characterized of
`these effects is the voltage-dependent K⫹ current down-
`regulation in lymphocytes and upregulation in cardiac
`myocytes (75,76). In both of these tissues, the cAMP/PKA
`
`S438
`
`DIABETES, VOL. 51, SUPPLEMENT 3, DECEMBER 2002
`
`MPI EXHIBIT 1090 PAGE 5
`
`

`

`signaling pathway has been implicated in the regulation of
`these channels (76,77). Reports suggest that cAMP can
`reduce voltage-dependent K⫹ currents in murine lympho-
`cytes (76) and a pituitary cell
`line (78) but enhance
`voltage-dependent K⫹ currents in cardiac myocytes (77), a
`finding that has been confirmed at the single-channel level
`in frog atrial myocytes (79) and the giant squid axon (80).
`Phosphorylation may occur directly on the channel, be-
`cause PKA phosphorylation of an atrial Kv channel near
`the NH2-terminus enhanced channel activity (81), and
`phosphorylation of Kv1 channel ␣-subunits regulates the
`extent of inhibition of these channels conferred by a
`regulatory ␤-subunit (82). Phosphorylation of ␤-subunits
`themselves may also modulate the regulatory interaction
`with pore-forming ␣-subunits (83). It has recently been
`demonstrated that regulation of a cardiac Kv channel
`(KvLQT) by cAMP requires the expression of AKAP15/18
`or AKAP79 (84). Additionally, an increase in voltage-
`dependent K⫹ current is implicated in epinephrine-in-
`duced inhibition of the glucose-dependent increase in
`[Ca2⫹]i in ob/ob and ⫹/⫹ mouse ␤-cells (85) since the
`effect was reversed by tetraethylammonium. Interestingly,
`the inhibitory effect of epinephrine on [Ca2⫹]i was also
`reversed by the adenylyl cyclase activator forskolin (85).
`Therefore, we believe there is mounting evidence to
`suggest that hormonal modulation of Kv currents is phys-
`iologically important. Specifically, GLP-1 inhibition of
`these currents is expected to lead to enhanced ␤-cell
`excitability.
`GLP-1 and other ␤-cell ion channels. Increases in
`intracellular cAMP have long been known to enhance Na⫹
`currents (86), an effe