European Journal of Pharmacology 451 (2002) 217 – 225
`
`www.elsevier.com/locate/ejphar
`
`NN2211: a long-acting glucagon-like peptide-1 derivative with anti-
`diabetic effects in glucose-intolerant pigs
`
`Ulla Ribel a,*, Marianne O. Larsen a, Bidda Rolin a, Richard D. Carr a, Michael Wilken a,
`Jeppe Sturis a, Lisbet Westergaard a, Carolyn F. Deacon b, Lotte Bjerre Knudsen a
`
`aPharmacological Research 1, Health Care Pharmacology, Novo Nordisk A/S, Novo Alle´, DK-2880 Bagsværd, Denmark
`bDepartment of Medical Physiology, The Panum Institute University of Copenhagen, Denmark
`
`Received 26 June 2002; accepted 30 July 2002
`
`Abstract
`
`Glucagon-like peptide-1 (GLP-1) is an effective anti-diabetic agent, but its metabolic instability makes it therapeutically unsuitable. This
`q
`study investigated the pharmacodynamics of a long-acting GLP-1 derivative (NN2211: (Arg34Lys26-(N-
`-(g-Glu(N-a-hexadecanoyl)))-GLP-
` 1 i.v.)
`1(7 – 37)), after acute and chronic treatment in hyperglycaemic minipigs. During hyperglycaemic glucose clamps, NN2211 (2 Ag kg
` 1 min
` 1). Insulin excursions were
`treated pigs required more ( P < 0.005) glucose than control animals (5.8 F 2.1 vs. 2.9 F 1.8 mg kg
` 1 min), and glucagon levels were suppressed ( P < 0.05). Once-daily
`higher ( P < 0.01) after NN2211 (15367 F 5438 vs. 9014 F 2952 pmol l
` 1 s.c.) reduced the glucose excursion during an oral glucose tolerance test, to 59 F 15% of pre-treatment
`injections of NN2211 (3.3 Ag kg
`values by 4 weeks ( P < 0.05), without measurable changes in insulin responses. Fructosamine concentrations were unaltered by vehicle, but
` 1, P = 0.14) after 4 weeks of NN2211. Gastric emptying was reduced ( P < 0.05) by
`decreased (from 366 F 187 to 302 F 114 Amol l
`NN2211. NN2211 acutely increases glucose utilization during a hyperglycaemic glucose clamp and chronic treatment results in better daily
`metabolic control. Therefore, NN2211, a GLP-1 derivative that can be administered once daily, holds promise as a new anti-diabetic drug
`with a minimal risk of hypoglycaemia.
`D 2002 Elsevier Science B.V. All rights reserved.
`
`Keywords: GLP-1 glucagon-like peptide-1 derivative; Minipig; Diabetes type 2; Glucose tolerance; Insulin sensitivity; Gastric emptying; Glucagon
`
`1. Introduction
`
`Glucagon-like peptide 1 (GLP-1), a potent incretin hor-
`mone secreted from the intestinal L-cells after ingestion of
`carbohydrate and fat (Holst, 1997; Kreymann et al., 1987;
`Ørskov, 1978), has a variety of physiological effects which
`support its potential use as an anti-hyperglycaemic agent.
`Firstly, it stimulates insulin and decreases glucagon secre-
`tion in a glucose-dependent manner (Gromada et al., 1998).
`Secondly, the hormone potently inhibits gastric emptying
`(Nauck et al., 1997) and suppresses appetite (Flint et al.,
`1998), and, lately, it has been shown that GLP-1 stimulates
`beta-cell growth (Buteau et al., 1999; Edvell and Lindstro¨m,
`1999; Gang et al., 1999) and inhibits apoptosis (Hansotia et
`al., 2001). This unique combination of properties provides
`
`* Corresponding author. Tel.: +45-44422014; fax: +45-44427488.
`E-mail address: ulr@novonordisk.com (U. Ribel).
`
`an unprecedented opportunity to develop an effective and
`safe anti-diabetic compound, particularly since both the
`insulinotropic and glucagonostatic effects are glucose-
`dependent (Qualmann et al., 1995). Consequently, since
`its discovery in 1984, GLP-1 has received much attention
`as a possible new treatment for type 2 diabetes (Gutniak et
`al., 1994; Larsen et al., 2001; Nauck et al., 1993, 1996;
`Rachman et al., 1997). However, the native sequence of
`GLP-1 is rapidly degraded and deactivated by dipeptidyl
`peptidase IV (DPPIV; EC 3.4.15.5) (Mentlein et al., 1993;
`Deacon et al., 1995) and eliminated through the kidneys
`(Deacon et al., 1996). These degradation pathways preclude
`the use of the native form of the GLP-1 molecule therapeuti-
`cally. Therefore, we have designed derivatives of GLP-1,
`which have a protracted pharmacodynamic profile due to
`binding to serum albumin, resistance towards DPPIV deg-
`radation and slow release from the injection site. NN2211
`q
`[(Arg34Lys26-(N-
`-(g-Glu(N-a-hexadecanoyl)))-GLP-1(7 –
`37)] is derivatised with a fatty acid side chain and a spacer
`
`0014-2999/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.
`PII: S 0 0 1 4 - 2 9 9 9 ( 0 2 ) 0 2 1 8 9 - 1
`
`MPI EXHIBIT 1087 PAGE 1
`
`MPI EXHIBIT 1087 PAGE 1
`
`

`

`218
`
`U. Ribel et al. / European Journal of Pharmacology 451 (2002) 217–225
`
`adding charge and solubility (Knudsen et al., 2000). The
`compound has a maintained affinity for the GLP-1 receptor,
`affinity for serum albumin and resistance to degradation by
`DPPIV (Knudsen et al., 2000), giving it a unique pharma-
`cokinetic and pharmacodynamic profile suitable for once-
`daily subcutaneous administration to humans.
`The present study was undertaken to investigate the
`insulinotropic and glucagonostatic effects of NN2211, both
`acutely (hyperglycaemic clamp) and chronically (4-week
`dosing study), in nicotinamide and streptozotocin (STZ)
`treated beta-cell-reduced minipigs (Larsen et al., 2000). This
`is a new model which was developed to mimic more closely
`the human conditions of impaired glucose tolerance (IGT)
`and mild type 2 diabetes. In addition, minipigs provide the
`additional advantage of resembling humans in terms of
`gastrointestinal (Miller and Ullrey, 1987; Tumbleson,
`1986) and skin physiology (Quist et al., 2000), which are
`important considerations in terms of nutrition and drug
`absorption from the skin.
`
`2. Materials and methods
`
`2.1. Animals
`
`All experiments were carried out in accordance with
`animal welfare guidelines provided by the Animal Experi-
`ments Inspectorate, Ministry of Justice Denmark.
`Male Go¨ ttingen Minipigs, purchased from Ellegaard
`Go¨ttingen Minipigs, Denmark were used in the study. The
`pigs were housed in single pens, and fed a diet as recom-
`mended by the supplier: 245 g twice daily of Special Diet
`Sciences (SDS, Witham, Essex, UK) pelleted fodder. After 2
`weeks of acclimatization, the pigs were anaesthetised with a
` 1, tiletamin 0.8 mg
`combination of zolazepam 0.8 mg kg
` 1, ketamin 0.8 mg kg
` 1 and
` 1, methadone 0.2 mg kg
`kg
` 1 i.m., and maintained on anaesthesia
`xylazin 0.9 mg kg
`with isoflurane 1 – 3%, after tracheal
`intubation. During
`anaesthesia, the animals were instrumented with two venous
`catheters (Certo 455, B. Braun, Melsungen, Germany)
`inserted in the external jugular vein and advanced to the
`superior vena cava. The catheters were exteriorised to the
`back of the neck and filled with saline containing 1000 U
` 1 heparin. Animals were allowed to recover from the
`l
`anaesthesia, and were given post-operative analgesia (bupre-
` 1
` 1 i.m. and carprophene 4 mg kg
`norphine 0.03 mg kg
`i.m. once daily) for 3 days. Catheters were flushed twice a
`week with saline.
`One to two weeks after surgery, the pigs underwent an
`oral glucoce tolerance test (OGTT), in which glucose (2 g
` 1) was mixed with 25 g pelleted fodder. This was
`kg
`offered to unrestrained pigs in a bowl (t = 0 min), and the
`animals were carefully supervised while eating the mixture.
`Blood samples for measurement of plasma glucose, insulin,
`and glucagon were taken at the following time points: 15,
` 10, 0, 15, 30, 45, 60, 90, 120, 150 and 180 min. All blood
`
`samples were drawn from the indwelling catheters while the
`pigs were moving freely in their pens.
`
`2.2. Induction of diabetes
`
`Animals had a body weight of 24 – 29 kg, at the time,
`they had their beta-cell mass reduced with two separate i.v.
` 1 (protocol 1) nicotinamide
`injections of either 100 mg kg
`(which prevents cellular energy depletion caused by STZ-
` 1 (protocol 2) and
`induced DNA damage) or 45 mg kg
` 1 i.v. (Larsen et al., 2000). After
`streptozotocin 125 mg kg
`8 days, a second OGTT was carried out as described above,
`in order to assess the severity of diabetes, according to the
`human criteria defined by the American Diabetes Associa-
`tion: impaired glucose tolerance (IGT), 2-h plasma glucose
` 1, type 2 diabetic, fasting
`during OGTT > 7.8 mmol l
` 1 and/or 2-h plasma glucose
`plasma glucose >7.0 mmol l
` 1.
`during OGTT >11.1 mmol l
`
`2.3. Experimental designs
`
`2.3.1. Protocol 1: hyperglycaemic clamp
`One month after nicotinamide and STZ treatment, a
`hyperglycaemic clamp was carried out in six animals. Prior
`to the experimental day, the animals were fasted for 18 h
`and had free access to water. The study was conducted in
`fasted animals, since it is known that GLP-1 infusion carried
`out during ingestion of a meal results in a diminished insulin
`response due to delayed intestinal absorption (Nauck et al.,
`1997).
`On 2 days, separated by 1 week, the animals were dosed
`i.v. via the implanted catheter, with either vehicle or with the
` 1, corresponding to
`GLP-1 derivative NN2211 (2 Ag kg
` 1), with each animal receiving both treatments
`0.6 nmol kg
`in random order. The injected volume was 1 ml and the
`catheter was flushed with saline. Two basal blood samples
`were collected at 30 and 1 min, after which NN2211 or
`vehicle were administered (t = 0 min). An i.v. glucose bolus
` 1) was given at t = 30 min, followed by an
`load (0.1 g kg
`i.v. infusion of a 20% glucose solution at a variable rate,
`which aimed to clamp the plasma glucose at a level 1.5 – 2
` 1 above the basal level of the individual animals.
`mmol l
`Blood samples were thereafter drawn at 5-min intervals until
`t = 60 min followed by 10-min intervals until t = 110 min,
`after which the glucose infusion was terminated and further
`blood samples were collected until t = 130 min. Plasma
`glucose was measured on the Yellow Springs Instruments
`glucose analyser (YSI, Ohio, USA). Blood (3 ml) for
`determination of plasma glucose,
`insulin, glucagon and
`NN2211 concentrations was collected into tubes containing
`the relevant additives for the specific assays, as described
`below. Samples were centrifuged (4 jC, 3000  g, 10 min)
`and plasma separated and stored at 20 jC until analysed.
`In order to study the glucose-dependence of NN2211
`further, data from another set of experiments carried out in a
`parallel group of pigs were analysed. These experiments
`
`MPI EXHIBIT 1087 PAGE 2
`
`MPI EXHIBIT 1087 PAGE 2
`
`

`

`U. Ribel et al. / European Journal of Pharmacology 451 (2002) 217–225
`
`219
`
`were identical in design except for the fact that they utilized
` 1
`a constant low dose i.v. glucoce infusion (3.3 mg kg
` 1), without clamping the blood glucose.
`min
`
`2.3.2. Protocol 2: chronic dosing study
`The effects of chronic (4 weeks) treatment with NN2211
`were examined in a second group of animals. OGTTs (as
`described above, but with the addition of 500 mg para-
`cetamol) were carried out prior to ( = pre-STZ) and 1 week
`after ( = post-STZ) induction of diabetes. A further two
`OGTTs were performed after 2- and 4-week treatment with
`once-daily s.c. injections (0.33 ml) of NN2211 (3.3 Ag
` 1 dissolved in phosphate-buffered saline; n = 6) or
`kg
`vehicle alone (n = 6). On OGTT days, animals were fasted
`overnight, but allowed free access to water. Animals
`received their usual
`injection of vehicle or NN2211 at
`07.30 h (t = 270 min), and the OGTT was performed at
`12:00 h (t = 0 min). Blood samples (4 ml) were taken at
` 15, 0, 15, 30, 45, 60, 90, 120, and 180 min for determi-
`nation of plasma glucose, insulin, glucagon, and paraceta-
`mol. Blood samples for determination of fructosamine and
`NN2211 concentrations were collected once a week at
`12:00 h.
`
`2.4. Formulation of compound
`
`q
`
`NN2211 is an acylated GLP-1(7 – 37): (Arg34Lys26-(N-
`-
`(g-Glu(N-a-hexadecanoyl)))-GLP-1(7 – 37) (Knudsen et al.,
` 1) was dissolved in 4 mmol
`2000). NN2211 (4.91 mg ml
` 1 phosphate buffer, containing 38 mg ml
` 1 mannitol,
`l
` 1 phenol in water. The preparation was
`and 5.5 mg ml
`stored at 4 jC, and diluted for the clamp experiment (1 + 99)
`in saline immediately before dosing. For the chronic dosing
` 1) was prepared in
`study, a solution of NN2211 (284 Ag ml
` 1 phosphate buffer.
`isotonic 0.2 mol l
`
`2.5. Analytical procedures
`
`The glucose concentration was measured by the glucose
`hexokinase method using a Cobas Mira auto-analyser Plus
`(Roche Diagnostic Systems, Basel, Switzerland) following
`the procedures as described by the manufacturer. Fructos-
`amine was measured by a reduction test with nitrobluete-
`trazolium (ABX Diagnostics, Montpellier, France), and
`paracetamol concentrations were determined by high pres-
`sure liquid chromatrography (Hewlett Packard HP1090, C8
`column) with UV detection at 245 nm following extraction
`from plasma with ethyl acetate.
`Plasma insulin was analysed using an in-house two site
`enzyme-linked immunosorbent assay (ELISA) based on
`two monoclonal antibodies as catcher and detector, respec-
` 1 and
`tively. The assay has a detection limit of 3.2 pmol l
` 1 of
`the inter-assay variations at 87, 235 and 342 pmol l
`15.3%, 9.9% and 14.6%, respectively. Intra-assay variations
`at the same concentration levels are 3.2%, 7.6% and 4.4%,
`respectively. Cross-reactivities were as follows (all por-
`
`cine): growth hormone 0.001%, glucagon 0.4%, pancreatic
`polypeptide 0.2% and C-peptide 0.01%. Plasma NN2211
`was analysed using an in-house two-site immunoassay with
`monoclonal antibodies. The catcher antibody was raised
`against intact GLP-1(7 – 37) coupled to KLH and reacted
`with the C-terminal half of the molecule while the detector
`antibody raised against an N-terminal fragment of GLP-1
`coupled to KLH reacted specifically with the N-terminus
`and not with elongated or truncated forms of GLP-1. The
`detector antibody was biotinylated and the detection system
`was streptavidin-peroxidase in combination with Super-
`Signal amplifying system (Pierce, cat.no. 37075). (Wilken
`et al., 2000). The assay measures the sum of free and
`albumin-bound NN2211, and has a detection limit of 3
` 1. The intra-assay variations at 85, 790 and 3220
`pmol l
` 1 are 5.9%, 6.5%, and 2.4%, respectively and
`pmol
`l
` 1 are
`inter-assay variations at 95, 840 and 3395 pmol l
`10.1%, 6.2% and 3.7%, respectively. Cross-reactivity to
`endogenous GLP-1 is < 4%. Plasma for insulin and
` 1; 35 Al
`NN2211 was stabilised with EDTA (0.18 mol l
` 1 blood). Plasma glucagon was analysed using a
`ml
`commercial
`radioimmunoassay kit
`(GL-32K, Linco
`Research St Charles, MO, USA). Plasma was stabilised
` 1 blood)
`with Trasylol (500 kallikrein inhibitory units ml
` 1; 35 Al ml
` 1 blood).
`and EDTA (0.18 mol l
`
`2.6. Data analysis
`
`Data were analysed using Statistical Analysis System
`software (6.11, SAS Institute, Cary, NC, USA). Data are
`presented as mean F S.D.; time courses as mean F S.E.M.
`Pair-wise group comparisons were performed using a paired
`t-test. In the chronic study, one-way analysis of variance
`followed by post-hoc analysis was used to demonstrate
`differences over the course of the study (pre-, post-STZ,
`2- and 4-week treatments). Values of P < 0.05 were consid-
`ered significant. To investigate the glucose dependency of
`NN2211’s ability to stimulate insulin secretion, correspond-
`ing approximate steady-state values of glucose and insulin
`(mean of time points 15 and 28 min after i.v. bolus of
`NN2211 and mean of time points 80, 90 and 100 min after
`i.v. bolus of NN2211) were calculated for each set of
`experiments. These data points were then plotted (insulin
`vs. glucose) for each condition (NN2211 or vehicle) and
`best-fit lines were estimated by linear regression analysis.
`The slopes of the best-fit lines were statistically compared as
`described by Zar (1984).
`
`3. Results
`
`3.1. Protocol 1
`
`After 8 days, five of the animals were characterised as
`IGT and one as type 2 diabetic, according to the human
`criteria defined by the American Diabetes Association.
`
`MPI EXHIBIT 1087 PAGE 3
`
`MPI EXHIBIT 1087 PAGE 3
`
`

`

`220
`
`U. Ribel et al. / European Journal of Pharmacology 451 (2002) 217–225
`
` 1) (Fig. 1),
`the vehicle-treated group (233 F 132 mg kg
`representing a 196 F 257% increase during NN2211 treat-
`ment ( P < 0.005). Despite this, plasma glucose concentra-
`tions during the clamp (70 – 110 min) tended to be slightly
` 1)
`lower during NN2211 treatment (6.1 F 0.5 mmol
`l
` 1; n.s.; Fig.
`than in the vehicle group (6.4 F 0.3 mmol l
`2A). In spite of this trend,
`the corresponding plasma
`insulin concentration profile (Fig. 2B) was markedly
`higher in the NN2211-treated animals (area under the
`plasma insulin curve (AUC30 – 110 min), 15367 F 5438 pmol
` 1
` 1 min for NN2211 compared to 9014 F 2952 pmol l
`l
`min) representing a 72 F 28% increase ( P < 0.01). After
`NN2211 injection, plasma glucagon was suppressed during
`the hyperglycaemic clamp (glucagon AUC70 – 110 min
` 1
`decreased by 31 F 14%, from 832 F 360 pmol l min
` 1 min P < 0.05), but
`(vehicle group) to 531 F 82 pmol l
`the suppression was lifted immediately after the glucose
`infusion was terminated and plasma glucose fell below
`fasting levels (Fig. 2C). Plasma NN2211 immunoreactivity
`remained relatively constant during the test period,
` 1 at
`6825 F 1155, 6603 F 1198 and 6318 F 1103 pmol l
`45, 60, and 130 min, respectively, reflecting a prolonged
`half-life of the compound also after i.v. administration
`(Fig. 2D). For both NN2211 (r2 = 0.97, P < 0.05) and
`
`Fig. 1. Hyperglycaemic glucose clamp in h-cell-reduced minipigs. NN2211
`(2 Ag kg
` 1 i.v.) treated pigs required a 196 F 257% increase in the glucose
`infusion rate compared to vehicle ( P < 0.005, mean F S.E.M., n = 6).
`
`NN2211-treated animals required more glucose during
`the hyperglycaemic clamp period (cumulated amount
` 1) compared to
`infused 30 – 110 min; 462 F 152 mg kg
`
`Fig. 2. (A) Plasma glucose profile before, during and after hyperglycaemic glucose clamp. The plasma glucose clamp level tends to be slightly lower during
` 1 i.v.) treatment (6.1 F 0.5 vs. 6.4 F 0.3 mmol l
`NN2211 (2 Ag kg
` 1, n.s., mean F S.E.M., n = 6). (B) Plasma insulin profiles show the area under the curve to
`be 72 F 28% higher after NN2211 (2 Ag kg
` 1 i.v.) compared to vehicle ( P < 0.01, mean F S.E.M., n = 6). (C) Plasma glucagon is suppressed during
`hyperglycaemic clamp by 31 F 14% after NN2211 (2 Ag kg
` 1 i.v.) compared to vehicle ( P < 0.05, mean F S.E.M., n = 6). (D) Plasma concentrations of
`NN2211 after i.v. injection of 2 Ag kg
` 1 remained constant during the clamp period (mean F S.E.M., n = 6).
`
`MPI EXHIBIT 1087 PAGE 4
`
`MPI EXHIBIT 1087 PAGE 4
`
`

`

`U. Ribel et al. / European Journal of Pharmacology 451 (2002) 217–225
`
`221
`
`in animals
`ened by nicotinamide and STZ treatment
`assigned to receive vehicle treatment, and remained constant
`over the 4-week vehicle treatment (Table 1). In animals
`allocated to receive NN2211, the deterioration of glucose
`tolerance was more pronounced, reflecting the fact that this
`group included animals with severe diabetes. However, in
`contrast to vehicle, treatment with NN2211 significantly
`improved glucose tolerance at both 2 and 4 weeks (Table 1),
`corresponding to reductions in the glucose excursion to
`74 F 12% and 59 F 15% of the post-STZ values at 2 and
`4 weeks, respectively. When the two subgroups of NN2211-
`treated animals were examined, the respective values were
`73 F 14% and 68 F 3% for the IGT group and 65% and
`51% for the diabetic group.
`Induction of diabetes resulted in an impairment of the
`insulin response to the OGTT in both vehicle and NN2211
`groups. The insulin response remained impaired over the 4-
`week study period in both groups (Table 2).
`The glucose to insulin ratios (calculated from the
`AUC0 – 120 min) did not differ between the groups prior to
`the induction of diabetes (Table 3). In the vehicle group,
`the ratio increased after STZ treatment and thereafter was
`unchanged over the 4-week period (Table 3). ANOVA
`analysis revealed an overall difference ( P < 0.05) between
`vehicle treatment and the NN2211 (IGT) subgroup,
`although post hoc analysis to compare the individual time
`points failed to reach statistical significance. However,
`NN2211 treatment did result in a reduction of the glucose
`to insulin ratio, corresponding to falls to 53% ( P < 0.001)
`and 60% ( P < 0.05) in the IGTs and 20% and 17% in the
`diabetic animals at 2 and 4 weeks, compared to their
`respective post-STZ values. Taking the NN2211 as a
`whole,
`the corresponding values were 25% at 2 weeks
`(n.s.) and 20% at 4 weeks ( P < 0.05) compared to NN2211
`(all) post-STZ values.
`
`Table 1
` 1 s.c. once daily) treatment on glucose
`Effect of NN2211 (3.3 Ag kg
` 1) before and after induction of
`tolerance during an OGTT (2 g kg
`diabetes in Go¨ttingen minipigs
`
`Glucose AUC0 – 120 min (mmol l
`
` 1 min)
`Pre-STZ
`Post-STZ
`2 Weeks
`4 Weeks
`974 F 75b
`879 F 111b
`689 F 21
`982 F 69a
`Vehicle
`787 F 53c
`1950 F 46d
`1437 F 797e
`1149 F 287e
`NN2211(All)
`777 F 82c
`1237 F 133f
`914 F 150e
`837 F 107g
`NN2211(IGT)
`NN2211(Dia)
`806
`3375
`2222
`1774
`Data are mean F S.D., n = 6 in Vehicle and NN2211(All) groups, n = 4 in
`NN2211(IGT) group and n = 2 in NN2211(Diabetic) group. No statistics is
`applied with diabetic group because of the low number of animals.
`a P < 0.01 vs. vehicle pre-STZ.
`b n.s. vs. vehicle post-STZ.
`c n.s. vs. vehicle pre-STZ.
`d P < 0.05 vs. NN2211 pre-STZ.
`e P < 0.05 vs. NN2211 post-STZ.
`f P < 0.001 vs. NN2211 pre-STZ.
`g P < 0.001 vs. NN2211 post-STZ.
`
`Fig. 3. Correlation between mean plasma glucose and plasma insulin at
`steady state. The slope of the best-fit line is significantly steeper after
`NN2211 treatment (2 Ag kg
` 1, i.v.) after vehicle ( P < 0.05) demonstrating
`that NN2211 stimulation of insulin secretion is glucose dependent. Two
`points from each condition (NN2211 or vehicle) are from a low-dose
`glucose infusion study in a separate group of pigs.
`
`vehicle (r 2 = 0.96, P < 0.05) conditions, there was a sig-
`nificant
`linear correlation between plasma insulin and
`glucose values. There was, however, a significant differ-
`ence in slope of the best-fit lines between the NN2211 and
`vehicle treated animals ( P < 0.05), demonstrating that
`NN2211 stimulation of insulin secretion is glucose depend-
`ent (Fig. 3).
`
`3.2. Protocol 2
`
`After treatment with nicotinamide and STZ, 2 animals
`were characterised as diabetic and 10 as IGT. Because of the
`severity of diabetes in the two diabetic animals (fasting
` 1), it was considered
`plasma glucose above 11 mmol l
`unethical to randomly allocate the animals to receive vehicle
`or NN2211 treatment. Therefore,
`these animals were
`assigned to receive NN2211, with random allocation of
`the remainder. The NN2211-treated group was statistically
`treated as one group (n = 6, four IGT pigs and two diabetic
`pigs). In addition, for the glucose and insulin responses, the
`NN2211-treated animals were further subdivided into those
`with IGT and those with diabetes, although because of the
`small number of diabetic animals, further statistical analysis
`was not possible.
`The dose of NN2211 used for s.c. administration was
`selected because it resulted in plasma levels similar to those
` 1, which was
`obtained after i.v. administration of 2 Ag kg
`shown to be effective in the acute study (protocol 1). This
` 1) was well tolerated, with all signs of
`dose (3.3 Ag kg
`well-being and behaviour being normal. Animals ate and
`drank normally, and there were no incidences of vomiting.
`Despite the fact
`that
`the animals were not randomly
`allocated to receive NN2211 or vehicle,
`there was no
`significant difference in their response to the pre-STZ
`OGTT (Table 1). Glucose tolerance was significantly wors-
`
`MPI EXHIBIT 1087 PAGE 5
`
`MPI EXHIBIT 1087 PAGE 5
`
`

`

`222
`
`U. Ribel et al. / European Journal of Pharmacology 451 (2002) 217–225
`
`Table 2
`Effect of NN2211 (3.3 Ag kg
`minipigs
`
` 1 s.c. once daily) treatment on insulin secretion during an OGTT (2 g kg
`
` 1) before and after induction of diabetes in Go¨ttingen
`
`Insulin AUC0 – 120 min (pmol l
`
` 1 min)
`Pre-STZ
`Post-STZ
`2 Weeks
`4 Weeks
`27181 F 24304b
`25014 F 20711b
`33078 F 12763
`21958 F 16327a
`Vehicle
`41219 F 5797
`15867 F 10835c
`22874 F 12867d
`19018 F 9545d
`NN2211(All)
`38873 F 5464e
`22212 F 5803f
`23336 F 16530d
`24372 F 6099d
`NN2211(IGT)
`NN2211(Dia)
`45912
`3177
`11132
`8311
`Data are mean F S.D., n = 6 in Vehicle and NN2211(All) groups, n = 4 in NN2211(IGT) group and n = 2 in NN2211(Diabetic) group. No statistics is applied
`with diabetic group because of the low number of animals.
`a P < 0.05 vs. vehicle pre-STZ.
`b n.s. vs. vehicle post-STZ.
`c P < 0.01 vs. NN2211 pre-STZ.
`d n.s. vs. NN2211 post-STZ.
`e n.s. vs. vehicle pre-STZ.
`f P < 0.001 vs. NN2211 pre-STZ.
`
`In the group of IGT’s allocated to receive NN2211, the
`ratio was similar to the vehicle group after STZ treatment
`whereas the diabetic animals showed a significantly higher
`ratio after STZ. However, NN2211 treatment (all) resulted
`in a reduction of the glucose to insulin ratio, corresponding
`to falls to 25% (n.s.) and 20% ( P < 0.05), respectively,
`compared to NN2211 post-STZ values. For the two
`NN2211 subgroups, the corresponding values were 53%
`and 60% in the IGTs and 20% and 17% in the diabetic
`animals, respectively.
`The glucagon response to the OGTT was not signifi-
`cantly affected by the induction of diabetes in either vehicle
` 1 min pre-STZ and
`(AUC0 – 120 min 2361 F 92 pmol
`l
` 1 min post-STZ) or NN2211
`2601 F 222 pmol
`l
` 1 min pre-STZ and 2786 F 267 pmol
`(2550 F 309 pmol l
` 1 post-STZ) groups. These responses did not
` 1 min
`l
`change significantly over the 4-week study period (vehicle,
` 1 min at 2 and 4
`2668 F 203 and 2768 F 140 pmol l
` 1
`weeks; NN2211, 2949 F 233 and 3349 F 353 pmol l
`min at 2 and 4 weeks).
`
`Gastric emptying (as determined by the appearance of
`paracetamol in the plasma) was not affected by induction of
`diabetes, and remained unchanged in the vehicle-treated
`group (Fig. 4A). In contrast, NN2211 treatment significantly
`reduced gastric emptying (Fig. 4B), corresponding to a
`reduced area under the paracetamol curve (AUC0 – 120 min,
`to 80 F 14% at 2 weeks, P < 0.05 and 51 F 24% at 4 weeks;
`P < 0.01).
`Fructosamine concentrations increased after induction of
` 1 pre-STZ to reach 250 F 7
`diabetes, from 221 F 10 Amol l
` 1 in the vehicle group and from 216 F 9 to 366 F 187
`Amol l
` 1 in the NN2211 group ( P < 0.05). In the vehicle
`Amol l
`group, fructosamine levels remained elevated throughout the
`4-week study period. In the NN2211 group, the rise in
`fructosamine concentrations was more marked, reflecting
`the greater severity of diabetes in these animals. However,
`during treatment, there was a tendency for fructosamine levels
` 1 at 4 weeks ( P = 0.14).
`to fall, reaching 302 F 114 Amol l
`Plasma NN2211 levels determined once a week at 12:00 h
`remained constant throughout the study period (6417 F 1711,
`
`Table 3
` 1 s.c. once daily) treatment on glucose to insulin ratios (calculated from the respective AUC0 – 120 min) during an OGTT
`Effect of NN2211 (3.3 Ag kg
` 1) before and after induction of diabetes in Go¨ttingen minipigs
`(2 g kg
`Glucose to insulin ratio
`
`4 Weeks
`2 Weeks
`Post-STZ
`Pre-STZ
`49  106 F 20  106b
`50  06 F 18  106b
`58  106 F 22  106a
`23  106 F 9  106
`Vehicle
`94  106 F 110  106f
`119  106 F 154  106e
`468  106 F 736  106d
`19  106 F 4  106c
`NN2211(All)
`36  106 F 6  106e
`32  106 F 11  106f
`60  106 F 19  106d
`20  106 F 3  106c
`NN2211(IGT)
`214  106
`250  106
`1275 106
`18  106
`NN2211(Dia)
`Data are mean F S.D., n = 6 in Vehicle and NN2211(All) groups, n = 4 in NN2211(IGT) group and n = 2 in NN2211(Diabetic) group. No statistics is applied
`with diabetic group because of the low number of animals.
`a P < 0.01 vs. vehicle pre-STZ.
`b n.s. vs. vehicle post-STZ.
`c n.s. vs. vehicle pre-STZ.
`d P < 0.01 vs. NN2211 pre-STZ.
`e n.s. vs. NN2211 post-STZ.
`f P < 0.05 vs. NN2211 post-STZ.
`
`MPI EXHIBIT 1087 PAGE 6
`
`MPI EXHIBIT 1087 PAGE 6
`
`

`

`U. Ribel et al. / European Journal of Pharmacology 451 (2002) 217–225
`
`223
`
`glucagon rose to contribute to the maintenance of normo-
`glycaemia. This suggests that the glucose-lowering effect of
`NN2211 is self-limiting and the drug is,
`therefore, not
`expected to cause serious hypoglycaemia regardless of dose,
`fully in keeping with the effects of native GLP-1 (Qualmann
`et al., 1995). The glucose-dependency of GLP-1’s action has
`been demonstrated in many studies in humans (Gutniak et
`al., 1994; Hvidberg et al., 1994; Nauck et al., 1993, 1996;
`Rachman et al., 1997; Willms et al., 1996), although in
`normal fasted subjects, it can (under the rather unphysio-
`logical conditions of an i.v. glucose load) temporarily lower
`blood glucose below normoglycaemia (Hvidberg et al.,
`1994; Toft-Nielsen et al., 1998). This is a natural conse-
`quence of its potent insulinotropic effect (particularly given
`the rapid increase in blood glucose following i.v. admin-
`istration), and since the inactivation time for insulin is
`considerable, enough may remain to lower blood glucose
`transiently below normoglycaemia. However, importantly,
`at lower blood glucose levels, GLP-1 no longer inhibits
`glucagon secretion, so that any tendency towards hypogly-
`caemia would immediately result in glucagon secretion, thus
`quickly restoring normoglycaemia. Any hypoglycaemia
`observed with GLP-1 in type 2 diabetic patients would,
`therefore, be expected to be mild and readily reversible by
`the body’s own counterregulatory mechanisms, especially
`given the relative insulin-resistance of such patients. Indeed,
`it has recently been demonstrated that a similar reactive
`hypoglycaemia cannot be induced in type 2 diabetic patients
`(Vilsboll et al., 2001).
`In order to maintain its full effect, it appears that GLP-1
`must be present throughout the 24-h dosing cycle. Thus, when
`GLP-1 is infused continuously in type 2 diabetic patients,
`both fasting and post-prandial glucose concentrations are
`almost normalised (Larsen et al., 2001; Rachman et al.,
`1997). However, once the infusion is halted, there is no
`sustained improvement of metabolic control (Rachman et
`al., 1997; Willms et al., 1998). Similarly, in a 1-week study
`where the GLP-1 infusion is given only during the day,
`metabolic control is poorer than when the infusion is main-
`tained for 24 h a day (Larsen et al., 2001). In the present study,
`a single injection of NN2211 significantly reduced the glu-
`cose excursion, even when given 270 min before the OGTT,
`confirming that the pharmacokinetic profile of NN2211 is
`suitable for maintaining sufficiently elevated plasma concen-
`trations to allow once-daily administration. In order to pro-
`long the plasma survival time of NN2211, GLP-1 was
`acylated to promote plasma albumin binding, resulting in a
`compound with a half-life of 14 h in pigs (Knudsen et al.,
`2000) and 10 – 12 h in man (Juhl et al., 2001). NN2211 is
`presumably inaccessible to the GLP-1 receptor when bound,
`so that albumin binding acts as a reservoir from which the
`active drug can dissociate, resulting in a sustained supply of
`agonist tone at the GLP-1 receptor over a full 24-h period.
`Moreover, this also explains why the high plasma levels of
`NN2211 were not associated with any adverse side-effects,
`since the assay measures both free and albumin-bound drug.
`
`Fig. 4. Plasma paracetamol concentrations (used as an index of gastric
`emptying) after ingestion of 500 mg/animal at pre-, post-STZ, 2 and 4
`weeks in vehicle treated (A) and NN2211 (3.3 Ag kg
` 1 s.c. once daily) (B)
`treated animals. There were no changes in the rate of gastric emptying in the
`vehicle treated animals, but NN2211 significantly reduced the rate of gastric
`emptying at 2 (to 80 F 14%, P < 0.05) and 4 weeks (to 51 F 24%, P < 0.01).
`Data are mean F S.E.M., n = 6.
`
`6973 F 1697, 6818 F 2063, and 5716 F 1635 pmol l
`2, 3 and 4 weeks, respectively).
`
` 1 at 1,
`
`4. Discussion
`
`In this study, the long-acting GLP-1 derivative NN2211
`increased glucose utilization during an acute hyperglycae-
`mic glucose clamp in glucose-intolerant minipigs. Further-
`more, the beneficial effect of the drug was maintained, with
`animals having markedly improved glucose tolerance dur-
`ing chronic dosing, supporting the therapeutic potential of
`this compound in the treatment of type 2 diabetes. During
`the clamp, NN2211 both induced insulin secretion and
`suppressed plasma glucagon, in agreement with its mecha-
`nism of action as a GLP-1 receptor agonist (Knudsen et al.,
`2000). These effects were glucose-dependent, as evidenced
`by the significantly steeper dose – response curve relating
`glucose and insulin, and the fact that at low glucose levels,
`
`MPI EXHIBIT 1087 PAGE 7
`
`MPI EXHIBIT 1087 PAGE 7
`
`

`

`224
`
`U. Ribel et al. / European Journal of Pharmacology 451 (2002) 217–225
`
`Other attempts to prolong the action of GLP-1 do not take
`advantage of interaction with albumin, relying instead solely
`upon introducing resistance to peptidases. Exendin-4 is one
`such analogue (plasma half-life of 26 min in man (Edwards et
`al., 2001)) with anti-hyperglycaemic activity in diabetic mice
`(Greig et al., 1999) and healthy humans (Edwards et al.,
`2001). Other s