Potent derivatives of glucagon-like peptide-1 with pharmacokinetic properties suitable for once daily administration

Article abstract of DOI:10.1021/jm9909645

A series of very potent derivatives of the 30-amino acid peptide hormone glucagon-like peptide-1 (GLP-1) is described. The compounds were all derivatized with fatty acids in order to protract their action by facilitating binding to serum albumin. GLP-1 had a potency (EC₅₀) of 55 pM for the cloned human GLP-1 receptor. Many of the compounds had similar or even higher potencies, despite quite large substituents. All compounds derivatized with fatty acids equal to or longer than 12 carbon atoms were very protracted compared to GLP-1 and thus seem suitable for once daily administration to type 2 diabetic patients. A structure-activity relationship was obtained. GLP-1 could be derivatized with linear fatty acids up to the length of 16 carbon atoms, sometimes longer, almost anywhere in the C- terminal part without considerable loss of potency. Derivatization with two fatty acid substituents led to a considerable loss of potency. A structure- activity relationship on derivatization of specific amino acids generally was obtained. It was found that the longer the fatty acid, the more potency was lost. Simultaneous modification of the N-terminus (in order to obtain better metabolic stability) interfered with fatty acid derivatization and led to loss of potency.

Full text of DOI:10.1021/jm9909645

1664
J . Med. Chem. 2000, 43, 1664-1669
Expedited Articles
P oten t Der iva tives of Glu ca gon -lik e P ep tid e-1 w ith P h a r m a cok in etic P r op er ties
Su ita ble for On ce Da ily Ad m in istr a tion
Lotte B. Knudsen,*,† Per F. Nielsen,‡ Per O. Huusfeldt,§ Nils L. J ohansen,§ Kjeld Madsen,§ Freddy Z. Pedersen,§
Henning Thøgersen,§ Michael Wilken,| and Henrik Agersø
Departments of Molecular Pharmacology, Protein Chemistry, Medicinal Chemistry, Assay & Cell Technology, and
Pharmacokinetics, Health Care Discovery and Preclinical Development, Novo Nordisk A/ S, Novo Park,
DK-2760 Maaloev, Denmark
Received December 3, 1999
A series of very potent derivatives of the 30-amino acid peptide hormone glucagon-like peptide-1
(GLP-1) is described. The compounds were all derivatized with fatty acids in order to protract
their action by facilitating binding to serum albumin. GLP-1 had a potency (EC50) of 55 pM for
the cloned human GLP-1 receptor. Many of the compounds had similar or even higher potencies,
despite quite large substituents. All compounds derivatized with fatty acids equal to or longer
than 12 carbon atoms were very protracted compared to GLP-1 and thus seem suitable for
once daily administration to type 2 diabetic patients. A structure-activity relationship was
obtained. GLP-1 could be derivatized with linear fatty acids up to the length of 16 carbon atoms,
sometimes longer, almost anywhere in the C-terminal part without considerable loss of potency.
Derivatization with two fatty acid substituents led to a considerable loss of potency. A structure-
activity relationship on derivatization of specific amino acids generally was obtained. It was
found that the longer the fatty acid, the more potency was lost. Simultaneous modification of
the N-terminus (in order to obtain better metabolic stability) interfered with fatty acid
derivatization and led to loss of potency.
In tr od u ction
ics are, among other things, characterized as having
insufficient â-cell capacity. No existing drugs for type 2
diabetes approach this problem. With this long list of
beneficial effects, GLP-1 compounds have a large po-
tential as a new class of drugs for the treatment of type
2 diabetes and impaired glucose tolerance (IGT).
Glucagon-like peptide-1 (GLP-1) is an incretin and is
produced by the L-cells of the intestine. Since its
discovery in 1984, GLP-1 has received much attention
as a possibly new treatment for type 2 diabetes.1-7
GLP-1 stimulates insulin secretion and biosynthesis and
inhibits glucagon release, each very favorable effects in
the treatment of hyperglycemia. Most importantly, both
these effects are glucose-dependent8,9 and therefore
represent a very safe way of lowering increased blood
glucose. In addition, GLP-1 inhibits gastric emptying,10-12
thereby decreasing postprandial glucose excursions.
Gastric acid secretion is inhibited, too,10 and thus,
GLP-1 compounds may provide protection against gas-
tric ulcers. Also, GLP-1 has been shown to be a potent
appetite suppressant,13-15 although effects on body
weight in humans still need to be shown. An effect on
body weight in clinical use would make the GLP-1
receptor an even more attractive target, since the
majority of type 2 diabetic patients are obese. Lately,
GLP-1 has been shown be able to stimulate growth and
proliferation of pancreatic â-cells.16 Overt type 2 diabet-
GLP-1 is a 30-amino acid peptide hormone. Generally,
peptide hormones this size are not orally available and
thus need to be administered by injection or through
an alternative way feasible for peptides (pulmonal,
buccal). Due to its strong tendency to fibrillate, GLP-1,
like its close analogue glucagon, is a very difficult
molecule to handle in solution. In a recent publication,
it was described how problematic it is to formulate a
GLP-1 compound and thus how difficult it may become
to ever make a drug out of GLP-1.17 The metabolic and
pharmacokinetic properties of GLP-1 add to these
problems. Dipeptidyl peptidase IV (DPP-IV) rapidly
degrades GLP-1(7-36)amide,18,19 rendering the rest of
the molecule, GLP-1(9-36)amide, inactive. Indeed,
GLP-1(9-36)amide may even act as an antagonist.20
Simultaneously, the kidneys clear GLP-1 quickly. The
half-life of GLP-1(7-36)amide in humans has been
determined to be 1.5 min after iv administration and
1.5 h after sc administration.3 N-Terminally modified,
DPP-IV-resistant analogues are of course still subject
to renal clearance, and as a result of this, they have
half-lives of only 4-5 min.21
* To whom correspondence should be addressed. Tel: +45 44 43
4788. Fax: +45 44 43 4587. E-mail: lbkn@novo.dk.
† Department of Molecular Pharmacology.
‡ Department of Protein Chemistry.
§ Department of Medicinal Chemistry.
| Department of Assay & Cell Technology.
Department of Pharmacokinetics.
10.1021/jm9909645 CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/18/2000

Potent Derivatives of GLP-1
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 9 1665
Ta ble 1. List of All Compounds and Their Potency Measured Using the Cloned Human GLP-1 Receptor Expressed in BHK Cellsa
compd
parent peptide
acyl site
acyl substituent
potency (EC50, pM)
1
2
GLP-1(7-37)
none
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
7000 ( 7
16700 ( 3700
3050 ( 350
177 ( 52
154 ( 66
72 ( 0.7
194 ( 24
22.0 ( 7.1
27.3 ( 8.4
687 ( 129
84.4 ( 22.1
113 ( 3
K8R26,34-GLP-1(7-37)
K18R26,34-GLP-1(7-37)
K23 R26,34-GLP-1(7-37)
R34-GLP-1(7-37)
K8
γ-Glu-C16
γ-Glu-C16
γ-Glu-C16
γ-Glu-C16
γ-Glu-C16
γ-Glu-C16
γ-Glu-C16
γ-Glu-C16
bis-C16-diacid
bis-γ-Glu-C16
bis-γ-Glu-C14
bis-C12-diacid
C16-diacid
C14-diacid
γ-Glu-C18
γ-Glu-C14
γ-Glu-C12
γ-Glu-C16
GABA-C16
â-Ala-C16
Iso-Nip-C16
γ-Glu-C16
C8
3
K18
K23
K26
K27
K34
K36
K38
K26,34
K26,34
K26,34
K26,34
K26
K26
K26
K26
K26
K26
K26
K26
K26
K34
K34
K34
K36
K36
K36
K38
K38
K38
K38
K38
K38
K38
K38
4
5
6
K27R26,34-GLP-1(7-37)
R26-GLP-1(7-37)
7
8
K36R26,34-GLP-1(7-36)
R26,34-GLP-1(7-38)
GLP-1(7-37)
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
GLP-1(7-37)
GLP-1(7-37)
GLP-1(7-37)
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)
desamino-H7R34-GLP-1(7-37)
R34-GLP-1(7-37)
R34-GLP-1(7-37)
R34-GLP-1(7-37)
410 ( 120
2360 ( 370
236 ( 66
169 ( 1
210 ( 14
7.89 ( 1.21
116 ( 3
desamino-H7R26-GLP-1(7-37)
desamino-H7R26-GLP-1(7-37)
desamino-H7R26-GLP-1(7-37)
K36R26,34-GLP-1(7-36)
K36R26,34-GLP-1(7-36)
K36R26,34-GLP-1(7-36)
R26,34-GLP-1(7-38)
R26,34-GLP-1(7-38)
R26,34-GLP-1(7-38)
R26,34-GLP-1(7-38)
G8R26,34-GLP-1(7-38)
E37R26,34-GLP-1(7-38)
E37G8R26,34-GLP-1(7-38)
E37G8R26,34-GLP-1(7-38)
γ-Glu-C8
C20-diacid
C16-diacid
γ-Glu-C18
C16-diacid
C12-diacid
γ-Glu-C18
γ-Glu-C14
γ-Glu-C16
γ-Glu-C16
γ-Glu-C16
γ-Glu-C18
5.60 ( 3.5
4.19 ( 0.98
115 ( 21
54 ( 1
328 ( 14
27.2 ( 0.1
135 ( 7
213 ( 30
a
Abbreviations used for acyl groups in lysine Nꢀ-acylated peptides: γ-Glu-C8 ) γ-L-glutamoyl(NR-octanoyl); γ-Glu-C14 ) γ-L-
glutamoyl(NR-tetradecanoyl); γ-Glu-C16 ) γ-L-glutamoyl(NR-hexadecanoyl); γ-Glu-C18 ) γ-L-glutamoyl(NR-octadecanoyl); C8 ) octanoyl;
C12-diacid ) ω-carboxyundecanoyl; C16-diacid ) ω-carboxypentadecanoyl; C20-diacid ) ω-carboxynonadecanoyl; GABA-C16 ) γ-ami-
nobutyroyl(Nγ-hexadecanoyl); Iso-Nip-C16 ) 1-(hexadecanoyl)piperidyl-4-carboxy. Data are given as mean ( SD of two individual
experiments with triplicate samples.
The principle of fatty acid derivatization has been
used to protract the action of insulin by facilitating
binding to serum albumin.22-25 In this article we
describe the structure-activity relationship (SAR) of a
large number of fatty acid derivatives of GLP-1. We
show that these compounds have pharmacokinetic
properties in pigs suitable for once daily administration.
In the absence of a nonpeptide agonist these compounds
may prove to be the best drug candidates for the treat-
ment of type 2 diabetes via the GLP-1 receptor target.
F igu r e 1. Amino acid sequence of GLP-1.
The amino acid sequence of GLP-1 can be seen in
Figure 1. GLP-1(7-37) and close analogues thereof were
derivatized on position 8, 18, 23, 26, 27, 34, 36, or 38
with fatty acids and optionally a spacer (Table 1). The
SAR of the compounds was investigated using a func-
tional assay employing the cloned human GLP-1 recep-
tor expressed in baby hamster kidney (BHK) cells.
Figure 2 shows examples of the dose-response curves
for a few selected compounds. All compounds tested
were full agonists and were shown to activate the GLP-1
receptor selectively (by using the cloned human gluca-
gon receptor, data not shown). Fatty acids have been
shown to interact with cell membranes as well as
albumin. However, the selective activation of the GLP-1
receptor evidences that this phenomenon is not impor-
tant for these compounds.
All compounds acylated with a fatty acid equal to or
longer than 12 carbon atoms were considerable pro-
tracted compared to GLP-1, which had a half-life after
sc administration of only 1.2 h. Table 2 shows plasma
half-lives after sc administration to pigs for a selection
of very potent compounds (4, 5, 7, 8, 18, 20, 21, 27, 35).
All had half-lives equal to or longer than 9 h. Bioavail-
ability was measured for selected compounds only and
was shown to be in the order of 50% and above (data
Discu ssion
The acylated compounds were all synthesized as part
of a program aimed at protracting the action of the
peptide hormone GLP-1. Fatty acids or fatty diacids,
optionally extended with a “spacer” between the ꢀ-amino
group of the lysine side chain and the carboxyl group of
the fatty acid, were used. Acylation with mono activated
esters of symmetrical diacids presented a potential
synthesis problem, which was solved by using an excess
of diacid versus N-hydroxysuccinimide in the activation
step. Acylation with simple fatty acids increases the net
negative charge of the resulting molecule by 1, whereas
peptides acylated with a L-glutamoyl spacer or with
diacids provide a net increase of the negative charge
by 2. In the two latter cases, an enhanced effect on
binding to albumin is predicted.22,26 This extra negative
charge added to the acylated molecule is also expected
to provide a higher solubility at physiological pH.

1666 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 9
Knudsen et al.
29 were important in the C-terminus.27 Even though
only position 8 in the N-terminus was acylated and this
compound found has a poor potency, the present results
nicely support those earlier reported from this group.27
Nu m ber of F a tty Acid Su bstitu en ts. The natural
sequence of GLP-1 has two lysines. When using an
analogue of a naturally occurring hormone, there is
always a risk of introducing a change in the amino acid
sequence, which can lead to an immune response to the
analogue. In order not to have to use analogues of GLP-
1, derivatization with two fatty acids was examined
(10-13). Only 13 with a C12 fatty acid had an accept-
able potency (177 pM). Derivatization with C14 or C16
fatty acids, either as diacids or as monoacids with a
γ-Glu spacer, resulted in compounds with a marked loss
of activity (10-12).
Com p ou n d s Der iva tized on Lysin e 26. The poten-
cies of the compounds were comparable when looking
at a series of different length diacids (14, 15) or fatty
acids with the same spacer (5, 16-18). Within the γ-Glu
spacer monoacid series (5, 16-18), derivatization with
a C18 acid (16, 194 pM) led to a significant loss of
activity compared to C16 (5, 61 pM), C14 (17, 22 pM),
and C12 (18, 27 pM). Within the diacid series (14, 15),
the diacid could be no longer than a C14 (15, 72 pM)
before a loss in potency (14, 154 pM), compared to the
γ-Glu spacer monoacid series (17, 18, 22-27 pM), was
seen. Earlier, this group and others have tried to modify
the amino terminus of GLP-1 in order to make the
molecule more resistant to enzymatic breakdown.21,28,29
Desaminohistidine7 represents one of the more potent
suggestions to a modification giving metabolic stabil-
ity.29 Nevertheless, as seen when comparing 19 (687
pM) to 5 (61 pM), considerably more potent compounds
could be obtained by not modifying the N-terminus
when a combination with acylation was desired. This
could of course be caused both by the position of the fatty
acid and because of the modified histidine. However as
shown below, further data from the position 34 acylation
series showed that acylation with a short chain fatty
acid (C8) led to a considerable more potent compound
than a corresponding C16 fatty acid.
F igu r e 2. Dose-response curves of GLP-1 and selected
compounds using the cloned human GLP-1 receptor expressed
in BHK cells: 1 (GLP-1) (9), 11 (0), 13 (b), 27 (O). Data are
from one representative experiment with triplicate samples
and are shown as mean ( SD.
Ta ble 2. Plasma Half-Lives in Pigs of GLP-1 and Selected
Potent Acylated Compoundsa
compd
plasma t1/2 (h)
compd
plasma t1/2 (h)
1 (GLP-1)
1.2
18
20
21
27
35
15 ( 3
31 ( 4
8.8 ( 1
13 ( 4
11 ( 1
4
5
7
8
20 ( 2
14 ( 2
13
12 ( 1
a
The half-lives were calculated from individual pigs after a
single sc injection. Each compound was injected in two pigs. Data
are shown as mean ( SD. For 7, half-life could only be calculated
from one pig.
Com p ou n d s Der iva tized w ith Differ en t Sp a cer s.
The γ-Glu spacer is optically active. Thus, it presents a
greater analytical challenge when upscaling the com-
pounds for good manufacturing production guidelines.
We therefore investigated other spacers without optical
activity. A GABA spacer (20, 84 pM) gave a compound
with equal affinity to the γ-Glu spacer (5, 61 pM). A
â-alanine spacer reduced the activity slightly (21, 113
pM) and a piperidyl-4-carboxy spacer resulted in a 6-fold
lower activity (22, 410 pM).
F igu r e 3. Pharmacokinetic profile of selected compounds
after sc administration to pigs: 1 (GLP-1) (9), 5 (O), 7 (b), 8
(0). Two pigs were dosed with compounds, and the following
immunoassay was performed in duplicate. Data are expressed
as mean ( SD.
not shown). Figure 3 illustrates the dramatic difference
in plasma half-lives between GLP-1 and three potent
acylated compounds (5, 7, 8).
Site of Acyla tion . Many different positions in the
C-terminal part of GLP-1 could be derivatized with quite
long fatty acids (3-9, 30-121 pM) without affecting the
potency. Binding affinity was not measured for these
compounds as they all bind to albumin as part of their
mechanism of protraction and it has not been possible
to set up a reproducible binding assay without albumin.
Derivatizing amino acids in the N-terminal part of the
peptide, as exemplified in 2 (1260 pM), led to a sub-
stantial loss of potency. Earlier findings from this group
showed that positions 7, 10, 12, 13, and 15 in the
N-terminal part were important for binding to and
activation of the receptor, whereas only positions 28 and
Com p ou n d s Der iva tized on Lysin e 34. Acylation
of lysine 34 (7, 23-25) generally led to compounds a
little less potent than those on lysine 26. Comparing 7
(121 pM) with 5 (61 pM) showed a 2-fold difference in
the potency when acylating with a γ-Glu spacer and a
C16 fatty acid. This difference can of course also be
explained by the respective arginine substitutions.
Acylating with a γ-Glu spacer and a C16 fatty acid when
combined with the above-discussed N-terminal modifi-
cation led to a 20-fold loss of potency (23, 2360 pM vs
7). Using a short chain fatty acid, the combination could
be employed without too much loss of activity (24, 236

Potent Derivatives of GLP-1
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 9 1667
The peptides were cleaved from the resin and side chain
deprotected in TFA/phenol/thioanisole/water/ethanedithiol
(83.25:6.25:4.25:4.25:2.0) for 180 min. The cleavage mixtures
were filtered and the filtrates were concentrated in a stream
of nitrogen. The crude peptides were precipitated from the
residual oil with diethyl ether and washed twice with diethyl
ether. After drying, the crude peptides were dissolved in 50%
aqueous acetic acid, diluted to 10% with water and purified
by semipreparative HPLC (Waters, Millipore) on a 25- × 250-
mm column packed with 7-µm C-18 silica. The column was
eluted with a gradient of acetonitrile against 0.05 M (NH4)2-
SO4, pH 2.5, at 10 mL/min at 40 °C. The peptide-containing
fractions were collected, diluted with 3 volumes of water and
applied to a Sep-Pak C18 cartridge (Waters part. 51910) which
was equilibrated with 0.1% aqueous TFA. The peptide was
eluted from the Sep-Pak cartridge with 70% acetonitrile/0.1%
TFA in water and isolated from the eluate by lyophilization
after dilution with water. The final product obtained was
characterized by amino acid analysis, analytical RP-HPLC and
MALDI-MS. Amino acid analysis and mass spectrometry data
agreed with the expected structure within the experimental
error of the methods (mass spectrometry ( 3 mass units,
amino acid analysis ( 10%). RP-HPLC showed a peptide purity
>95%.
pM vs 25, 169 pM). However, the C8 fatty acid will not
lead to much protraction (data not shown).
Com p ou n d s Der iva tized on Lysin e 36. Lysine 36
could be derivatized with quite long fatty acids (26, 210
pM) without too much loss of activity. The longer the
fatty acid, irrespective of whether diacids or γ-Glu-
spaced monoacids were used, the more loss of activity
was observed. Going from a C16 diacid to a C20 shifts
the potency from 7.9 to 210 pM (27 vs 26). A C16 γ-Glu-
spaced monoacid and the corresponding C18 decreases
potency from 36 to 116 pM (8 vs 28). Acylation of lysine
36 with a γ-Glu-spaced C16 acid (8, 36 pM) led to a more
potent compound than the same modification on position
26 (5, 61 pM) or 34 (7, 121 pM).
Com p ou n d s Der iva tized on Lysin e 38. The com-
pounds derivatized on lysine 38 (9, 29-36) were also
very potent. In this series we tried some further
modifications. Substitution of glycine instead of alanine
at position 8 gives stability toward enzymatic break-
down.21 However, as expected from earlier studies,21 this
also leads to loss of activity (33, 328 pM vs 9, 53 pM).
Substitution of glycine in position 37 with glutamic acid
was investigated because adding a negative charge
adjacent to the fatty acid may increase binding of fatty
acid-derived peptides to albumin.22,26 This substitution
was possible without loss of activity; in fact, it actually
seemed to increase activity (34, 27 pM vs 9), but when
combined with the glycine substitution, again some
activity was lost (35, 135 pM).
The RP-HPLC analyses were performed using UV detection
at 214 nm and a Vydac 218TP54 4.6- × 250-mm, 5-µm C-18
silica column which was eluted at 1 mL/min at 42 °C. Two
different elution conditions were used: (A) a gradient of 5-60%
acetonitrile against 0.1 M ammonium sulfate in water, pH 2.5
and (B) a gradient of 5-60% acetonitrile, 0.1% TFA against
0.1% TFA in water. In this manner the following starting
peptides were prepared:
R26-GLP-1(7-37)
E37G8R26,34-GLP-1(7-38)
K27R26,34-GLP-1(7-37)
K23R26,34-GLP-1(7-37)
K18R26,34-GLP-1(7-37)
K8R26,34-GLP-1(7-37)
E37R26,34-GLP-1(7-38)
R34-GLP-1(7-37)
Con clu sion
K36R26,34-GLP-1(7-36)
R26,34-GLP-1(7-38)
We found that the peptide hormone GLP-1 could be
derivatized almost anywhere in the C-terminal part of
the peptide. Derivatization with both short and long
fatty acids and amino acid-derived spacers led to
compounds that were highly potent. Several compounds
were both very potent and had plasma half-lives above
10 h, making them suitable as drugs for treatment of
type 2 diabetes using once daily administration.
desamino-H7R26-GLP-1(7-37)
desamino-H7R34-GLP-1(7-37)
G8R26,34-GLP-1(7-38)
Syn th esis of In ter m ed ia tes. Alk a n ed ioic Acid Mon o-
su ccin im id yl Ester . The alkanedioic acid monoesters were
prepared from the corresponding alkanedioic acids according
to the procedure described in the literature.30
Nr-Alk a n oyl-L-glu ta m ic Acid r-ter t-Bu tyl Ester γ-Su c-
cin im id yl Ester . The NR-alkanoyl-L-glutamic acid R-tert-butyl
ester γ- succinimidyl esters were prepared form L-glutamic acid
R-tert-butyl ester and the corresponding alkanoic acids accord-
ing to the procedure described in the literature.31
Exp er im en ta l Section
In str u m en ta tion . Analytical HPLC analysis of acylated
GLP-1 analogues was performed using UV detection at 214
nm and a Vydac 214TP54 4.6- × 250-mm, 5-µm C-4 silica
column at 42 °C. The column was eluted at 1 mL/min with a
gradient of 0-100% acetonitrile, 0.07% TFA against 0.1% TFA
in water.
Mass spectral data were obtained using a Voyager RP
MALDI-TOF (matrix assisted laser desorption ionization time-
of-flight) instrument (Perseptive Biosystems Inc., MA) equipped
with a nitrogen laser (337 nm). The instrument was operated
in linear mode with delayed extraction and the accelerating
voltage in the ion source was 25 kV. Sinapinic acid was used
as matrix and calibration was performed using external
standards with a resulting mass accuracy of 0.1%.
Syn th esis of GLP -1 An a logu e Sta r tin g P ep tid es. Solid-
phase methodology was used to synthesize these analogues.
The peptides were synthesized according to the Fmoc strategy
on an Applied Biosystems 431A peptide synthesizer in 0.25-
mmol scale using the manufacturer supplied FastMoc UV
protocols starting with either a Fmoc-Gly-Wang resin, Fmoc-
Lys(Boc)-Wang resin or Rink-amide resin (NovaBiochem). The
protected amino acid derivatives used were commercially
obtained Fmoc amino acids. The derivatives used, where side
chain protection was needed, were: Fmoc-Arg(Pmc), Fmoc-
Trp(Boc), Fmoc-Glu(OBut), Fmoc-Lys(Boc), Fmoc-Gln(Trt),
Fmoc-Tyr(But), Fmoc-Ser(But), Fmoc-Thr(But), Fmoc-His(Trt),
Adoc-imidazolylpropionic acid and Fmoc-Asp(OBut).
γ-(Nγ-Hexa d eca n oyla m in o)bu tyr ic Acid Su ccin im id yl
Ester . A mixture of hexadecanoic acid succinimidyl ester (3
g, 8.48 mmol), prepared as described in the literature,32 and
4-aminobutyric acid (0.87 g, 8.48 mmol) in DMF (200 mL) was
stirred at 25 °C for 60 h. The reaction mixture was filtered
and the filtrate was added drop by drop to a 10% aqueous
solution of citric acid (500 mL). The precipitate was collected
and dried in vacuo. To a suspension of the residue in DMF
(35 mL) was added a solution of DCC (1.45 g, 7 mmol) and
HONSu (0.89 g, 7.7 mmol) in dichloromethane (20 mL). The
resulting mixture was stirred at 25 °C for 20 h and then
filtered. The solvent was evaporated to give a solid residue,
which after recrystallization from a mixture of n-heptane (50
mL) and 2-propanol (2.5 mL) gave the title compound (2.5 g,
75%).
Syn th esis of Acyla ted GLP -1 An a logu es 2-36. All acyl-
modified peptides, except 20 and 24, were synthesized accord-
ing to general Methods A and B. Peptide sequence, site of
acylation, acyl substituent, yields, and physical data are given
in Tables 1 and 3.
Meth od A. A mixture of starting peptide (1 equiv), DIPEA
(28 equiv), NMP (70 µl/µmol starting peptide) and water (66
µL/µmol starting peptide) was gently shaken for 5 min at 25
°C. To the resulting mixture was added a solution of al-
kanedioic acid monosuccinimidyl ester in NMP (22 µL/µmol

1668 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 9
Knudsen et al.
Ta ble 3. List of All Compounds with Method of Synthesis,
Purification Yield, and Estimated Purity as Determined by
Analytical HPLC and Measured and Theoretical Molecular
Weight
containing the product were isolated, and lyophilized to give
the final product. Characterization was performed by HPLC
and mass spectral analysis (Table 3).
R26K34-(NE-(γ-Am in obu tyr oyl(Nγ-h exa d eca n oyl)))-GLP -
1(7-37)-OH, 20. To a mixture of R26-GLP-1(7-37)-OH (41.1
mg, 12.2 µmol), DIPEA (44 mg, 341 µmol), NMP (5.76 mL)
and water (2.88 mL) was added a solution of γ-(Nγ-hexade-
canoyl)aminobutyric acid succinimidyl ester (16 mg, 36.5
µmol), prepared as described above, in NMP (400 µL). The
reaction mixture was gently shaken at 25 °C for 5 min and
then allowed to stand at 25 °C for an additional 1.5 h. The
reaction was quenched by the addition of a solution of glycine
(20 mg, 268 µmol) in water (200 µL). The solvent was
evaporated and the residue purified by preparative chroma-
tography using a cyanopropyl column (Zorbax 300SB-CN) and
a standard acetonitrile/TFA system. The column was heated
to 65 °C and the acetonitrile gradient was 0-100% in 60 min.
Fractions containing the product were isolated and lyophilized
to provide 20 (15.8 mg, 35%). Characterization was performed
by HPLC and mass spectral analysis (Table 3).
Desa m in o-H7R26K34-(NE-octa n oyl)-GLP -1(7-37)-OH, 24.
To a mixture of desamino-H7R26-GLP-1(7-37)-OH (19.8 mg,
5.89 µmol), DIPEA (21.2 mg, 164 µmol), NMP (1.38 mL) and
water (1.38 mL) was added a solution of octanoic acid succin-
imidyl ester (20.7 mg, 36.5 µmol), prepared as described in
the literature,33 in NMP (106 µL). The reaction mixture was
gently shaken for 5 min at 25 °C and then allowed to stand
for an additional 1.5 h at 25 °C. The reaction was quenched
by the addition of a solution of glycine (9.7 mg, 129 µmol) in
water (97 µL. The solvent was evaporated and the residue
purified by preparative chromatography using a cyanopropyl
column (Zorbax 300SB-CN) and a standard acetonitrile/TFA
system. The column was heated to 65 °C and the acetonitrile
gradient was 0-100% in 60 min. Fractions containing the
product were isolated and lyophilized to provide 24 (9.2 mg,
45%). Characterization was performed by HPLC and mass
spectral analysis (Table 3).
Recep tor Exp er im en ts. Baby hamster kidney (BHK) cells
expressing the cloned human GLP-1 receptor (BHK 467-12A)
were grown in DMEM media with the addition of 100 IU/mL
penicillin,100 µL/mL streptomycin, 10% fetal calf serum and
1 mg/mL Geneticin G-418 (Life Technologies). Plasma mem-
branes were prepared by homogenization in buffer (10 mM
Tris-HCl, 30 mM NaCl and 1 mM dithiothreitol, pH 7.4,
containing, in addition, 5 mg/L leupeptin (Sigma), 5 mg/L
pepstatin (Sigma), 100 mg/L bacitracin (Sigma), and 16 mg/L
aprotinin (Calbiochem-Novabiochem, La J olla, CA)). The ho-
mogenate was centrifuged on top of a layer of 41% w/v sucrose.
The white band between the two layers was diluted in buffer
and centrifuged. Plasma membranes were stored at -80 °C
until used.
The functional receptor assay was carried out by measuring
cAMP as a response to stimulation by GLP-1 or derivatives.
Incubations were carried out in 96-well microtiter plates in a
total volume of 140 mL and with the following final concentra-
tions: 50 mM Tris-HCl, 1 mM EGTA, 1.5 mM MgSO4, 1.7 mM
ATP, 20 mM GTP, 2 mM 3-isobutyl-1-methylxanthine (IBMX),
0.01% w/v Tween-20, pH 7.4. Compounds were dissolved and
diluted in buffer. GTP was freshly prepared for each experi-
ment; 2.5 µg of membrane was added to each well and the
mixture was incubated for 90 min at room temperature in the
dark with shaking. The reaction was stopped by the addition
of 25 mL of 0.5 M HCl. Formed cAMP was measured by a
scintillation proximity assay (RPA 542, Amersham, UK).
Dose-response curves were plotted for the individual com-
pounds and EC50 values calculated using GraphPad Prism
software.
yield
mg
MW
compd method
%
HPLC purity (%) calcd found
1
2
3
B
B
B
B
B
B
B
B
A
B
B
A
A
A
B
B
B
B
a
28
15
24
6
16
30
23
47
14
20
9
9
12
43 11.5
34 15.5
29
36
41
35 15.8
55 11.6
44 12
35
45
27
8
11
27
26
16
11
26
38
26
29
32
3.2
1.9
3.1
0.2
2.1
6.0
3.1
7.2
4.8
2.4
27.3
1.0
0.6
>95
>95
>90
>95
>95
>90
>95
>80
>90
>90
>90
>90
>90
>90
>80
>90
>85
>95
>90
>95
>95
>95
>95
>90
>90
>90
>95
>90
>90
>90
>90
>95
>95
>90
>90
3837
3820
3779
3751
3778
3751
3694
3880
3893
4091
4035
3780
3653
3624
3779
3723
3695
3737
3707
3694
3733
3737
3496
3625
3651
3595
3722
3808
3752
3935
3880
3894
3981
3967
3995
3835
3823
3779
3751
3780
3751
3695
3880
3893
4091
4035
3780
3653
3626
3780
3722
3693
3738
3705
3694
3734
3737
3496
3625
3651
3595
3722
3808
3752
3933
3880
3894
3980
3966
3994
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
6.1
8.2
9.1
A
A
B
a
5.6
9.2
4.1
0.5
0.6
3.7
3.9
0.9
3.6
3.9
5.0
5.1
2.9
3.2
B
A
A
B
A
A
B
B
B
B
B
B
a
Specific method described below.
starting peptide); the reaction mixture was gently shaken for
5 min at 25 °C and then allowed to stand at 25 °C for an
additional 2.5 h. The reaction was quenched by the addition
of a solution of glycine (22 equiv) in 50% aqueous ethanol (165
µL/µmol starting peptide). The reaction mixture was purified
by preparative chromatography using a cyanopropyl column
(Zorbax 300SB-CN) and a standard acetonitrile/TFA system.
The column was heated to 65 °C and the acetonitrile gradient
was 0-100% in 60 min. Fractions containing the product were
isolated and lyophilized to give the final product. Character-
ization was performed by HPLC and mass spectral analysis
(Table 3).
Meth od B. A mixture of starting peptide (1 equiv), DIPEA
(28 equiv), NMP (245 µl/µmol starting peptide) and water (245
µL/µmol starting peptide) was gently shaken at 25 °C for 5
min. To the resulting mixture was added a solution of
NR-alkanoyl-L-glutamic acid R-tert-butyl ester γ-succinimidyl
ester (5.9 equiv), prepared as described above, in NMP (75 µL/
µmol starting peptide); the reaction mixture was gently shaken
at 25 °C for 5 min and then allowed to stand at 25 °C for an
additional 2 h. The reaction was quenched by the addition of
a solution of glycine (22 equiv) in 50% aqueous ethanol (165
µL/µmol starting peptide). A 0.5% aqueous solution of am-
monium acetate (12.5 mL/µmol starting peptide) was added,
and the resulting mixture was eluted onto a preequilibrated
Varian C8 Mega Bond Elut cartridge. The immobilized com-
pound was washed with 5% aqueous acetonitrile (3.75 mL/
µmol starting peptide) and finally liberated from the cartridge
by elution with TFA (1.5 mL/µmol starting peptide). The eluate
was allowed to stand at 25 °C for 1.75 h and then evaporated.
The residue was purified by preparative chromatography using
a cyanopropyl column (Zorbax 300SB-CN) and a standard
acetonitrile/TFA system. The column was heated to 65 °C and
the acetonitrile gradient was 0-100% in 60 min. Fractions
P h a r m a cok in etic Exp er im en ts. The experiments were
performed in Landrace × Duroc × Yorkshire pigs weighing
between 25 and 50 kg. All the GLP-1 analogues were admin-
istered sc at a dose of 0.5 nmol/kg; GLP-1 was administered
at a dose of 15 nmol/kg. Blood samples were collected by means
of a catheter placed in the jugular vein, according to the
following schedule: before dosing and 2, 4, 6, 8, 24, 36, 48 and

Potent Derivatives of GLP-1
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 9 1669
(10) Wettergren, A.; Schjoldager, B.; Mortensen, P. E.; Myhre, J .;
Christiansen, J .; Holst, J . J . Truncated GLP-1 (proglucagon 72-
107amide) inhibits gastric and pancreatic functions in man. Dig.
Dis. Sci. 1993, 38, 665-673.
(11) Nauck, M. A.; Niederereichholz, U.; Ettler, R.; Holst, J . J .;
Ørskov, C.; Ritzel, R.; Schmigel, W. H. Glucagon-like peptide-1
inhibition of gastric-emptying outweighs its insulinotropic effects
in healthy humans. Am. J . Physiol. 1997, 273, E981-E988.
(12) Nauck, M. A. Is GLP-1 an incretin hormone?. Diabetologia 1999,
42, 373-379.
(13) Flint, A.; Raben, A.; Astrup, A.; Holst, J . J . GLP-1 promotes
satiety and suppresses energy intake in humans. J . Clin. Invest.
1998, 101, 515-520.
(14) Na¨slund, E.; Gutniak, M. K.; Skogar, S.; Ro¨ssner, S.; Hellstro¨m,
P. M. GLP-1 increases the period of postprandial satiety and
slows gastric emptying in obese humans. Am. J . Clin. Nutr.
1998, 68, 525-530.
(15) Gutzwiller, J . P.; Go¨ke, B.; Drewe, J .; Hildebrand, P.; Ketterer,
S.; Handscin, D.; Winterhalder, R.; Conen, D.; Beglinger, C.
Glucagon-like peptide-1: a potent regulator of food intake in
humans. Gut 1999, 44, 81-86.
(16) Edvell, A.; Lindstro¨m, P. Initiation of increased pancreatic islet
growth in young normoglycaemic mice (Umeå +/?). Endocrinol-
ogy 1999, 140, 778-783.
(17) Clodfelter, D. K.; Pekar, A. H.; Rebhun, D. M.; Destrampe, K.
A.; Havel, H. A.; Myers, S. R.; Brader, M. L. Effects of nonco-
valent self-association on the subcutaneous absorption of a
therapeutic peptide. Pharm. Res. 1998, 15, 254-262.
(18) Deacon, C. F.; J ohnsen, A. H.; Holst, J . J . Degradation of
glucagon-like peptide-1 by human plasma in vitro yields an
N-terminally truncated peptide that is a major endogenous
metabolite in vivo. J . Clin. Endocrinol. Metab. 1995, 80, 952-
957.
(19) Kieffer, T. J .; McIntosh, C. H. S.; Pederson, R. A. Degradation
of GIP and truncated GLP-1 in vitro and in vivo by Dipeptidyl
Peptidase IV. Endocrinology 1995, 136, 585-3596.
(20) Knudsen, L. B.; Pridal, L. GLP-1(9-36)amide is major metabolite
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, 429-435.
(21) Deacon, C. F.; Knudsen, L. B.; Madsen, K.; Wiberg, F. C.;
J acobsen, O.; Holst, J . J . Dipeptidyl peptidase IV resistant
analogues of GLP-1 which have extended metabolic stability and
improved biological activity. Diabetologia 1998, 41, 271-278.
(22) Kurtzhals, P.; Havelund, S.; J onassen, I.; Kiehr, B.; Larsen, U.
D.; Ribel, U.; Markussen, J . Albumin-binding of insulins acylated
with fatty-acids - characterization of the ligand protein-interac-
tion and correlation between binding-affinity and timing of the
insulin effect in-vivo. Biochem. J . 1995, 312, 725-731.
(23) Markussen, J .; Havelund, S.; Kurtzhals, P.; Andersen, A. S.;
Halstøm, J .; Hasselager, E.; Larsen, U. D.; Ribel, U.; Schaffer,
L.; Vad, K.; J onassen, I. Soluble, fatty-acid acylated insulins bind
to albumin and show protracted action in pigs. Diabetologia
1996, 39, 281-288.
72 h after dosing. For GLP-1 additional blood samples were
collected 0.25, 0.5, 0.75, 1, 1.5 and 3 h after dosing; 1 mL of
blood samples were stabilized with 35 µL of 0.18 M EDTA,
pH 7.4, with the addition of 15000 KIE/mL aprotinin (Novo
Nordisk A/S) and 3% (w/v) bacitracin (Sigma). Plasma was
separated and stored at -18 °C until assayed. Plasma samples
were measured in a radiommunoassay with a polyclonal rabbit
antibody directed against the N-terminus of native GLP-1
(HER4, D. Bataille, Marseilles). The antibody was specific for
bioactive GLP-1. Samples were tested without prior extraction
by incubating a 30-µL sample or calibrator with 100 µL of
antibody dilution (1:2000) in phosphate buffer, pH 7.5, with
human serum albumin (1 g/L). The relevant peptide (GLP-1
or analogue) added to normal pig plasma was used as calibra-
tor (concentrations from 5000 to 39 pM). After incubation for
3 days at 4 °C, 200 µL of tracer solution was added ([125I]GLP-
1(7-36)amide labeled by the peroxidase method and diluted
in phosphate buffer with human serum albumin (Behring) to
app. 6000 cpm/200 µL). After a further 2 days of incubation,
free and antibody-bound tracer were separated by adding 1.5
mL of charcoal suspension to each tube, incubation for 60 min
at 4 °C and centrifugation. Supernatants were transferred to
new tubes and counted. Concentrations were calculated from
the calibrators using a four-parameter logistic curve fit using
the MultiCalc software from Wallac. Data were analyzed for
the individual pigs in each group by use of noncompartmental
methods using the PC-based software WinNonlin (version 2.1,
Scientific Consulting Inc.).
Ack n ow led gm en t. The authors thank Edson Celso
dos Santos, Lene von Voss, Lotte Gottlieb, Anne-Grethe
J uul, Dorte M. Gundesen, Mette Frost, and Karin
Hamborg Albertsen for their excellent technical as-
sistance. Bernard Thorens is thanked for cloning the
receptor, and Finn C. Wiberg is thanked for making the
stable cell line. Susanne Roug, Ole Kirk, Lone Pridal,
Peter Høngård Andersen, Søren E. Bjørn, Niels Fiil,
J ohan Selmer, Peter Kristensen, and the SoLonGLIP
and NN2211 Project Teams are all thanked for their
enthusiasm.
(24) Kurtzhals, P.; Havelund, S.; J onassen, I.; Kiehr, B.; Ribel, U.;
Markussen, J . Albumin-binding and time action of acylated
insulins in various species. J . Pharm. Sci. 1996, 85, 304-308.
(25) Myers, S. R.; Yakubumadus, F. E.; J ohnson, W. T.; Baker, J .
E.; Cusick, T. S.; Williams, V. K.; Tinsly, F. C.; Kriauciunas, A.;
Manetta, J .; Chen, V. J . Acylation of human insulin with
palmitic acid extends the time action of human insulin in
diabetic dogs. Diabetes 1997, 46, 637-642.
Refer en ces
(1) Nauck, M. A.; Heimesaat, M. M.; Ørskov, C.; Holst, J . J .; Ebert,
R.; Creutzfeldt, W. 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, 301-307.
(2) Nathan, D. M.; Schreiber, E.; Fogel, H.; Mojsov, S.; Habener, J .
F. Insulinotropic action of GLP-1-(7-37) in diabetic and non-
diabetic subjects. Diabetes Care 1992, 15, 270-276.
(3) Gutniak, M. K.; Linde, B.; Holst, J . J .; Efendic, S. Subcutaneous
injection of the incretin hormone GLP-1 abolishes postprandial
glycemia in NIDDM. Diabetes Care 1994, 17, 1039-1044.
(4) Nauck, M. A.; Kleine, N.; Ørskov, C.; Holst, J . J .; Willms, B.;
Creutzfeldt, W. Normalization of fasting hyperglycemia by
exogenous GLP-1 (7-36) amide in type 2-diabetic patients.
Diabetologia 1993, 36, 741-744.
(26) Kenyon, M. A.; Hamilton, J . A. 13C NMR studies of the binding
of medium-chain fatty acids to human serum albumin. J . Lipid
Res. 1994, 35, 458-467.
(27) Adelhorst, K.; Heedegaard, B. B.; Knudsen, L. B.; Kirk, O.
Structure activity studies of GLP-1. J . Biol. Chem. 1994, 269,
6275-6278.
(28) Buckley, D. I.; Habener, J . F.; Mallory, J . B.; Mojsov, S. GLP-1
analogues useful for diabetes treatment. U.S. Patent 5,545,618.
(29) Chen, V. J .; DiMarchi, R. D.; Smiley, D. L.; Stucky, R. D.;
Kriauciunas, A. V. Glucagon-Like insulinotropic peptide ana-
logues, compositions, and methods of use. U.S. Patent 5,512,-
549.
(5) Kreymann, B.; Ghatai, M. A.; Williams, G.; Bloom, S. R. GLP-1
7-36: a physiological incretin in man. Lancet 1987, II, 1300-
1304.
(6) Rachman, J .; Barrow, B. A.; Levy, J . C.; Turner, R. C. Near-
normalization of diurnal glucose concentrations by continuous
administration of GLP-1 in subjects with NIDDM. Diabetologia
1997, 40, 205-211.
(7) Larsen, J .; J allad, N.; Damsbo, P. One-week infusion of GLP-1
improves the glycemic control in NIDDM. Diabetes 1996, 45, 860.
(8) Qualmann, C.; Nauck, M.; Holst, J . J .; Ørskov, C.; Creutzfeldt,
W. Insulinotropic actions of intravenous glucagon-like peptide-1
(7-36) amide in the fasting state in healthy subjects. Acta
Diabetol. 1995, 32, 13-16.
(9) Toft-Nielsen, M.; Madsberg, S.; Holst, J . J . Exaggerated secretion
of GLP-1 could cause reactive hypoglycaemia. Diabetologia 1998,
41, 1180-1186.
(30) Ebashi, I.; Takigawa, T.; Inoue, M. Long chain carboxylic acid
imide ester. EP511600.
(31) Knudsen, L. B.; Nielsen, P. F.; Sørensen, P. O. GLP-1 derivatives
which have lipophilic substituent - exhibit protracted profiles
of action relative to known GLP-1 compounds and are useful in
treatment of diabetes. WO9808871.
(32) Paquet, A. Succinimidyl esters of fatty-acids for amino acid
acylations. Can. J . Chem. 1976, 54, 733-737.
(33) Suskovic, B.; Vajtner, Z.; Naumski, R. Synthesis and biological
activities of some peptidoglycan monomer derivatives. Tetrahe-
dron 1991, 47, 8407-8416.
J M9909645