`Copyright © 2001 PROUS SCIENCE
`CCC: 0377-8282/2001
`
`GLP-1 derivatives as novel compounds
`for the treatment of type 2 diabetes:
`selection of NN2211 for clinical development
`
`Review Article
`
`L. Bjerre Knudsen’, H. Agerso, C. Bjenning,
`S. Bregenholt, R.D. Carr, C. Godtfredsen,
`J.J. Holst', P.O. Huusfeldt, M.O. Larsen,
`P.J. Larsen*, P.F. Nielsen, U. Ribel, B. Rolin,
`J. Romer, J. Sturis, M. Wilken and
`P. Kristensen
`Novo Nordisk, Novo Park, DK-2760 Maaloev, Denmark;
`*Dept. of Medical Physiology, University of Copenhagen,
`DK-2200 Copenhagen, Denmark; *Present address:
`Lab. Obesity Research, CCRB, DK-2760 Ballerup, Denmark.
`* Correspondence
`
`found to also lower plasma glucagonin a glucose-depen-
`dent manner, decrease the rate of gastric emptying, pro-
`mote fullness/satiety and stimulate insulin biosynthesis,
`as well asproliferation of B-cells. All of these effects taken
`together position GLP-1 as an obvious drug candidate for
`the treatment of type 2 diabetes. Apart from the food
`intake lowering effect, the glucagonostatic effect is partic-
`ularly interesting as type 2 diabetic patients are charac-
`terized by increased plasma glucagonlevels, which again
`lead to an increased hepatic glucose output (4, 5).
`The natural hormone GLP-1 is found as both
`GLP-1(7-37) and GLP-1(7-36)amide. It is metabolized by
`dipeptidyl peptidase IV (DPP-IV) and rapidly cleared by
`the kidneys (6, 7). The primary metabolite, GLP-1(9-37)
`or GLP-1(9-36)amide, may even act as an antagonist (8).
`The plasmahalf-life of GLP-1 after i.v. administration is
`less than 2 min and 1-2 h after s.c. administration (9). The
`dynamic half-life of GLP-1 has been estimated to be 3h
`(10). Thus, the natural hormone is not very useful as a
`drug. No small-molecule agonists have been described
`for any membersof the glucagon-secretin G-protein cou-
`pled receptor family,
`to which the GLP-1 receptor
`belongs. Thus, in order to make a GLP-1-based drug, one
`must either make a protracted formulation of the natural
`hormone, design analogs or derivatives with improved
`pharmacokinetic properties or find a non-peptide agonist.
`This review describesthe biological aspects of GLP-1
`and gives a detailed description of a series of GLP-1
`derivatives designed for once-daily administration.
`
`Mechanism of action of GLP-1
`
`GLP-1 is an incretin, responsible for the difference in
`blood glucose levels obtained after oral as compared to
`
`MPI EXHIBIT 1031 PAGE 1
`
`CONTENTS
`
`cnx ome cos wre own ome a mx ey we ee ETE Be 677
`PTSGUGHOR:
`Mechanism of.action of GLP=1 «52:56 see0+waxensae ews 677
`Desired pharmacodynamicprofile of an effective
`GLPRT GIO iin ee veer run nce ene im oa Rue Re ies eum a 679
`Derivatives of GLP-1 for once-daily administration ...... 679
`Site of acylation ........ 0.0... e eee ee eee 680
`Compoundsderivatized on lysine 26 ............... 680
`Compoundsderivatived with different spacers ........ 680
`Biological characterization of NN2211
`.............. 681
`Conclusions ......... 0.60 cece cee eee 682
`References .........-. 2c e eee eee ete eee eee eae 682
`
`Introduction
`
`Current drugs for treatment of type 2 diabetes are
`classified into four major categories:
`insulin secreta-
`gogues (sulfonylureas and the shorter acting glinides),
`biguanides,
`insulin sensitizers and insulin (1). The first
`three groupsare orally active but are of limited efficacy.
`Insulin is very effective but has to be dosed four times
`daily for optimal efficacy, thus presenting the risk of seri-
`ous hypoglycemia to the patient. Furthermore, type 2 dia-
`betes is most commonly associated with obesity, and of
`the four drug groups listed, only the biguanides do not
`cause increases in body weight. Thereis, therefore, an
`increasing awareness of the need for new efficacious
`drugs for the treatment of type 2 diabetes, especially
`drugs that do not lead to an increase in body weight.
`Glucagon-like peptide-1 (GLP-1) is a peptide hormone
`derived from proglucagon which wasdiscovered in 1983
`(2, 3). GLP-1 wasfirst described as an incretin, a sub-
`stance released from the gastrointestinal system upon
`ingestion of food, promoting glucose-dependentinsulin
`release from the pancreatic B-cells. Later, GLP-1 was
`
`MPI EXHIBIT 1031 PAGE 1
`
`
`
`678
`
`GLP-1 derivatives for the treatment of type 2 diabetes
`
`intravenous administration of the same amount of glu-
`cose (11-14). Upon ingestion of a meal, GLP-1 is
`released from the L-cells in the intestine and stimulates
`insulin release via specific receptors on pancreatic
`B-cells.
`The GLP-1 receptor was cloned in 1992 (15) and has
`been shownto be present in numerous tissues with the
`highest expression levels in the B- and 5-cells of the pan-
`creas and in the lungs (16). Only one subtype is known.
`The receptor has also been found in parts of the gas-
`trointestinal tract and in many regions in the CNS, includ-
`ing the hypothalamic regions, as well as specific areas in
`the brain stem.
`Equally important to the insulinotropic action, GLP-1
`also potently inhibits glucagon secretion (17-20). The
`mechanism by which GLP-1 exerts its glucagonostatic
`effect is not fully understood. It has been suggested to be
`paracrine, via neighboring somatostatin cells. Because of
`the combination of
`increased insulin and decreased
`glucagon secretion, hepatic glucose production is effi-
`ciently decreased (21). In spite of the insulinotropic effect,
`GLP-1 infusions carried out during ingestion of a meal
`actually result in diminished insulin responses. This is
`due to an inhibition of gastric emptying effectively reduc-
`ing the delivery rate of ingested nutrients to the absorp-
`tive segments of the GI tract (22).
`GLP-1 has been investigated in a substantial number
`of small clinical trials and has been shownto efficiently
`lower blood glucosein patients with type 2 diabetes and
`also in patients with poor metabolic control, referred to as
`secondary failures to sulfonylurea treatment
`(9-14).
`GLP-1 is also capable of lowering blood glucosein type 1
`diabetic patients without
`residual beta-cell secretory
`capacity (23, 24), reflecting its effects on lowering of plas-
`ma glucagon andinhibition of gastric emptying. Indeed,
`glucagon antagonism alone has been suggested as an
`alternative treatment for type 2 diabetes but no drugs
`have reached the marketyet.
`One study has suggestedthat the effect of GLP-1 can
`decline overnight, perhaps as a result of receptor desen-
`sitization (25). However, these results are in contrast to
`another study which showed that a 7-day infusion of
`GLP-1 lowered blood glucose levels for the entire dura-
`tion of the study (26, 27).
`the
`Because of
`the glucose-dependent nature of
`insulinotropic and glucagonostatic actions of GLP-1, the
`glucose lowering effect of this hormoneis self-limiting
`and, therefore,
`it does not cause serious hypoglycemia
`regardless of dose (28). The mechanism underlying the
`glucose dependency is closely related to the glucose
`signaling events being dependent on a high ATP/ADP
`ratio (29). Numerous studies in the literature, where
`GLP-1 has been infused or injected subcutaneously in
`humans, support the glucose dependency of GLP-1’s
`actions (9-14, 23, 25). However, there are some reports
`that GLP-1 can lower blood glucose to below normo-
`glycemia (30, 31). This ability of GLP-1 to transiently
`lower blood glucose to subnormallevels is a natural con-
`sequenceofits effect on insulin secretion. The inactiva-
`
`tion time for insulin is considerable, and after GLP-1 stim-
`ulation there may be enough insulin around (and enough
`activated insulin receptors) to lower blood glucose tem-
`porarily, even if secretion of new insulin has ceased
`becauseofthe falling glucose levels. The overall conclu-
`sion, therefore, is that GLP-1 can cause blood glucoseto
`decrease temporarily to below normal, but never leads to
`profound and lasting hypoglycemia.
`Apart from its direct effect on insulin secretion, GLP-1
`has been shownto increasetherate of insulin biosynthe-
`sis (32, 33) and to restore thefailing ability of the beta-cell
`to respond to glucosein old rats (34-36). Thus, it is spec-
`ulated that GLP-1 may preventthe transition from IGT to
`full-blown diabetes (37), presumably because B-cell func-
`tion is preserved. Also, several publications now show a
`direct effect of GLP-1 compounds on growth andprolifer-
`ation of B-cells in animals (38-40), as well as an ability to
`prevent apoptosis in B-cells (41). GLP-1 has even been
`found to be able to cause differentiation of pancreatic
`stem cells into functional B-cells (42).
`GLP-1 has also been reported to exert peripheral
`effects by promoting glucose uptake and glycogen
`storage in fat cells, skeletal muscle andliver cells in rats
`(43). A few studies in humans have confirmed that there
`maybe a peripheral effect (44, 45) while others have not
`(46, 47). Mostlikely, if this effect exists, it is quantitatively
`unimportant.
`All of the effects described above are aimed at direct-
`ly lowering blood glucose. However, GLP-1 in peripheral
`circulation has also been shown to lower food intake in
`both rodents and humans, indirectly leading to improved
`glucose controi via loss of body weight (48-53). Several
`receptors may be involved in the food intake lowering
`effect. First, centrally acting GLP-1 has been suggested
`to have a role as a physiological satiety factor (53-57) via
`receptors localized in arcuate hypothalamic areas.
`However,
`it has been shown that these receptors are
`probably not involved in mediating the food intake lower-
`ing effect of peripherally administered GLP-1 (52).
`Binding sites in the area postrema and the subfornical
`organ have been shownto be accessible for GLP-1 in the
`peripheral circulation, and these representlikely sites for
`mediating the food intake effect of peripheral GLP-1 com-
`pounds (58). As mentioned above, GLP-1 inhibits gastric
`emptying, and the vagus nerve has been shownto be
`involved in mediating this effect (22, 59-62). The areasin
`the brainstem containing the GLP-1 receptors are known
`to receive afferent
`inputs from gastrointestinal organs.
`Thus, the mostlikely mechanism for the food intake low-
`ering effect of a peripherally administered GLP-1 com-
`pound that does not pass the blood-brain barrieris that a
`sensation of fullness is obtained via small amounts of
`receptor in the intestine which then projects via the vagus
`nerve to receptors in the area postrema and the subforni-
`cal organ or directly via receptors in the blood-brain bar-
`rier free areas of the CNS.
`Obese subjects have been shownto have anattenu-
`ated GLP-1 release in response to meals (63, 64), sug-
`gesting that a decreased peripheral GLP-1 signal may
`
`MPI EXHIBIT 1031 PAGE 2
`
`MPI EXHIBIT 1031 PAGE 2
`
`
`
`Drugs Fut 2001, 26(7)
`
`679
`
`contribute to the developmentof obesity. Also, type 2 dia-
`betes patients have been shown to have a reduced GLP-
`1 release in response to food ingestion (65) and they
`have a defect in their response to the other incretin hor-
`mone, GIP (11). Thus, there is a very good rationale for
`treating type 2 diabetes and obesity with a GLP-1-based
`drug as a replenishment treatment.
`A significant inhibitory effect of GLP-1 on waterintake
`has been observed together with profound stimulation of
`diuresis in rats (54). The profound diuretic response to
`acute GLP-1
`administration is due to natriuresis.
`However, chronic administration of a GLP-1 derivative
`was shown notto lead to disturbed water homeostasis in
`rats (52). In humans, GLP-1 does not acutely affect water
`homeostasis (66).
`Last, GLP-1 potently inhibits pentagastrin and meal-
`induced gastric acid secretion and pancreatic enzyme
`secretion (59, 60). Along with the inhibited gastric empty-
`ing, these effects of GLP-1 are most
`likely mediated
`vagally (61). The effect of GLP-1 gastric acid secretion
`mayalso point to a role for GLP-1 in protection against
`gastric ulceration.
`A recent very comprehensive review has a thorough
`description of the biology of GLP-1 (67).
`In conclusion, the mode of action of GLP-1 seemsto
`be ideal for the treatment of type 2 diabetes, especially
`obese type 2 diabetes patients. GLP-1 compoundsglu-
`cose-dependently stimulate insulin secretion and inhibit
`glucagon secretion, decrease gastric emptying and medi-
`ate increased fullness and/or appetite, and directly stimu-
`late growth and rescue of pancreatic B-cells. With this
`important spectrum of effects of GLP-1, it is conceivable
`that a GLP-1-based compound could potentially be more
`effective than any of the current blood glucose lowering
`drugs available today.
`
`Desired pharmacodynamic profile
`of an effective GLP-1 drug
`
`Several studies have shown that GLP-1 can normal-
`ize blood glucosein patients and remainsvery effective in
`the so-called sulfonylurea failure patients, but it has a
`very short duration of action. It has been suggested that
`repeated subcutaneous administration of GLP-1 (9-10,
`25) or buccal tablets (68) might be an effective treatment.
`However, a study by Larsen et al. (26) showsthat a long
`dynamic half-life is needed in order to maintain good glu-
`cose control on a GLP-1-based drug.
`In this study,
`patients were administered GLP-1 by infusion for either
`16 h or 24 h a dayfor 1 week. Patients in the 24-h infu-
`sion group had a muchbetter glucose control than those
`in the 16-h infusion group (26). In patients in the 16-h infu-
`sion group, blood glucose rapidly reverted to preinfusion
`levels after GLP-1 infusion. Thus, the difference between
`these two groupsstrongly indicates that GLP-1 has to be
`present in an effective dose at all times. At least three
`daily administrations of natural GLP-1 by buccal or s.c.
`administration would be needed in order to obtain this
`kind of pharmacodynamic profile.
`
`As mentioned above, GLP-1 is metabolized rapidly
`by DPP-IV and the major metabolite, GLP-1(9-36)amide
`or GLP-1(9-37) is inactive or may even act as an antago-
`nist (8). Thus, analogs of GLP-1 stabilized against meta-
`bolic breakdown have been proposedas possible drugs
`(69-72), However, as GLP-1 is also cleared very rapidly
`from the kidneys, such drugs would only have half-lives of
`around 5 min after i.v. administration (69). Even though
`s.c. administration would improve the dynamic half-lives
`of such compounds, probably to a range of 3-5 h (72, 73),
`they still require a tremendous formulation challenge if
`multiple daily dosing is to be avoided.
`The only knownside effect of GLP-1 treatmentis nau-
`sea, occurring at plasma concentrations high enough to
`markedly slow gastric emptying. Thus, the most optimal
`pharmacokinetic profile of a GLP-1 drug from this per-
`spective would again be a long half-life with as few as
`possible peak concentrations. Thus, compounds adminis-
`tered several times a day have a disadvantage over com-
`poundswith only once-daily administration.
`Derivatizing GLP-1 with long fatty acids has been
`shown to result in half-lives exceeding 10 h in healthy
`humans as well as type 2 diabetic patients after s.c.
`administration. Such derivatives of GLP-1 could be dosed
`once daily and provide active plasma concentrations
`throughout the day.
`
`Derivatives of GLP-1 for once-daily administration
`
`Fatty acid derivatization has been used successfully
`to protract the action of insulin by facilitating binding to
`plasma albumin (74-77). The same principle has been
`used to design derivatives of GLP-1 with half-lives longer
`than 10 h, thereby being optimal for once-daily adminis-
`tration (78). Fatty acids or fatty diacids, optionally extend-
`ed with a “spacer” between the epsilon-amino group of
`the lysine side chain and the carboxyl group of the fatty
`acid, were used. Acylation with simple fatty acids increas-
`es the net negative charge of the resulting molecule with
`one (by blocking the epsilon-amino group of the lysine),
`whereas peptides acylated with a L-glutamoyl-spacer or
`with diacids provides a further increase of the negative
`charge. The addition of a negative charge to the acylated
`molecule is expected to improve solubility at physiological
`pH.
`
`The amino acid sequence of GLP-1 can be seen in
`Figure 1. GLP-1(7-37) and close analogsthereof, as well
`as an extended molecule, were derivatized on position 8,
`18, 23, 26, 27, 34, 36 or 38 with fatty acids and optional-
`ly a spacer (Table 1). The structure-activity relationship
`(SAR) of the compounds wasinvestigated using a func-
`tional assay employing the cloned human GLP-1 receptor
`expressed in baby hamster kidney cells. All compounds
`tested were full agonists and were shownto selectively
`activate the GLP-1 receptor. Table II shows plasma haif-
`lives after s.c. administration to pigs for a selection of very
`potent compounds4, 5, 7, 8, 18, 20 and 21. Ail com-
`poundsacylated with a fatty acid equal to or longer than
`
`MPI EXHIBIT 1031 PAGE 3
`
`MPI EXHIBIT 1031 PAGE 3
`
`
`
`680
`
`GLP-1 derivatives for the treatment of type 2 diabetes
`
`Compoundsderivatized on lysine 26
`
`Lys-Gly-Arg-Gly*’
`
`His?-Ala-Glu-Gly'°-Thr-Phe-Thr-Ser-Asp'5-Val-Ser-Ser-Tyr-
`Leu?°-Glu-Gly-Gin-Ala-Ala?>-Lys-Glu-Phe-lle-Ala°°-Trp-Leu-Val-
`
`Fig. 1. Amino acid sequence of GLP-1.
`
`The potency of the compounds was comparable when
`looking at a series of different length diacids (14, 15,
`Table !) or fatty acids with the same spacer (5, 16-18).
`Within the y-Glu spacer monoacid series (5, 16-18),
`derivatization with a C18 acid (16, 194 pM)led to a sig-
`nificant loss of activity compared to C16 (5, 68 pM), C14
`(17, 22 pM) and C12 (18, 27 pM). Within the diacid series
`12 carbon atoms were considerable protracted compared
`(14, 15), the diacid could be no longer than a C14 (15, 72
`to native GLP-1, which had a half-life after s.c. adminis-
`pM) before a loss in potency (14, 154 pM), compared to
`tration of only 1.2 h. Bioavailability was measured for
`the +-Glu spacer monoacid series (17, 18, 22-27 pM) was
`selected compounds and was shownto be on the order
`seen.
`In earlier studies from our group and others,
`of 50% and above. Figure2illustrates the dramatic dif-
`attempts were made to modify the amino terminus of
`ference in plasma half-lives between GLP-1 and three
`GLP-1 in order to make the molecule more resistant to
`potent acylated compounds5, 7 and 8.
`enzymatic breakdown (69, 80-81). Desamino His’ repre-
`sents one of the more potent suggestions to a modifica-
`tion giving metabolic stability (81). Nevertheless, as seen
`when comparing 19 (687 pM) to 5 (68 pM), considerably
`more potent compoundscould be obtained by not modi-
`fying the N-terminus when a combination with acylation
`wasdesired. This could be caused both by the position of
`the fatty acid and by the modified histidine.
`
`Site of acylation
`
`Many different positions in the C-terminal part of
`GLP-1 could be derivatized with quite long fatty acids,
`visualized with compounds 3-9 (EC,,, 30-121 pM) without
`affecting the potency. Binding affinity was not measured
`for these compoundsastheyall bind to albumin as part
`of their mechanism of protraction and it has not been pos-
`sible to set up a reproducible binding assay without albu-
`min. Derivatizing amino acids in the N-terminalpart of the
`peptide, as exemplified in compound 2 (1260 pM), led to
`a substantial loss of potency. This is in agreement with
`earlier findings showing the importance of the N-terminal
`region for affinity (79).
`
`Compoundsderivatized with different spacers
`
`The y-Glu spacer is optically active. Thus,it presents
`a greater analytical challenge when upscaling the com-
`pounds for good manufacturing production guidelines.
`Wetherefore investigated other spacers without optical
`
`Table |: GLP-1 compounds andtheir potency measured using the cloned human GLP-1 receptor expressed in baby hamsterkidneycells.
`
`Cod.
`
`Parent peptide
`
`Acylsite
`
`1
`2
`3
`4
`5
`6
`7
`8
`3
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`GLP-1(7-37)
`K8R26,34-GLP-1(7-37)
`K18R26,34-GLP-1(7-37)
`K23 R26,34-GLP-1(7-37)
`R34-GLP-1(7-37)
`K27R26,34-GLP-1(7-37)
`R26-GLP-1(7-37)
`K36R26,34-GLP-1 (7-36)
`R26,34-GLP-1(7-38)
`R34-GLP-1(7-37)
`R34-GLP-1(7-37)
`R34-GLP-1(7-37)
`R34-GLP-1 (7-37)
`R34-GLP-1 (7-37)
`Des-amino-H7R34-GLP-1(7-37)
`R34-GLP-1 (7-37)
`R34-GLP-1 (7-37)
`R34-GLP-1 (7-37)
`
`-
`K8
`K18
`K23
`K26
`K27
`K34
`K36
`K38
`K26
`K26
`K26
`K26
`K26
`K26
`K26
`K26
`K26
`
`Acyl substituent
`None
`¥-Glu-C16
`-Glu-C16
`y-Giu-C16
`y-Glu-C16
`y-Glu-C 16
`y-Giu-C16
`y-Glu-C16
`y-Glu-C16
`C16-diacid
`C14-diacid
`
`y-Glu-C18
`y-Glu-C 14
`-Glu-C12
`y-Glu-C 16
`GABA-C16
`B-Ala-C16
`Iso-Nip-C16
`
`Potency (EC,,, pM)
`55+ 19
`1260 + 210
`35.2 + 6.2
`30.1 + 3.3
`61.0+7.1
`36.3 + 0.3
`121 + 26
`36.4 + 2.1
`53.0 + 2.8
`154 + 66
`72 +07
`
`194 + 24
`22.0+7.1
`27.3+84
`687 + 129
`84.4 + 22.1
`11343
`410 + 120
`
`Abbreviations used for acyl groups in lysine-N-e-acylated peptides: y-Glu-C12 = y-L-glutamoyl(N-a-dodecanoyl); y-Glu-C14 = y-L-gluta-
`moyl(N-c-tetradecanoyl); y-Glu-C16 = y-L-glutamoyl(N-o-hexadecanoy)); y-Glu-C18 = y-L-glutamoyl(N-a-octadecanoyl); C14-diacid = w-
`carboxytridecanoyl; C16-diacid = w-carboxypentadecanoyl; GABA-C16 = y-amino- butyryi(N-y-hexadecanoyl); Iso-Nip-C16 = 1-(hexade-
`canoyl)piperidyl-4-carboxy. Data are given as mean + SD of 2 individual experiments with triplicate samples. (Reprinted in part with
`permission from J Med Chem 2000, 43; 1664-1669. Copyright 2000 American Chemical Society.)
`
`MPI EXHIBIT 1031 PAGE 4
`
`MPI EXHIBIT 1031 PAGE 4
`
`
`
`Drugs Fut 2001, 26(7)
`
`681
`
`Table Il: Plasma half-lives in pigs of GLP-1 and selected potent
`acylated compounds.
`
`Cpd.
`
`1(GLP-1)
`4
`§
`7
`8
`18
`20
`21
`
`Plasmat' (h)
`
`1.2
`20+2
`14+2
`13
`1241
`156+3
`3144
`8.8+1
`
`The half-lives were calculated from individual pigs after a single
`s.c. injection. Each compound wasinjected in 2 pigs. Data are
`shown as mean + SD.For(7), half-life could only be calculated
`from 1 pig. (Reprinted in part with permission from J Am Chem
`2000, 43: 1664-1669. Copyright 2000 American Chemical
`Society.)
`
`10000
`
`2 1000
`
`100
`
`10
`
`s®
`
`= 85
`
`oO
`
`ly stable compound. Compound 5 had anin vitro half-life
`of 20 h, whereas compoundsderivatized on position 34
`(7) and position 36 (8) had in vitro half-lives of 6.3 and 6.9
`h, respectively (82). For comparison, GLP-1(7-36)amide
`hadanin vitro half-life of 0.12 h. Amino acid substitutions
`in position 8 can give better metabolic stability against
`DPP-IV. However, since quite a substantial protection
`against DPP-IV was obtained by acylation alone, and
`since any amino acid substitution posesa risk of immuno-
`genicity, and since compound 5 was equipotent with
`GLP-1 and had the half-life required to be dosed once
`daily, compound 5 was selected for clinical development
`under the name of NN2211.
`The long half-life in pigs, adequate for once-daily
`administration, has as mentioned above been confirmed
`in man, where the half-life was determined to be 12 + 2h
`following a single injection to healthy humans and 10.0 +
`3.5 h following a single injection to type 2 diabetic
`subjects (83-84).
`
`Biological characterization of NN2211
`
`As shown above, NN2211 was a full agonist with
`equal potency to GLP-1 on the cloned human GLP-1
`receptor. Its mechanism of protraction involves binding to
`albumin, metabolic stability towards DPP-IV and slow
`release from the injection site. Because of the binding to
`albumin, NN2211 appears less potent than GLP-1 in the
`presence of high concentrations of albumin. NN2211 has
`been shown to be selective, as it has no activity on the
`very closely related glucagon receptor, and also did not
`bind to any receptors in a broad Panlabs receptor binding
`screen (data not published).
`NN2211 has been studied extensively in pigs, repre-
`senting the only model where s.c. administration normal-
`ly results in pharmacokinetic profiles comparable to those
`of human subjects.
`In STZ-induced glucose intolerant
`minipigs, clamp experiments showed that NN2211
`increased glucose utilization while at
`the same time
`increasing insulin secretion and inhibiting glucagon
`release, both in a glucose-dependent manner. Gastric
`emptying was decreased as well (85).
`In STZ-induced
`diabetic minipigs, 0.004 mg/kg/day (0.28 mg/75 kg)
`NN2211 lowered blood glucose to near normal
`levels
`after 2-4 weeks of once-daily administration and normal-
`ized prandial glucose tolerance (86).
`In order to obtain full efficacy in rodents, all pharma-
`cology studies have been carried out with higher doses
`than in pigs and monkeys. Rodents, in general, seem to
`have a lower sensitivity to GLP-1 compounds and a more
`aggressive metabolism of GLP-1.
`In normal rats, 0.2 mg/kg twice-daily NN2211 has
`been shown to significantly decrease body weight by
`15%, inhibit food and water intake and stimulate diuresis
`(52). In subchronic experiments, water intake was com-
`pensatorily increased to account for the increased diure-
`sis. The mechanism for the decreased food intake appar-
`ently does not involve hypothalamic GLP-1 receptors, as
`
`MPI EXHIBIT 1031 PAGE 5
`
` 0
`
`10
`
`20
`
`30
`
`40
`
`50
`
`60
`
`Time (hours)
`
`Fig. 2. Pharmacokinetic profile of selected compoundsafter s.c.
`administration to pigs. 1 (GLP-1) (MM), 5 (C ), 7 (@) and 8 (L). Two
`pigs were dosed with compoundsafter which the immunoassay
`wasperformedin duplicate. Data are expressed as mean + SD.
`(Reprinted with permission from J Med Chem 2000, 43: 1664-
`1669. Copyright 2000 American Chemical Society.)
`
`activity. A GABA spacer (20, 84 pM) gave a compound
`with equalaffinity to the y-Glu spacer (5, 68 pM). A B-Ala
`spacer reduced the activity slightly (21, 113 pM) and a
`piperidyl-4-carboxy spacer resulted in a 6-fold lower
`activity (22, 410 pM).
`As described above, several compounds had half-
`lives making them suitable for once-daily administration.
`Compound 5, y-.-glutamoyl(N-o-hexadecanoyl)-Lys?6,
`Arg**-GLP-1(7-37) was selected as the best compound
`for several reasons. A y-L-Glu spacer was preferred
`because it gave the most potent compound. Position 26
`or 34 were the most preferred amino acids to derivatize
`because they were natural
`lysines and could thus be
`derivatized easily and without introducing any potential
`immunogenic changes in the amino acid sequence.
`Position 26 gave the most potent compound. Further-
`more, acylation at position 26 gave the most metabolical-
`
`MPI EXHIBIT 1031 PAGE 5
`
`
`
`682
`
`GLP-1 derivatives for the treatment of type 2 diabetes
`
`the same inhibition of food intake was observed in rats
`with defective hypothalamic GLP-1 receptors due to
`monosodium glutamate injections. In this study, a reduc-
`tion in plasmatriglyceride levels was observed as well
`(52). Moreover, in normal rats, NN2211 has been shown
`to affect carbohydrate intake more than fat intake (87).
`Doses of 0.003 mg/kg NN2211 and highersignificant-
`ly reduced blood glucose in an acute study in diabetic
`ob/ob mice. At the same doses, food intake wasalso sig-
`nificantly reduced (88).
`In both ob/ob and db/db mice,
`subchronic studies of 14 days duration led to continuous-
`ly lowered blood glucose levels. However, in these animal
`models of diabetes, no effect on food intake was obser-
`ved after day 1. The blood glucose lowering effect was
`maintained throughout the study but was mostsignificant
`on day 1 (89). A comparative study to exendin-4, another
`GLP-1-based compound in clinical development, was
`also conducted. This study showed a more pronounced
`glucose lowering effect of NN2211 than exendin-4 (90). In
`ob/ob mice, a tendency to an increased beta-cell prolifer-
`ation was observed (data not published). In db/db mice,
`there were
`significant
`effects
`on
`both
`f-cell
`proliferation and mass (91). NN2211 effects were greater
`than exendin-4. These differences between NN2211 and
`exendin-4 may be explained by the longer half-life of
`NN2211.
`Twice-daily doses of 0.15 mg/kg s.c. NN2211 very
`effectively delayed the progression of diabetes in a
`6-week preventive study in young Zucker diabetic fatty
`rats. NN2211 was significantly more effective than pair
`feeding alone
`and decreased HbAtic
`by 3.1%.
`Cholesterol and free fatty acid plasma levels were also
`reduced in this model. No increased B-cell proliferation
`was observed compared to the control group, but
`increased B-cell volume was observed in both NN2211-
`treated and pair-fed animals. Food intake was decreased
`for the entire duration of the study (92).
`NN2211 and GLP-1 have been shown to inhibit
`cytokine-induced apoptosis in vitro using isolated B-cells
`(41). These results may explain why NN2211 in someani-
`mal models increases B-celi mass without significantly
`affecting B-cell proliferation.
`In summary,
`in a broad spectrum of animal models
`NN2211 has been shown to lower blood glucose and
`body weight, and to increase or maintain B-cell mass. All
`the effects are consistent with the known physiological
`effects of GLP-1. In humans, a study in type 2 diabetic
`patients has shownthat a single injection of 0.010 mg/kg
`NN2211 lowered blood glucose from 8.1 + 1.0 mM to
`6.9 + 1.0 mM (p = 0.004) (82). NN2211 is currently in
`phase2 clinical trials where it’s long-term efficacy will be
`evaluated.
`
`Conclusions
`
`GLP-1 compounds form a new classof drugsinclini-
`cal developmentfor the treatment of type 2 diabetes. This
`new class of drugs is especially interesting because
`
`GLP-1 has been shownto lower blood glucose as well as
`food intake and body weight. However, to be clinically
`useful the dynamic half-life needs to be significantly pro-
`longed. We found that the peptide hormone GLP-1 could
`be derivatized almost anywhere in the C-terminal part of
`the peptide and that derivatization with both short and
`long fatty acids and amino acid-derived spacers led to
`compounds that were highly potent. A number of com-
`pounds were both very potent and had plasmahalf-lives
`above 10 h, making them suitable as drugs for the treat-
`mentof type 2 diabetes using once-daily administration.
`NN2211 has been selected for clinical development and
`is currently in phase 2 clinicaltrials.
`NN2211 is a metabolically stable compound with
`potency equal to GLP-1. It has been characterized to act
`as a GLP-1 compoundin several animal models, includ-
`ing the ability to lower body weight. NN2211 is currently
`the only GLP-1 compoundin clinical development that
`has been shown to possess pharmacokinetic properties
`applicable to once-daily administration. The only study
`carried out thus far in type 2 diabetic patients has con-
`firmed its efficacy. Ongoing phase 2 clinical trials will
`reveal the potential of NN2211 as a promising new treat-
`mentfor type 2 diabetes.
`
`References
`
`1. ADA:Clinical practice recommendations 2001. Diabetes Care
`2001, 24(S1): $134.
`
`2. Bell, G.l., Santerre, R.F., Mullenbach, G.T. Hamster pre-
`proglucagon contains the sequence of glucagon and two related
`peptides. Nature 1983, 302: 716-8.
`
`3. Bell, G.I., Sanchez-Pescador, R., Laybourn, P.J., Najarian,
`R.C. Exon duplication and divergence in the human pre-
`proglucagon gene. Nature 1983, 304: 368-71.
`
`E.
`4. DeFronzo, R.A., Bonadonna, R.C., Ferrannini,
`Pathogenesis of NIDDM. A balanced overview. Diabetes Care
`1992, 15: 318-68.
`
`5. Unger, R.H., Orci, L. The essential role of glucagon in the
`pathogenesis of diabetes mellitus. Lancet 1975, 1: 14-6.
`
`6. Deacon, C.F., Johnsen, A.H., Holst, J.J. Degradation of
`glucagon-like peptide-1 by human plasmain vitro yields an N-ter-
`minally truncated peptide that is a major endogenous metabolite
`in vivo. J Clin Endocrinol Metab 1995, 80: 952-7.
`
`7. Kieffer, T.J., Mcintosh, C.H.S., Pederson, R.A. Degradation of
`GIP and truncated GLP-1 in vitro andin vivo by dipeptidy! pepti-
`dase IV. Endocrinology 1995, 136: 3585-96.
`
`8. Knudsen, L.B., Pridal, L. GLP-1(9-36)amide is major metabo-
`lite of GLP-1(7-36)amide after in vivo administration to dogs, and
`it acts as an antagonist on the pancreatic receptor. Eur J
`Pharmacol 1996, 318: 429-35.
`
`9. Gutniak, M.K., Linde, B., Holst, J.J. et al. Subcutaneousinjec-
`tion of the incretin hormone GLP-1 abolishes postprandial
`glycemia in NIDDM. Diabetes Care 1994, 17: 1039-44.
`
`10. Nauck, M.A, Wolischiager, D., Werner, J. et al. Effects of sub-
`cutaneous GLP-1(7-36)amide
`in patients with NIDDM.
`Diabetologia 1996, 39: 1546-53.
`
`MPI EXHIBIT 1031 PAGE 6
`
`MPI EXHIBIT 1031 PAGE 6
`
`
`
`Drugs Fut 2001, 26(7)
`
`683
`
`11. Nauck, M.A., Heimesaat, M.M., Orskov, C. et al. Preserved
`incretin activity of GLP-1(7-36amide) but not of synthetic human
`GIP in patients with type 2-diabetes mellitus. J Clin Invest 1993,
`91: 301-7.
`
`27. Zander, M., Madsbad, S., Holst, J.J. GLP-7 for six weeks
`reduces weight and improves insulin sensitivity and glycemic
`control in patients with type 2 diabetes. Diabetes 2001, 50(S2):
`A31.
`
`12. Nathan, D.M., Schreiber, E., Fogel, H. et al. insulinotropic
`action of GLP-1-(7-37) in diabetic and non-diabetic subjects.
`Diabetes Care 1992, 15: 270-6.
`
`13. Nauck, M.A., Kleine, N., Orskov C. et al. Normalization of
`fasting hyperglycemia by exogenous GLP-1(7-36amide) in type
`2 diabetic patients. Diabetologia 1993, 36: 741-4.
`
`14. Kreymann, B., Ghatai, M.A., Williams, G. et al. GLP-1(7-36):
`A physiological incretin in man. Lancet 1987, 2: 1300-4.
`
`15. Thorens, B. Expression clon