`
`J. Med. Chem. 2000, 43, 1664-1669
`
`Expedited Articles
`
`Potent Derivatives of Glucagon-like Peptide-I with Pharmacokinetic Properties
`Suitable for Once Daily Administration
`
`Lotte B. Knudsen,*,r Per F. Nielsen/ Per 0. Huusfeldt,§ Nils L. Johansen,§ Kjeld Madsen,s Freddy Z. Pedersen)
`Henning Th0gersen ,9 Michael Wilken} a nd Henrik Agerso_j_
`
`Departments of.A1olecular Pharmacology. Protein Chemistry, lvfedicinal Chemistry. Assay & Cell Technolo:,.,ry, and
`Pharmacokinetics, Health Care D iscovery and Preclinical Development, Novo Nordisk A / S, Novo Park,
`DK-2760 Maaloev, Denmark
`
`Received December 3, 1999
`
`A series ofvery potent derivatives of the 30-amino acid peptide hormone glucagon-Iike peptide-]
`(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 (EC 50 ) of 5 5 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 ,vith 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-tenninal part without considerable loss of potency.
`Derivatization with two fa tty acid substituents led to a considerable loss of potency. A structure(cid:173)
`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.
`
`Introduction
`Glucagon-like peptide-! (GLP-I) 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-dependent 8·9 and therefore
`represent a very safe way of lowering increased blood
`glucose. In addition , GLP-1 inhibits gas tric emptying,10 12
`thereby decreasing postprandial glucose excursions.
`Gastric acid secretion is inhibited, too, 10 and thu.s,
`GLP-1 compounds may provide protection against gas(cid:173)
`tric ulcers . . Also, GLP-I 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 ,veight 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 p-cells 16 Over t type 2 dia bet-
`
`* To whom correspondence should be addressed. Tel: +45 44 43
`4788 . Fax: +45 44 43 4587. E -mail: lbkn(qhovo.dk.
`1 Deparl.ment of Molecular Pharmacology.
`i Department of Protein Ch emistry.
`§ Depa r tment of Medicinal Chemistry.
`11 Department of Assay & Cell Technology.
`J. Department of Pharmacokinetics.
`
`ics are, among other things , characterized as having
`insufficient µ-cell capacity. No existing drugs for type 2
`diabetes approach this problem . With this long list of
`beneficial effects, GLP-l compounds have a large po(cid:173)
`tentia l as a new class of drugs for the treatment of type
`2 diabet es and impaired glucose tolerance (1GT).
`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,
`bu.ccal). 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 vvas described how problematic it is to formulate a
`GLP-1 compound and t hus 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)a.mide , 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
`I .5 h after sc administration 3 N-Term.inally modified ,
`DPP-I V-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
`
`10.1021 /jm 9909645 CCC: Sl 9 .00 © 2000 American Chemical Society
`Published on Web 04/18/2000
`
`
`
`Novo Nordisk A/S Ex. 2011, P. 1
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`
`
`
`Journal ofAfedicinal Chemistry, 2000, Vol. 43, No. 9
`1665
`
`Expressed in BHK Cells"
`List of All Compounds and Their Potency Measured Using the Cloned Human GLP-1 Receptor
`
`
`Table 1.
`
`acyl site acyl substituent potency (ECso, pl\'[)
`com pd
`
`parent peptide
`55 ± 19
`
`GLP-1 (7 ·--3 7)
`none
`1
`KsR26.34_GLP-1(7-37)
`Ks
`1260J,2]0
`y-Glu-Cl6
`2
`K isR26.34_ffLP-l (7-37)
`Kl8
`35.2 ± 6.2
`y-Glu-Cl6
`3
`K2-'
`EY R26·34-GLP-1(7---
`·37)
`30. l± 3.3
`y-Glu-Cl6
`4
`K26
`R.34-GLP-1(7-37)
`61.0±7.1
`y-Glu-C16
`5
`K27
`K27R26,34_GLP-1(7-·37)
`36.3 ± 0.3
`y-Glu-Cl6
`6
`K34
`R26-GLP-1(7-37)
`121:±:26
`7
`y-Glu-Cl6
`K36R26,34-GLP-1(7-36
`!(36
`y-Glu-Cl6
`36.4 ± 2.l
`)
`8
`K3S
`R26·34-GLP-1(7--38)
`53.0 ± 2.8
`
`y-Glu-C 16
`9
`K2G,34
`7000 ± 7
`bis-C 16-diacid
`GLP-1(7-37)
`10
`K20,:,4
`16700 ± 3700
`bis-y-Glu-C16
`GLP-1(7---37)
`11
`K::6,34
`3050 J: 350
`bis-y-Glu-Cl4
`GLP-1(7-37)
`12
`!(26,34
`177 ± 52
`bis-C'l2-diacid
`GU'-1(7-37)
`l3
`K26
`
`R 34-GLP-1(7--3 7)
`C16-diacid
`154 ± 66
`14
`K26
`R34-GLP-1(7-37)
`72 ± 0.7
`Cl4-diacid
`15
`K2c
`R 34-GLP-1(7-
`---37)
`y-Glu-Cl8
`194 ± 24
`16
`K26
`R34-GLP-1(7-37)
`y-Glu-Cl4
`22.0c±:7.l
`17
`!(26
`R34-GLP-1(7-37)
`27.3 ± 8.4
`y-Glu-Cl2
`l8
`K26
`R 3<GLP-H7 ---3 7)
`y-Glu-Cl6
`687 ± 129
`desamino-II'
`, ,
`19
`K2G
`R.34-GLP-1(7-37)
`GABA-Cl6
`84.4 ± 22. l
`20
`R 34-GLP-1(7---37)
`K26
`f:/-Ala-Cl6
`113 ± 3
`21
`K2"
`R34-GLP-1(7-37)
`
`Iso-Nip-C 16
`410 + 120
`22
`K34
`dcsamino-H7R 26-GLP-1(7-3
`7)
`y-Glu-Cl6
`2360 ± 370
`23
`K34
`des am in o-H7R26-GLP-1(7 ---3 7)
`236 ± 66
`24
`des am in o-H7R 26-GLP-l
`K34
`y-Glu-C8
`(7-37)
`169 ± 1
`25
`K3c
`K36R26,34-GLP-1(7
`-·36\
`C20-diacid
`210 ± 14
`26
`K36
`K36R26,34_GLP-1(7-36)
`Cl6-diacid
`7.89 :±: 1.21
`27
`K36R26,34-GLP-1(7-36
`!(36
`116 ± 3
`y-Glu-Cl8
`)
`28
`K3S
`R26·34-GLP-1(7--38)
`C16-diacid
`5.60 ± 3.5
`29
`.K38
`R26•34-GLP-l
`(7-38)
`CI 2-diacid
`4.19+0.98
`30
`K3s
`R26·34-G LP-1(7---38)
`y-Gh1-Cl8
`115+21
`31
`K38
`
`R 26 34-GLP-l (7-3 8)
`y-Glu-Cl4
`54 ,±: 1
`32
`GsR26,34_GLP-l (7-38)
`K1s
`y-Glu-Cl6
`328 ± 14
`33
`K3S
`E 37R 26.34-GLP-1(7
`----38)
`y-Glu-Cl6
`27.2 ± 0.1
`34
`.K38
`E37GsR 26,34_GLP-l (7-3 8)
`135±7
`y-Glu-Cl6
`35
`K3s
`E11asR2c,34_GLP-l (7---3 8)
`y-Glu-Cl8
`213 ± 30
`36
`
`
`
`
`
`
`
`"Abbreviations used for acyl groups in lysine N'-acylatcd peptides: y-Glu-C8 = y-L-glutamoyl(N°--octanoyl); y-Glu-Cl4 = y-L
`
`
`
`
`
`glutamoyl(N"-tetradecanoyl); y-Glu-Cl6 ""' y-L-glutamoyl(JV"-hexadecanoyl); y-Gh1-Cl8 = y-L-glutamoyl(N"-octadecanoyl); CS"'' octanoyl;
`
`
`
`
`
`
`
`
`
`nonadecanoyl: C20-diacipcntadecanoyl: Cl6-diacid Cl2-diacid undecanoyl; = oJ-carboxy = m-carboxy d = oJ-carboxy GABA-Cl6 = y-ami
`
`
`
`
`
`nobutyroyl(1VY-hexadecanoyl): Iso-Nip-Cl6 = l-(hexadecanoyl)piperidyl-4-carbm,y. Data are g iven as mean J, SD of two individual
`
`experiments v,ith triplicate samples.
`The principle of fatty acid derivatization has been
`
`
`
`
`
`used to protract the action of insulin by facilitating
`
`binding to serum alburnin.22 25 Jn this article we
`
`
`Gly-Gln-Ala-Ala25-Lys-Glu-Phe-lle-Ala"'-Trp-Leu-Val-Lys-Gly-Arg-Gly"
`
`
`
`describe the structure-activity relationship (SAR) of a
`
`Figure L Amino acid sequence ofGLP-1.
`
`large number of fatty acid derivatives of GLP-1. We
`The amino acid sequence of GLP-1 can be seen in
`
`show that these compounds have pharmacokinetic
`
`
`
`Figure 1. GLP-1(7-37) and close analogues thereof were
`
`
`
`properties in pigs suitable for once daily administration.
`
`
`deriva tized on position 8, 18, 23, 26, 27, 34, 36, or 3 8
`
`
`In the absence ofa nonpeptide agonist these compounds
`
`with fatty acids and optionally a spacer (Table 1). The
`
`may prove to be the best drug candidates for the treat
`
`
`SAR of the compounds was investigated using a func
`
`
`ment of type 2 diabetes via the GLP-1 receptor target.
`
`tional assay employing the cloned human GLP-1 recep
`
`tor expressed in baby hamster kidney (BHK) cells.
`Discussion
`
`
`Figure 2 shows examples of the dose--response curves
`
`
`
`for a fo,v selected compounds. All compounds tested
`
`
`The acylated compou.nds were all synthesized as part
`
`
`were full agonists and were shown to activate the GLP-1
`
`of a program aimed at protracting the action of the
`peptide hormone GLP-1. Fatty acids or fatty diacids,
`
`
`receptor selectively (by using the cloned human gluca
`
`gon receptor, data not shown)_ Fatty acids have been
`
`
`
`optionally extended with a ''spacer" between the E-am ino
`
`shoYvn to interact with cell membranes as well as
`
`group of the lysine side chain and the carboxyl group of
`
`
`
`albumin. However, the selective activation of the GLP-1
`
`the fatty acid, were used. Acylation with mono activated
`
`
`
`receptor evidences that this phenomenon is not impor
`
`
`
`esters of symmetrical diacids presented a potential
`tant for these compounds.
`
`
`synthesis problem, which was solved by using an excess
`
`
`All compounds acylated with a fatty acid equal to or
`
`ofdiacid versus /'./-hydroxysuccinirnide in the activation
`
`
`longer than 12 carbon atoms vvere considerable pro
`
`
`step. Acylation with simple fatty acids increases the net
`
`
`tracted compared to GLP-1, which had a half-life after
`
`
`
`negative charge ofthe resulting molecule by 1, whereas
`
`sc administration of only 1 .2 h. Table 2 shows plasma
`
`
`
`peptides acylated with a L-glutamoyl spacer or with
`
`
`half-lives after sc administration to pigs for a selection
`
`
`
`
`diacids provide a net increase of the negative charge
`by 2. In the two latter cases, an enhanced effect on
`
`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
`
`binding to albumin is predicted.
`6 This extra negative
`
`
`
`ability was measured for selected compounds only and
`
`
`charge added to the acylated molecule is also expected
`
`was shown to be in the order of 50'1/,, and above (data
`
`
`
`to provide a higher solubility at physiological pH.
`
`
`
`His7-Ala-Glu-Gly10-Thr-Phe-Thr-Ser-Asp 15-Val-Ser-Ser-Tyr-Leu20-Glu
`
`,2
`
`22
`
`
`
`Potent Derivatives ofGLP-1
`
`cs
`
`
`
`
`
`Novo Nordisk A/S Ex. 2011, P. 2
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`
`
`1666 Journal ofMedicin al Chemistry, 2000, Vol. 43. No. 9
`
`Knudsen el al.
`
`50
`
`40
`
`o.. 30
`2
`<(
`" 20
`
`10
`
`0
`
`-14
`
`-13
`
`-12
`
`-8
`-9
`-10
`-11
`Concentration (log(M))
`
`-7
`
`-6
`
`-5
`
`Figure 2. Dose-response curves of GLP-1 and selected
`compounds using the cloned human GLP-1 receptor expressed
`in BHK cells: 1 (GLP-1) ((cid:127)), 11 ((cid:143)) , 13 (e), 27 (0). Data are
`from one representative experiment with triplicate samples
`and are shown as mean ± SD .
`
`Table 2. Plasma Half-Lives in Pigs ofGLP -1 and Selected
`Potent Acylated Compounds 0
`
`compd
`1 (GLP -1)
`4
`5
`7
`8
`
`plasma t112 (h)
`1.2
`20 ± 2
`14 ± 2
`13
`12 ± 1
`
`compd
`18
`20
`21
`27
`35
`
`plasma t112 (h)
`15 ± 3
`31 ± 4
`8.8 ± 1
`13 ± 4
`11 ± 1
`
`"The half-lives were calcu lated 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.
`
`10000·
`
`2 1000
`s
`
`C
`0
`
`Q)
`
`~ 100
`c
`~
`0 u
`
`10
`
`0
`
`10
`
`20
`
`40
`
`50
`
`60
`
`30
`Time (hours)
`Figure 3. Pharmacokinetic profile of selected compounds
`after sc administration to pigs: 1 (GLP-1) ((cid:127)), 5 (0), 7 (e), 8
`((cid:143))
`. Two pigs were dosed with compounds, and the follo,ving
`immunoassay was performed in duplicate. Data are expressed
`as mean± SD .
`
`not shown). Figure 3 illustrates the dramatic difference
`in plasma hal.f~lives between GLP-1 and three potent
`acylated compounds (5 , 7, 8).
`Site of Acylation. Many different positions in the
`C-terminal part ofGLP-1 could be derivatized with quite
`long fatty acids (3-9, 30-121 pM) without a ffecting the
`potency . Binding affinity was not measured for these
`compounds as they all bind to albumin as part of their
`mechanism of protract ion 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(cid:173)
`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 ofthe receptor, whereas only positions 28 and
`
`29 were important in the C-terrninus. 27 Even though
`only position 8 in the N-terminus ,vas acylated and this
`compound found has a poor potency , the present results
`nicely support those earlier reported from this group .27
`Number of Fatty Acid Substituents. 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 . Jn order not to have to use analogues of GLP-
`1, derivatization with two fatty acids was examined
`(10-13). Only 13 with a Cl2 fatty acid had an accept(cid:173)
`able potency (l 77 pM). Derivatization with CI 4 or CI 6
`fatty acids 0 either as diacids or as monoacids with a
`y-G lu spacer, result ed in com pounds with a marked loss
`of activity (10-12).
`Compounds Derivatized on Lysine 26. The poten(cid:173)
`cies of the compounds were comparable when looking
`at a series of differ ent length diacids (14, 15) or fatty
`acids with the same spacer (5, 16-18). Within the y-Glu
`spacer monoacid series (5, 16-18), derivatization with
`a Cl8 acid (16, 194 pM) led to a significant loss of
`activity compared to Cl6 (5, 61 pM), C 14 (17, 22 pM),
`and Cl2 (18 , 27 pM) . Within the diacid series (14, 15),
`the diacid could be no longer than a Cl4 (15, 72 pM)
`before a loss in pot ency (14 , 154 pM), compared to the
`y-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- l in order to make the
`molecule more resistant to enzymatic breakdown. 21,28 -29
`Desaminohistidine 7 represents one of the more potent
`su.ggestions to a modification giving metabolic stabil(cid:173)
`ity.29 Nevertheless, as seen when comparing 19 (687
`pM) to 5 (61 pM), considerably rn ore potent compounds
`could be obtained by not moditying the N-tenn inus
`when a combination with acylation was desired . This
`could of course be caused both by the position ofthe 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 C 16 fatty acid .
`Compounds Derivatized with Different Spacers.
`The y-Glu spacer is optically active. Thu.s , it presents a
`greater analytical challenge when upscaling the com(cid:173)
`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 y-Glu spacer (5, 61 pM). A
`µ- a lanine spacer reduced the activity slightly (21, 113
`pM) and a piperidyl-4-ca rboxy spacer resulted in a 6-fold
`lower activity (22, 410 pM).
`Compounds Derivatized on Lysine 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 y-Glu spacer and a
`Cl6 fatty acid. This difference ca n of course also be
`explained by the respective a r ginine substitutions .
`Acylating with a y-Giu spacer and a Cl6 fatty acid when
`combined ,vith the above-discussed N-term ina 1 modifi(cid:173)
`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
`
`
`
`Novo Nordisk A/S Ex. 2011, P. 3
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`
`
`Potent Derivatives ofGLP-1
`
`Journal ofAfedicinal Chemistry, 2000, Vol. 43, No. 9 1667
`
`pM vs 25, 169 pM). However, the C8 fatty acid wiil not
`lead to much protraction (data not shown).
`Compounds Derivatized on Lysine 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 y-Glu(cid:173)
`spaced monoacids ,vere used , the more loss of activity
`was observed . Going from a C16 diacid to a C20 shifts
`the potency from 7. 9 to 2 l O pl'vl (27 vs 26). AC 16 y-Glu(cid:173)
`spa ced monoacid and the corresponding Cl8 decreases
`potency from 36 to I 16 pM (8 vs 28). Acylation of lysine
`36 with a y-Glu-spaced Cl6 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).
`Compounds Derivatized on Lysine 38. The com(cid:173)
`pounds derivatized on lysine 38 (9, 29-36) were also
`very potent. In this series vve
`tried some further
`modifications. Substitution of glycine instead of alanine
`at position 8 gives stability toward enzymatic break(cid:173)
`down.2i However, as expected from earlier studies,2i 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 alburnin. 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 ,vas lost (35, 135 pM).
`
`Conclusion
`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.
`
`J<:xperimen ta! Section
`Instrumentation. Analytical HPLC analysis of acylated
`GLP-1 analogues was performed using UV detection at 214
`nm and a Vydac 214TP54 4.6- x 250-mm, 5-µm C-4 silica
`column at 42 ''C. The column was eluted at l mL/min with a
`gradien t of0---100% acetonitrile, O.OH{, TF A against 0.1 °/4, TF A
`in water.
`Mass spectral data were obtained using a Voyager RP
`MALDI-TOF (matrix assisted laser desorption ionization time(cid:173)
`oJ~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 ,vas used
`as matrix and calibration was performed using external
`standards with a resulting mass accuracy of0.1%.
`Synthesis of GLP-l .Analogue Starting Peptides. Solid(cid:173)
`phase methodology was used to synthesize these analogues.
`The peptides were synthesized according to the Fmoc strategy
`on an App lied Biosystems 431A peptide synthesizer in 0.25-
`mmol scale using the manufacturer supplied FastMoc UV
`protocols starting with either a Fmoc-Gly-\Vang resin, Fmoc(cid:173)
`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 n eeded, were : Fmoc-Arg(Pmc), Fnrnc(cid:173)
`Trp(Boc), Fmoc-Glu(OBut), Fmoc-Lys(Boc), Fmoc-Gln(T rt ),
`Fmoc-Tyr(But), Fmoc-Ser(But). Fmoc-Thr(But), Fmoc-His(Trt),
`Adoc-imidazolylpropionic acid and F moc-Asp(OBu t).
`
`The peptides were cleaved from the resin and side chain
`deprotected in TF Aiphenol/thioanisole/water /ethanedith iol
`(83.25 :6.25:4.25:4 .25:2.0) for 180 min. The cleavage mixtures
`were filtered and t he 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 semi preparative HPLC (Waters, Mi11ipore ) on a 25- x 250-
`mm column packed with 7-pm C-18 silica. The column was
`eluted with a gradient of acetonitrile against 0.05 M (NH 4 )r
`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 toa Sep-Pak Cl8 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.l %
`TFA in water and isolated from the eluate by lyophilization
`atler dilution with water. The final product obtained was
`characterized by a mino acid analysis, analytical RP-HP LC and
`MALDI-N[S . Amino acid analysis and mass spectrometry data
`agreed ,vith the expected structure within the experimental
`error of the methods (mass spectrometry ± 3 mass units,
`am ino acid analysis± 10%). RP-HPLC showed a peptide purity
`>95%1.
`The RP-HP LC analyses were performed using UV detection
`at 214 nm and a Vydac 218TP54 4 .6- x 250-rnm, 5-/tm C-18
`silica column which was eluted at l mL/min at 42 cc_ Two
`different elution conditions were used: (A) a gradient of 5-60'}0
`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 -l(7-37)
`R34-GLP-1(7-37)
`K3 6R 26 ,34-GLP-l (7---36)
`R 26 •34-GLP-1(7-38)
`des am in o-H 7R 26-GLP-l (7----3 7)
`des am ino-H7 R 34-GLP-l (7-3 7)
`G 8R 26 34 -GLP -1 (7-38)
`
`E37GsR26,J•1_GLP-l (7-3 8)
`K27R 26,34_GLP-1(7-37)
`K23R26,34_G LP-1(7---37)
`K 18R26 •3 -1-GLP-1(7-37)
`K 8R 26 ,34-GLP-1(7---3 7)
`E 37R 26,3"-GLP-1 (7-38)
`
`Synthesis of Intermediates. Alkanedioic Acid ]\fono(cid:173)
`succinimidyl Ester. The alkanedioic acid monoesters were
`prepared from the corresponding alkanedioic acids according
`to the procedure described in the literature. 30
`N'-'-AJkanoyl-L-glutamic Acid o.-tert-Butyl Ester y-Suc(cid:173)
`cinimidyl Ester. The N'"-alkanoyl-L-glutamic acid a-tert-butyl
`ester y-succinimidy l esters were prepared form L-glutamic acid
`a-tert-bu tyl ester and the corresponding alkanoic acids accord(cid:173)
`ing to the procedure described in the literature .31
`y-(.iVY-Hexadecanoyiamino)hutyric Acid Suxcinimidyl
`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 11111101) in DMF (200 mL)was
`stirred at 25 c,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 rnL). 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 a t 25 c,c for 20 h and then
`filtered. The solvent was evaporated to give a solid residue,
`which after recrystallization from a mixture ofn-heptane (50
`mL) and 2-propanol (2 .5 m L ) gave the title compound (2 .5 g,
`7 5'1/i, ).
`Synthesis of Acylated GLP-1 Analogues 2--36. All acyl(cid:173)
`modified peptides, except 20 and 24, were synthesized accord(cid:173)
`ing to general Methods A and B . Peptide sequence, site of
`acylation, acyl substitucnt, yields, and physical data are given
`in Tables land 3.
`Method A. A mixture of starting peptide (1 equiv), DIP EA
`(28 equiv), NMP (70 ftll,umol starting peptide) and water (66
`,uUµrnol starting peptide) was gently shaken for 5 min at 25
`°C. To the resulting mixture was added a solution of al(cid:173)
`kanedioic acid monosuccinimidyl ester in NMP (22 /iL/f-lmol
`
`
`
`Novo Nordisk A/S Ex. 2011, P. 4
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`
`
`1668
`
`.Journal ofA1edicinal Chemistry, 2000, Vol. 43. No. 9
`
`Knudsen et al.
`
`Table 3. List of A.ll Compounds with Method of Synthesis,
`Purification Yi eld, and Estimated Purity as Determined by
`Analytical HPLC and Measured and Theoretical Molecular
`Vleight
`
`yie ld
`
`M'W
`
`com pd method
`
`o/0
`
`m.g HPLC purity(%)
`
`calcd
`
`found
`
`1
`2
`3
`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
`
`B
`B
`B
`B
`B
`B
`B
`B
`A
`B
`B
`A
`A
`A
`B
`B
`B
`B
`a
`A
`A
`B
`a
`B
`A
`A
`B
`A
`A
`B
`B
`B
`B
`B
`B
`
`28
`15
`24
`6
`16
`30
`23
`47
`14
`20
`9
`9
`12
`43
`34
`29
`36
`41
`35
`55
`44
`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
`~..,
`/ , h
`4.8
`2.4
`27.3
`1.0
`0.6
`11.5
`15.5
`6.1
`8.2
`9.1
`15 .8
`11.6
`12
`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
`
`>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
`
`0 Specific method described below.
`
`starting peptide); the reaction mixture was gently shaken for
`5 min at 25 cc and then allowed to stand at 25 c•c for an
`add itional 2.5 h . The reaction was quenched hy the addition
`ofa solution of glycine (22 equiv) in 50% aqueous ethanol (l.65
`µU11mol starting peptide). The react ion mixture was purified
`by preparative chromatography using a cyanopropyl column
`(Zorbax 300S B-CN) and a standard acetonitrile/TF A system .
`The column was heated to 65 °C and the acetonitrile gradient
`was 0-100°1,, in 60 min . Fractions containing the product were
`isolated and lyophilized to give the final product. Character(cid:173)
`ization was performed by HPLC and mass spectral analysis
`(Table 3 ).
`Method El. Am ixture of starting peptide (I equiv), DIP EA
`(28 equiv), NMP (245 ,ul/µmol starting peptide) and water (245
`/tL!,umol starting peptide) was gently shaken at 25 °C for 5
`min . To the resulting mixture was added a solution of
`N"-alkanoyl-L-glutam ic acid a.-tert-bu.tyl ester y-su.ccinimidyl
`ester (5.9 equiv), prepared as described above , in NMP (75 µLI
`µmo! starting peptide); the reaction mixture was gently shaken
`at 25 cc ti)r 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 glycin e (22 equiv) in 50% aqueous ethanol (165
`,uL/,umol starting pept ide). A 0.5% aqueous solution of am(cid:173)
`monium acetate (12 .5 mL/,umol starting peptide) was added,
`and the resulting mixture was eluted onto a preequilibrated
`Varian C8 Mega Bond Elut cartridge. The immobilized com(cid:173)
`pound ,vas washed with 5% aque ous acetonitrile (3 .75 mL/
`,um ol starting peptide) and finally .liberated from the cartridge
`by elution with TFA (1 .5 mL~umol starting peptide). The eluate
`was allowed to stand at 25 °C for 1.75 hand then evaporated.
`The residue was purified by preparative chromatography using
`a cyanopropyl column (Zorbax 300SB-CN) and a standard
`acetonitrile/TF A system . The column was heated to 65 cc and
`the acetonitrile gradient was 0-100% in 60 min . Fractions
`
`containing t he product were isolated, and lyophilized to give
`the final product. Characterization was performed by HPLC
`and mass spectral analysis (Table 3).
`R2"K34-(N'-(y-Am in ob u ty roy l(NY-hcx adecan oy l)))-GLP-
`1 (7 -37)-0H, 20. To a mixture ofR 26-GLP-1(7-37)-OH (41.l
`mg, 12.2 ,umol), DIPEA (44 mg, 341 ,umol), NMP (5.76 mL)
`and water (2 .88 mL) was a dded a solution of y-(NY-hexade(cid:173)
`canoyl)aminobutyric acid succinimidyl ester (16 mg, 36 .5
`µmo!), prepared as described above, in NMP (400 ,uL). 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 ofa solution of glycine
`(20 mg, 268 ,umol) in water (200 pL). The solvent ,vas
`evaporated and the residue purified by preparative chroma(cid:173)
`tography using a cyanopropyl column (Zorbax 300SB-CN) and
`a standard acetonitri!e1TFA system. The column was heated
`to 65 °C and the acetonitrile gradient was 0-100% in 60 min .
`Fractions containing the product 'Nere isolated and lyophilized
`to provide 20 (15 .8 mg, 35%). Characterization was performed
`by HPLC and mass spectral analysis (Table 3).
`Des am ino-IFR26K3';-(N'-octanoy l)-GLP-1 (7-37)-0H, 24.
`To a mixture of desamino-H 7R 26-GLP-1(7-37)-OH (19.8 mg.
`5 .89 /tmol), DlPEA (21.2 mg, 164 µmol), NMP (1.38 mL) and
`water (1.38 mL) was added a solution of octanoic acid succin(cid:173)
`imidyl ester (20.7 mg, 36.5 ,umol), prepared as described in
`the literature, 33 in Nl'v1P (106 ,uL). The reaction mixtu re was
`gently shaken for 5 min at 25 ,-;c and th en allowed to stand
`for an additional 1.5 hat 25 'T. The reaction was quenched
`by the addition of a solution of glycine (9 .7 mg, 129 ,umol) in
`water (97 /tL. The solvent was evaporated and the residue
`purified by preparative chromatography using a cyanopropyl
`column (Zorbax 300SB-CN) and a standard acetonitrile/TF A
`system. The column was heated to 65 cc and the acetonitrile
`gradient was 0-100% in 60 min . Fractions containing the
`product were isolated and lyophilized to provide 24 (9 .2 mg,
`45'1/i,). Characterization was performed by HPLC and mass
`spectral analysis (Table 3 ).
`Receptor Experiments. Baby hamster kidney (BHK) cells
`expressing the doned human GLP-1 receptor (BHK 467 -12A)
`were grown in DMEM media with the addition of 100 IU /mL
`penicillin,100 ftL /mL streptomycin, lO'h1 fetal calf serum and
`l mg/mL Geneticin G-418 (Life Technologies). Plasma mem (cid:173)
`branes were prepared by homogenization in buffer (10 mM
`Tris-HCl, 30 ml'vl NaCl and 1 mM dithiothreitol, pH 7.4,
`containing, in addition , 5 mg/L leupeptin (Sigma), 5 mg iL
`pepstatin (Sigma), 100 mg1T, bacitracin (Sigma), and 16 mg!L
`aprotinin (Calbiochem-Novabiochem, La Jolla, CA)). The ho(cid:173)
`mogenate was centrifuged on top ofa layer of41 % 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 (cid:173)
`tions: 50 m M Tris-HCl, 1 mM EGTA, 1.5 mM MgSO 4 , 1.7 mM
`ATP , 20 mM GTP, 2 mM 3-isobutyl-1-methylxanthine (IBMX),
`0 .0 1% w/v Tween-20 , pH 7.4 . Compounds were dissolved and
`diluted in buffer. GIP was freshly prepared f<.)f each experi(cid:173)
`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(cid:173)
`pounds and EC 50 val ues calculated using GraphPad Prism
`software .
`Pharmacokinetic Experiments. The experiments were
`performed in Landrace x Du.roe :< Yorkshire pigs weig hing
`between 25 and 50 kg. All the GLP-1 analogues wer