Macrocyclic prodrugs of a selective nonpeptidic direct thrombin inhibitor display high permeability, efficient bioconversion but low bioavailability

Article abstract of DOI:10.1021/acs.jmedchem.5b01871

The only oral direct thrombin inhibitors that have reached the market, ximelagatran and dabigatran etexilat, are double prodrugs with low bioavailability in humans. We have evaluated an alternative strategy: the preparation of a nonpeptidic, polar direct

Full text of DOI:10.1021/acs.jmedchem.5b01871

Subscriber access provided by - Access paid by the | UCSB Libraries
Article
Macrocyclic Prodrugs of a Selective Non-peptidic Direct Thrombin Inhibitor
Display High Permeability, Efficient Bioconversion but Low Bioavailability
Vincent Andersson, Fredrik Bergström, Jonas Brånalt, Gunnar Grönberg, David
Gustafsson, Staffan Karlsson, Magnus Polla, Joakim Bergman, and Jan Kihlberg
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01871 • Publication Date (Web): 27 Jun 2016
Downloaded from http://pubs.acs.org on June 28, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Macrocyclic Prodrugs of a Selective Non-peptidic Direct Thrombin
Inhibitor Display High Permeability, Efficient Bioconversion but Low
Bioavailability
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
1
3
Vincent Andersson, Fredrik Bergström, Jonas Brånalt, Gunnar Grönberg, David
4
1
1
1,
5,
Gustafsson, Staffan Karlsson, Magnus Polla, Joakim Bergman * and Jan Kihlberg *
1
2
3
4
5
CVMD iMed, AstraZeneca R&D Mölndal, SEꢀ431 83 Mölndal, Sweden.
CVMD iMed DMPK, AstraZeneca R&D Mölndal, SEꢀ431 83 Mölndal, Sweden.
RIA iMed, AstraZeneca R&D Mölndal, SEꢀ431 83 Mölndal, Sweden.
Emeriti Pharma AB, AZ Bioventure Hub, c/o AstraZeneca, Sꢀ431 83 Mölndal, Sweden.
Department of Chemistry ꢀ BMC, Uppsala University, Box 576, SEꢀ751 23 Uppsala,
Sweden.
Abstract
The only oral direct thrombin inhibitors that have reached the market, ximelagatran and
dabigatran etexilat, are double prodrugs with low bioavailability in humans. We have
evaluated an alternative strategy ꢀ the preparation of a nonꢀpeptidic, polar direct thrombin
inhibitor as a single, macrocyclic esteraseꢀcleavable (acyloxy)alkoxy prodrug. Two
homologous prodrugs were synthesized and displayed high solubilities and Cacoꢀ2 cell
permeabilities, suggesting high absorption from the intestine. In addition, they were rapidly
and completely converted to the active zwitterionic thrombin inhibitor in human hepatocytes.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Unexpectedly, the most promising prodrug displayed only moderately higher oral
bioavailability in rat than the polar direct thrombin inhibitor, most likely due to rapid
metabolism in the intestine or the intestinal wall. To the best of our knowledge, this is the
first in vivo ADME study of macrocyclic (acyloxy)alkoxy prodrugs, and it remains to be
established if the modest increase in bioavailability is a general feature of this category of
prodrugs or not.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Introduction
Plasma coagulation, activated during vascular damage, aims at reducing blood loss. It is a
cascade activation of trypsinꢀlike serine proteases with delicate control due to feedꢀback
loops, coꢀfactors and inhibitors. In the final step of the cascade thrombin (coagulation factor
IIa) cleaves fibrinogen to fibrin, which stabilizes the clot during the wound healing process.
In addition, platelet aggregation helps plasma coagulation and fibrinolysis counteracts it.
Plasma coagulation can also be activated in diseases like venous thrombosis, pulmonary
embolism, stroke and myocardial infarction and oral anticoagulants are an important group of
pharmaceuticals used in the prevention and treatment of these diseases. For many years
vitamin K antagonists such as warfarin were the only oral anticoagulants, but recently direct
1
, 2
thrombin inhibitors (DTIs) and direct coagulation factor Xa inhibitors have been approved.
3
4
The two oral DTIs that have reached the market, ximelagatran and dabigatran etexilate,
both have hydrophilic groups at the ends of the molecules; a benzamidine in the P1 position
and a carboxylic acid attached to the P3 group. The result of this zwitterionic character is
poor absorption from the gastrointestinal tract. Both DTIs are therefore prepared as doubleꢀ
prodrugs with a hydroxyl group or an Oꢀnꢀhexylcarbamate on the amidine, while the acid is
protected as an ethyl ester. Still, oral bioavailability in humans is in the range of 20 and 6 %
ACS Paragon Plus Environment

Page 3 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
for ximelagatran and dabigatran etexilate, respectively. Moreover, doubleꢀprodrugs are
complicated to develop since it often happens that one of the prodrug arms is metabolized,
3
but not both. Another liability common to ximelagatran and dabigatran etexilate is their low
5
, 6
selectivity against trypsin, which may result in GIꢀside effects.
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In the search for novel DTIs with improved properties we developed a series of potent
inhibitors based on a P2 dihydropyridꢀ2ꢀone scaffold, designed from the unique 5,6ꢀ
dihydroxyoctahydroindoleꢀ2ꢀcarboxylic acid core present in naturally occurring thrombin
7
inhibitors like chlorodysinosin. The dihydropyridꢀ2ꢀones were designed to bind in the P1ꢀP3
pockets of thrombin and had low nanoꢀmolar potency against thrombin and high selectivity
against trypsin. However, highly potent inhibitors of human thrombin such as 1 (Figure 1)
had liabilities such as high metabolism in rat liver microsomes and some inhibition of CYP
8
3
A4 (cf. Table 1). We reasoned that introduction of a carboxylic acid in the meta position of
the P3 moiety, which points towards solution in models of 1 and thrombin, would decrease
lipophilicity and thus decrease both metabolism and CYP 3A4 inhibition, but zwitterions
such as 2 require conversion into prodrugs to improve bioavailability. To avoid complications
with doubleꢀprodrugs we decided to investigate if a macrocyclic prodrug that was
transformed to an active DTI in only one enzymatic reaction could be an alternative and
9
advantageous approach. The esteraseꢀsensitive (acyloxy)alkoxy linker has been developed
for transformation of linear peptides into cyclic prodrugs, that have shown promising ADME
10ꢀ15
properties in vitro, and was therefore chosen for use in our study (cf. 3 and 4, Figure 1).
We were also interested in investigating the physicochemical properties and cell permeability
of the macrocyclic prodrugs as compared to a linear analogue. Herein we describe the
synthesis of zwitterionic direct thrombin inhibitor 2, the corresponding macrocyclic prodrugs
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3 and 4, and acyclic analogue 5 (Figure 1). We also report the potency, serine protease
selectivity as well as in vitro and in vivo ADME properties of these compounds.
R
CO Me
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
NH2
NHCO Me
NH
N
2
P3
NH
N
O
O
HN
HN
O
P2
O
P1
Cl
Cl
5
1
2
: R = H
: R = CO H
2
O
CO R
2
R
O
O
I
O
O
R
O
O
NH
NHBoc
O
HN
N
HN
N
O
O
+
HN
HN
NO2
O
Cl
8
: R = H
Cl
9: R = Me
3
4
: R = H
: R = Me
6
7
R = H (Key P1-3 intermediate)
R = Me
TMS
NH2
NHBoc
P1
OHC
CO Me
N
O
2
O
+
+
H N
2
P3
O
P2
1
0
12
Cl
1
1
Figure 1. Structures of direct thrombin inhibitors 1 and 2, and a retrosynthetic analysis of 2,
macrocyclic prodrugs 3 and 4, and acyclic analogue 5.
Results and Discussion
ACS Paragon Plus Environment

Page 5 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Retrosynthesis. We envisaged that macrocyclic prodrugs 3 and 4 could be synthesized from
key P1ꢀ3 intermediate 6 and (acyloxy)alkoxy synthetic equivalents 8 and 9 through a oneꢀpot
cyclisation procedure (Figure 1). Thus, the carboxylic acid of 6 would first be alkylated with
8
and 9 while having the P3ꢀamine protected with a Boc group. Cyclization would then be
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
achieved by deprotection of the amine, basification and an intramolecular attack to form
macrocycles 3 and 4. Direct thrombin inhibitor 2 would be obtained by deprotection of 6,
while the acyclic analogue 5 of the two prodrugs would be obtained by exchange of the
protective group on the P3ꢀamine of 7. Key intermediate 6 could, in turn, be prepared from
P1ꢀ3 fragments 10ꢀ12.
Synthetic chemistry. P3 fragment 10 is commercially available, whereas P1 fragment 12
16
was synthesized as previously reported. P2 fragment 11 was prepared via a route similar to
7
that developed for the corresponding ethyl ester. This started from NꢀBocꢀprotected
piperidone 13, which was prepared using a threeꢀstep process starting from 4ꢀ
17
methylpiperidine (Scheme 1). Alkylation of 13 with 14 afforded TMS ester 15, in 85 %
yield. The dihydropyridꢀ2ꢀone functionality was obtained by a Sharpless selenoxide
18
elimination involving stepꢀwise αꢀaddition with PhSeBr followed by an oxidative
elimination with hydrogen peroxide. Initial attempts gave 16 with a poor regioisomeric
distribution (1:4 endo:exo) and consequently the yield of desired product 16 was low.
However, the unwanted exoꢀcyclic product could be rearranged with DBU in toluene at reflux
to give 16 in 45 % yield, over three steps. Compound 16 was subsequently NꢀBoc
deprotected with TFA and then reacted with tertꢀbutyl nitrite to form the Nꢀnitroso compound
17 in 89 % yield. The Nꢀnitroso group was then reduced with Zn to give the hydrazidic P2
fragment 11 in 65 % yield.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Scheme 1. Synthesis of P2 fragment 11
TMS
O
TMS
Boc
N
Br
Boc
N
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
O
O
O
1
4
a
O
13
15
b, c
TMS
TMS
NO
N
Boc
N
O
O
O
O
e-f
O
O
17
16
d
g
+
TMS
Boc
N
TMS
NH2
N
O
O
O
O
P2
O
O
exo-16
11
a
Reagents and conditions: (a) 14, LHMDS, THF, ꢀ78 to ꢀ20 °C, 90 min, 85%; (b) PhSeBr,
LHMDS, THF, ꢀ78 to ꢀ20 °C, 210 min. (c) H O in excess, CH Cl , 10 °C, 30 min; (d) DBU,
2
2
2
2
toluene, reflux, 12 h, 45% over three steps; (e) TFA, CH
pyridine, Et O, reflux, 24 h, 89% over two steps; (g) Zn nanopowder, acetic acid, methanol, 0
to 25 °C, 65%.
2 2
Cl , rt, 8 h. (f) tertꢀBuONO,
2
An attempted, oneꢀpot reductive amination of P2 fragment 11 with the aldehyde in P3
fragment 10 did not provide acceptable yields. Instead, imine 18 was first isolated and then
3
sequentially reduced with NaBH CN, which gave 19 in 90 % yield, based on recovered
ACS Paragon Plus Environment

Page 7 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
starting material (brsm, Scheme 2). Next, selective cleavage of the TMS ester in 19 with
TBAF gave 20 in 93 % yield (brsm). Formation of the amide bond between carboxylic acid
20 and monoꢀprotected diꢀamine 12 (P1 fragment) was performed using HATU as coupling
reagent in the presence of DIPEA to give 7 in 72 % yield. Finally, a convenient ester
hydrolysis with LiOH in 1,4ꢀdioxane and water for 8 h provided key P1ꢀ3 intermediate 6 in
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
95 % yield.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 2. Synthesis of key P1ꢀ3 intermediate 6, direct thrombin inhibitor 2 and acyclic
a
prodrug analogue 5 from P1ꢀP3 fragments 10ꢀ12
CO Me
CO Me
2
2
OHC
CO Me
2
TMS
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
NH2
N
10 (P3)
O
O
TMS
TMS
N
N
NH
N
a
O
O
b
O
O
O
O
O
19
1
1 (P2)
18
c
CO H
CO Me
2
2
NHBoc
CO Me
2
NH2
BocHN
BocHN
12
(P1)
NH
NH
N
NH
N
O
O
Cl
N
O
HN
HN
e
d
CO H
2
O
O
Cl
Cl
7
2
0
6
(Key P1-3 intermediate)
f-g
h
CO H
CO Me
2
2
NH2
NHCO Me
2
NH
N
NH
N
O
O
HN
HN
O
O
2
5
Cl
Cl
a
Reagents and conditions: (a) methyl 3ꢀformylbenzoate (P3ꢀfragment), acetic acid, methanol,
rt, 12 h, 81%; (b) NaBH CN, acetic acid, methanol, rt, 6 h, 90% based on recovered starting
material (brsm), 66% conversion of 18; (c) TBAF, THF, rt, 6 h, 93% brsm, 92 % conversion
of 19; (d) HATU, DIPEA, 12, NMP/CH Cl , 40 °C, 6 h, 72%; (e) LiOH, methanol, 1,4ꢀ
3
2
2
ACS Paragon Plus Environment

Page 9 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dioxane/water, rt, 8 h, 95%. (f) TFA, CH
2
Cl
2
, rt, 6 h; (g) methyl carbonochloridate, CH
2 2
Cl , 0
°C, 20 min, 71% over two steps. (h) TFA, CH
2
Cl , rt, 40 min, 97%.
2
Acyclic prodrug analogue 5 was synthesized from 7 via an NꢀBoc deprotection with TFA
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
followed by acylation of the liberated benzylic amino group with methyl carbonochloridate at
low temperature to provide 5 in 71 % yield, over two steps. Compound 2 was prepared as a
trifluoroacetate in almost quantitative yield by treatment of 6 with TFA.
Synthesis of (acyloxy)alkoxy linkers 8 and 9 was performed from commercially available
11
chloroformates 21 and 22 (Scheme 3). Reaction with paraꢀnitrophenol in the presence of 4ꢀ
methylmorpholine as base in CH Cl gave 23 and 24 in 90 and 81 % yields, respectively. A
2
2
Finkelstein halide exchange reaction with NaI in dry acetone at reflux then provided iodoꢀ
compounds 8 and 9 in 85 and 47 % yields, respectively. A competing E2ꢀelimination of
ethene to release carbon dioxide and reform paraꢀnitrophenol was observed during the
synthesis of 9. For this reason, the reaction to give 9 needed to be carefully monitored and
stopped at 75 % conversion, which led to a lower yield.
a
Scheme 3. Synthesis of (acyloxy)alkoxy linkers 8 and 9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
R
O
R
O
a
Cl
O
Cl
Cl
O
O
2
2
1: R = H
: R = Me
2
3
: R = H
: R = Me
2
24
NO2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
b
R
O
I
O
O
8
9
: R = H
: R = Me
NO2
a
Reagents and conditions: (a) pꢀnitrophenol, 4ꢀmethylmorpholine, CH
2 2
Cl , 0 to 25 °C, 12 h,
2
3: 90% and 24: 81%; (b) NaI, acetone, 50 °C, 24 to 48 h, 8: 85% and 9: 47%.
Finally, formation of the macrocycles was done in a fourꢀstep procedure, the first step of
which involved generation of a cesium carboxylate anion by mixing 6 and cesium carbonate
(Scheme 4). Alkylation of the cesium carboxylate using one equivalent of either of 8 and 9 at
0
°C in DMF for 4 h gave approximately 49 and 25 % of alkylated intermediates 25 and 26,
respectively, according to LCꢀMS. Steric effects are a plausible explanation for the lower
degree of reaction in the case of iodoꢀethyl linker 9. It has been reported that this type of
activation with (acyloxy)alkoxy substrates (R = H) gives yields around 30% when using
11
+
+
stoichiometric amounts of alkylating agents. Other cations such as Ag and Hg have been
reported to mediate substitutions with similar alkylating agent but with less functionalized
19
acids in good yields. Therefore attempts were performed using silver carbonate and 6, but
they showed little or no reaction in comparison with cesium carbonate. Dimethyl sulfoxide
and toluene that have been used as solvents in the literature were also tested, unfortunately
without any improvements. Alkylated intermediates 25 and 26 were then subjected to in situ
ACS Paragon Plus Environment

Page 11 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2 2
NꢀBoc deprotection with TFA in CH Cl to give free amines as trifluoroacetates and the
solution was subsequently diluted to 0.7 mM and basified with aqueous potassium carbonate
after which it was left to macrocyclize for 72 h. An intramolecular attack of the primary
amine on the acyloxy carbonyl group then leads to formation of macrocyclic prodrugs 3 and
4. The overall isolated yields of 3 and 4 from 6 was 16 and 8 %, over 4 steps respectively,
which connoted that the final macrocyclization proceeded with an effective yield of 32ꢀ33 %.
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Scheme 4. Synthesis of macrocyclic prodrugs 3 and 4
NO2
NO2
R
O
O
R
O
CO H
2
I
O
O
O
O
O
O
8: R = H
BocHN
9: R = Me
NH
NHBoc
HN
N
N
O
HN
O
a, b
HN
O
Cl
Cl
6
2
2
5
6
: R = H
: R = Me
c, d
O
R
O
O
O
O
NH
3
4
: R = H
: R = Me
HN
N
O
HN
Cl
a
Reagents and conditions: (a) Cs
2 3
CO , methanol, rt, 1 h; (b) 8 or 9, DMF, 0 °C, 4 h. 25: 49%
and 26: 25% over two steps based on LCMS analyses; (c) TFA, CH Cl , rt, 6 h; (d) CH Cl ,
2
2
2
2
K
2
3
CO (aq), rt, 72 h, 3: 33% and 4: 32% over two steps. Overall yields of 16% and 8% for 3
and 4, respectively, over four steps from 6.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Structure verification. Highꢀresolution mass spectrometry (HRꢀMS) showed a single
molecular ion having the expected mass for each of 3 and 4 within the standard deviation, ±5
ppm. In theory, macrocyclization of activated intermediates 25 and 26 could have occurred at
another position, for instance at the amide bond, and use of only mass spectrometry for
structural confirmation was therefore not judged to be sufficient. Structural verification of
macrocycles 3 and 4 by NMR spectroscopy was not straightforward since each of them
generated two sets of signals at room temperature, suggesting slow interconversion of
conformational isomers. Structural confirmation was, however, achieved through a complete
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
13
assignment of the NMR spectra of 3 and 4 based on H, C, COSY, NOESY, HSQC, HMBC
and NꢀHMBC experiments (cf. Supporting information). The HMBC spectrum of 4 provided
the key information, i.e. the couplings between carbonyl carbon C21 and linker proton H34,
between H34 and carbonyl carbon C36, and between C36 and proton H30 (Figure 2). These
couplings verified that the macrocyclization had occurred at the correct position and that the
cyclic structure was intact after purification. The structure of macrocycle 3 was elucidated
and confirmed in the same manner.
O
21
O
O
3
6
Me 34
O
4
NH
HN
N
30
O
HN
O
Cl
Figure 2. Arrows indicate the key couplings in the HMBC spectrum that confirmed the
structure of 4.
ACS Paragon Plus Environment

Page 13 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Potency, in vitro metabolism and selectivity towards serine proteases. As mentioned in
the introduction direct thrombin inhibitor 1 had liabilities that included high metabolism in
rat liver microsomes and inhibition of CYP3A4 (Table 1). Introduction of a carboxylic acid
in the meta position 1, to give compound 2, reduced both of these liabilities. Satisfactorily,
this was accomplished with only a minor loss in functional in vitro anticoagulant potency in
human plasma, as measured by the concentration required to double the activated partial
thromboplastin time potency (APTT), and the in vivo efficacy in prevention of venous
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
0
thrombosis in the rat caval vein thrombosis model (VT). Accordingly, compound 2 was
somewhat less potent than 1 as an inhibitor of human thrombin in vitro (Table 2). The
selectivity of 1 and 2 towards other serine proteases was also examined, and both compounds
displayed a >1000ꢀfold selectivity window against trypsin I, Factors VIIa, IXa, Xa and XIa as
well as against plasmin. Prodrug 3 was a weak inhibitor of thrombin and did not show any
detectable inhibition of trypsin I. Direct thrombin inhibitor 2 thus displayed high potency
both in vitro and in vivo, excellent selectivity against a panel of serine proteases as well as
satisfactory metabolic properties. The ADMET properties of the corresponding prodrugs 3
and 4 were therefore investigated further.
Table 1. Potency and in vitro metabolism of direct thrombin inhibitors 1 and 2
Compound
RLM Clint
ꢀL/min/mg)
150
CYP3A4 IC50
APTT IC50
VT IC50
a
b
c
d
(
(ꢀM)
(ꢀM, rat)
(nM)
1
2
13
0.68
1.2
72
97
<15
>50
a
b
Metabolic stability in rat liver microsomes. Inhibition of human CYP3A4 in vitro.
c
d
Activated partial thromboplastin time. Venous thrombosis model in the rat.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 2. Selectivity of direct thrombin inhibitors 1 and 2, and macrocyclic prodrug 3 towards
serine proteases
Enzyme
Compound 1
(nM)
SD, n)
.5
0.15, 2)
Compound 2
(nM)
Compound 3
(nM)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
K
i
K
i
K
i
(
(SD, n)
4.6
0
200
Thrombin
(
(2.3, 3)
>5.4 x10
a
3
3
3
Trypsin I
>5.4 x10
>5.4 x10
3
3
FVIIa
FIXa
>33 x10
>29 x10
>17 x10
>19 x10
>19 x10
>33 x10
>29 x10
>17 x10
>19 x10
>19 x10
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
3
3
3
3
3
3
3
3
FXa
FXIa
Plasmin
Based on two repeats.
a
Physicochemical properties and in vitro ADME. To investigate if prodrugs 3 and 4 had
suitable in vitro ADME properties for further development their solubility, lipophilicity, cell
permeability and in vitro metabolism was investigated. Data was also acquired for acyclic
analogue 5 and direct thrombin inhibitor 2. Solubility was determined in phosphate buffer at
pH 7.4 using amorphous material, whereas stock solutions in DMSO were employed for
21
determination of LogD using a modified version of the shakeꢀflask method. Cacoꢀ2 cells
were used for determination of cell monolayer permeability with a pH gradient from 6.5 on
the apical side of the cellular monolayer to 7.4 on the basolateral side. This is a common
22
model for absorption of drugs from the small intestine into the bloodstream. The passive
ACS Paragon Plus Environment

Page 15 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
cell monolayer permeability was determined by inclusion of an inhibitor cocktail (quinidine,
sulfasalazine and benzobromarone) that targets the three major efflux transporters in the
intestinal epithelium: Pꢀglycoprotein, (MDR1/Pꢀgp; ABCB1), breast cancer resistance protein
(BCRP; ABCG2) and multidrugꢀresistance associated protein 2 (MRP2, ABCC2). Cacoꢀ2
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
cell permeability was also determined in the absence of the inhibitor cocktail with a pH of 7.4
on both sides of the cellular monolayer to provide insight into the degree of efflux for
compounds 2ꢀ5.
Table 3. Solubility, lipophilicity and cell monolayer permeability for compounds 2ꢀ5
a
b
Compound MW TPSA
Solubility
LogD7.4
Cacoꢀ2 Papp
Efflux
ratio
2
c
(Å )
(ꢀM)
(SD, n)
AB
ꢀ6
(x 10
(SD, n)
cm/s)
0.10
2
3
4
5
470
526
540
542
125
126
126
126
>100
120
340
160
ꢀ0.5
2.4
(0.5, 4)
5.6
(
(
(
(
0.1, 3)
2.3
48
45
0.1, 3)
2.6
(1.5, 4)
4.8
0.1, 3)
2.9
(1.1, 4)
3.3
54
0.1, 3)
(0.4, 4)
a
b
TPSA was calculated by Molinspiration (www.molinspiration.com). Based on three repeats for 3ꢀ5,
c
with SD <0.17 for Log solubility, and one measurement for 2. Based on three repeats for 2ꢀ5, with
SD <0.14 for Log [Papp AB].
a
As expected from its zwitterionic character (pK = 3.94 and 9.27) at physiological pH (7.4)
direct thrombin inhibitor 2 had a low lipophilicity (Table 3). Consequently, the Cacoꢀ2 cell
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
permeability was in the range where the in vivo fraction absorbed is usually too low for oral
ꢀ6
administration (<1 x 10 cm/s). In contrast, cyclic prodrugs 3 and 4, as well as acyclic
analogue 5, had lipophilicities that are optimal for cell permeability and oral delivery (LogD
23
1
ꢀ3). Accordingly, 3ꢀ5 had high solubilities (>100 ꢀM) as well as high passive cell
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ6
permeabilities (appr. 50 x 10 cm/s) that can be expected to translate into close to complete
absorption from the intestine. Compounds 3ꢀ5 did display some transporter mediated efflux
(efflux ratios: 2.4ꢀ5.6) but not to the extent where oral administration can be expected be
severely compromised. In fact, all classes of macocyclic drugs and clinical candidates that are
delivered orally for systemic use have been linked to efflux transporters, either as substrates
24
or inhibitors. Macrocyclization often results in increased cell permeability as compared to
25, 26
nonꢀmacrocyclic controls,
but this was not observed for 3 and 4 as compared to acyclic
analogue 5, most likely because macrocyclization did not result in that polar moieties were
shielded e.g by formation of intramolecular hydrogen bonds.
Prodrug 3 was stable when incubated in human plasma for 4 h, indicating that 3 and 4 are not
substrates for human plasma esterases. In contrast, when incubated with human hepatocytes
at 37 °C both 3 and 4 were rapidly and completely metabolized to direct thrombin inhibitor 2
in 40 min (Figure 3). Identical results were obtained with rat hepatocytes (data not shown).
Metabolism of prodrug 3 results in formation of one equivalent of formaldehyde, which is a
potential safety concern. However, even at a dose of 1 g of 3/day the liberated formaldehyde
is comparable to the daily intake from food sources and <0.001% of the daily endogenous
27
turnover of formaldehyde. The methyl ester of acyclic analogue 5 was also cleaved to a
high extent in human hepatocytes, but the methyl carbamate turned out to be almost
completely resistant to metabolism during these conditions. When incubated with human
liver microsomes both 3 and 4 where degraded at equal rates in the presence and absence of
ACS Paragon Plus Environment

Page 17 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
NADPH, strongly indicating that conversion to 2 was mediated by esterases but not by
cytochrome P450 oxidases. In summary, prodrugs 3 and 4 had solubilities and permeabilities
across Cacoꢀ2 cells that suggested that they would display high absorption on oral
administration. Their rapid metabolism to direct thrombin inhibitor 2 in human and rat
hepatocytes also supported that they were progressed to in vivo ADME studies in rat.
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
R
O
O
R
HO
O
O
OH
OH
O
O
O
NH
NH
NH2
HN
N
HN
N
HN
N
Esterase
hydrolysis
Decomposition
O
O
O
HN
HN
HN
-
-
CO2
RCHO
O
O
b-4b
O
Cl
Cl
Cl
3
4
: R = H
: R = Me
3
2
>99% from
3
and in 40 min
4
CO Me
CO H
2
2
NHCO Me
NHCO Me
2
2
NH
N
NH
N
O
O
HN
Esterase
hydrolysis
HN
O
O
Cl
Cl
5
>
95% from
5
6
in 40 min
Figure 3. Esterase mediated cleavage of (acyloxy)alkoxy linked prodrugs 3 and 4 in human
9
hepatocytes, followed by chemical decomposition to give direct thrombin inhibitor 2. The
methyl ester of acyclic analogue 5 is also cleaved by esterases in the hepatocytes.
Bioavailability in rat. Prodrug 3 was selected for in vivo pharmacokinetic studies because
homologue 4 displayed a higher acid lability during purification, indicating that it might
undergo acid catalysed decomposition on oral administration. Pharmacokinetic data for direct
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
thrombin inhibitor 2 and prodrug 3 were determined in male Han Wistar rats after single 0.5
µmol/kg i.v. and 1 µmol/kg p.o. doses and calculated using nonꢀcompartmental analysis
(Table 4 and Figure 4). Plasma levels of 2 declined rapidly after i.v. dosing with a terminal
halfꢀlife of approximately 0.35 h. The total plasma clearance was 32 mL/min/kg, which is
about half the liver blood flow. The volume of distribution at steady state was determined to
be 0.23 L/kg, which is smaller than that of the exchangeable water and implies a low tissue
distribution. The systemic bioavailability (F) after p.o. dosing of 2 was 0.5%. Prodrug 3 was
not detected in plasma samples at any time point after p.o. dosing, which indicates a rapid
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
conversion to 2. The area under the plasma concentrationꢀtime curve at infinite time (AUCinf)
for 2 was approximately 5ꢀfold higher after p.o. administration of 3 compared to p.o.
administration of 2. This increase in systemic exposure illustrates the potential of using a
double prodrug macrocycle approach, but the absolute bioavailability of 2 after
administration of 3 was only 2.5%. Although, this is higher than when 2 is administered it is
still low and there could be several reasons for this. For instance, hydrolysis of the prodrug
could occur in the acidic environment during passage through the stomach and thereby limit
the bioavailability. However, administration of 3 directly in the duodenum in the rat did not
show an increased bioavailability of 2, and thus acid induced degradation does not explain
the low bioavailability. Instead we speculate that rapid metabolic degradation of prodrug 3
occurs either in the duodenum, or during uptake across the intestinal wall, and that
metabolism constitutes the major reason for the low bioavailabilities obtained on oral
administration of prodrug 3. In order to reduce the rate of metabolism, structurally different,
but still esterase sensitive macrocyclic prodrugs could be explored, e.g. prodrugs based on
2
8, 29
coumarinic or oꢀhydroxyphenylpropionic acid derivatives.
ACS Paragon Plus Environment

Page 19 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Table 4. Pharmacokinetic parameters for compound 2 after dosing of compounds 2 or 3 via
different routes of administration.
Intravenous
Per oral
Per oral
Intraduodenal
administration of 2 administration of 2 administration of 3 administration of 3
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
0.5 ꢀmol/kg)
32.4
(1 ꢀmol/kg)
(1 ꢀmol/kg)
(1 ꢀmol/kg)
Rat CL
mL/min/kg)
Rat Vss
ꢀ
ꢀ
ꢀ
(
(SD=1.3)
0.23
ꢀ
ꢀ
ꢀ
(L/kg)
(SD=0.05)
0.35
Terminal t1/2
(h)
0.56
(SD=0.006)
0.0025
1.12
(SD=0.23)
0.0127
0.11
(SD=0.01)
0.0081
(SD=0.04)
0.258
AUCinf of 2
(
h × ꢀmol/L)
F of 2 (%)
Cmax
(SD=0.010)
ꢀ
(SD=0.0005)
0.5
(SD=0.0012)
2.5
(SD=0.0026)
1.6
0.004
0.011
0.108
ꢀ
(
ꢀmol/L)
(SD=0.002)
(SD=0.002)
(SD=0.041)
a
Values for all data except the bioavailability (F) are means from n=3 experiments. F is
calculated from the average AUC values.
inf
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Plasma concentration over time for direct thrombin inhibitor 2 after intravenous
(i.v.), peroral (p.o.) or intraduodenal (i.d.) administration of 2 or prodrug 3 to rat.
Conclusions
Drug targets that require zwitterionic or other biꢀpolar ligands are ¨difficult to drug¨ as such
drug candidates often suffer from poor absorption from the gastrointestinal tract. In such
cases bioavailability may be improved by conversion into double prodrugs, but they are often
complicated to develop because of difficulties to obtain efficient cleavage of both prodrug
moieties. With this in mind we developed a synthetic route to zwitterionic, nonꢀpeptidic
direct thrombin inhibitor 2, which displayed high antithrombotic potency, excellent
selectivity against serine proteases and satisfactory in vitro metabolic properties. Inhibitor 2
was converted into macrocycles 3 and 4, each of which carried an (acyloxy)alkoxy prodrug
11
moiety across the carboxyl and amino group of 2. The two macrocyclic prodrugs displayed
high solubilities and also had high cell monolayer permeabilities, in contrast to lowꢀ
ACS Paragon Plus Environment

Page 21 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
permeable inhibitor 2. Moreover, the prodrug moieties of 3 and 4 were cleaved rapidly in
human hepatocytes, most likely by liver esterases. These in vitro studies suggested that both 3
and 4 would be highly absorbed from the intestine and then cleaved to liberate potent direct
thrombin inhibitor 2 in the liver. Unfortunately, the bioavailability of inhibitor 2 after oral
administration of prodrug 3 in rat was low (2.5%) and only moderately higher than on direct
oral administration of 2 (0.5%). We conclude that the low bioavailability after administration
of 3 most likely results from rapid metabolism of the (acyloxy)alkoxy prodrug moiety in the
intestine, or during uptake across the intestinal wall. To the best our knowledge, only in vitro
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10ꢀ15
data has been reported in previous studies
of cyclic (acyloxy)alkoxy prodrugs. Thus, our
study is the first that involves in vivo ADME studies and therefore we can not speculate if the
modest increase in bioavailability is a general feature of this category of prodrugs.
Furthermore, chemical developability at large scale constitutes a challenge for macrocyclic
(acyloxy)alkoxy prodrugs as the macrocyclization step requires high dilution and proceeds in
modest yields.
Experimental
General chemistry. Reagents and solvents used were of reagent grade and purchased from
commercial suppliers and used as received. Unless stated otherwise, reactions were
performed under nitrogen atmosphere, additions of reagents and/or solvents with sterilized
disposable syringes and needles, Silicone Oil (550 Fluid) was used for heatings, stirring was
made with magnetic bars, evaporation of solvents was done under vacuum with a rotavapor.
Glassware were dried in oven (150 °C), when necessary. Building blocks for synthesis, i.e.
aldehyde 10 and chloroformates 21-22 were from SigmaꢀAldrich, while amine 12 was
purchased from Syngene. Reactions were monitored with (i) thinꢀlayer chromatography
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
(TLC), (ii) highꢀperformance liquid chromatography mass spectrometry (HPLC/MS), (iii) H
NMR or combinations of (i), (ii) and (iii). (i) TLC analysis was carried out on silica gel plates
4
(Merck 60 F254). The plates were stained with basic KMnO or CAM (ceric ammonium
molybdate) manually prepared solutions from available protocols. (ii) Analytical HPLC/MS
was conducted on a QTOF mass spectrometer using a UV detector monitoring either at (a)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
10 nm with a BEH C18 column (2.1 × 100 mm, 1.7 ꢁm, 0.7 mL/min flow rate), using a
gradient of 2% v/v CH CN in H O (ammonium carbonate buffer pH 10) to 98% v/v CH
in H O or at (b) 230 nm with an HSS C18 column (2.1 × 100 mm, 1.8 ꢁm, 0.7 mL/min flow
3
2
3
CN
2
rate) or C18 CSH 1.7ꢁm column (2.1x50mm, 0.7 mL/min flow rate), using a gradient of 2%
v/v CH CN in H O (ammonium formate buffer pH 3) to 98% v/v CH CN in H O. Samples
3
2
3
2
3
diluted in MeOH or CH CN, 2 or 4 min measurements. (iii) NMR measurements were
recorded on a Bruker Avance DPX400 (400 MHz), AV500 (500 MHz) or AV600 (600 MHz)
instrument at 25 °C as indicated. The chemical shifts (δ) are expressed as parts per million
(ppm) relative to residual solvent signal as internal reference at 7.26 ppm (CDCl ) or at 3.30
3
ppm (MeODꢀd ) and coupling constants (J) values are reported in Hertz (Hz). Splitting
4
pattern abbreviations are as follows: s = singlet; d = doublet; t = triplet; q = quartet; p =
pentet; m = multiplet; br = broad; om = overlapping multiplet. Silica gel column
chromatography was carried out on preꢀpacked silica gel columns SNAP (KPꢀSil) cartridges
supplied by Biotage and using Biotage automated flash systems with UV detection.
Preparative Highꢀperformance liquid chromatography (HPLC) was performed by a Waters
Xbridge C18 OBD 5 ꢁm column (19 × 150 mm, flow rate 30 mL/min, or 30 × 150 mm, flow
3 3
rate 60 mL/min) using a varying gradient of CH CN with 0.2% NH at pH 10 with UV
detection. For purification of final compounds 3ꢀ4, Waters Sunfire ODB MSꢀtriggered
fraction collector was utilized with a UV/VIS detector 155 on a C18 5 ꢁm column (19 × 150
mm, flow rate 19 mL/min, or 50 × 250 mm, flow rate 100 mL/min) using a varying gradient
ACS Paragon Plus Environment

Page 23 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of CH
3
CN and 0.1 M NH
4
OAc. The purity of each compound tested was determined with
analytical method (b), relative UV absorbance at 230 nm. All tested compounds were ≥ 95%
pure. Highꢀresolution mass spectrometry (HRMS) was determined as for (ii) on a Waters
Xevo QTOF, except for compounds 8, 23-24, which were determined on a GCꢀFIꢀTOFꢀMS
setup Waters GCT HPꢀ5MS 0.25µm column (30 x 0.32 mm, 2 mL/min flow rate) using He as
carrier gas.
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3-((3-(2-(2-(Aminomethyl)-5-chlorobenzylamino)-2-oxoethyl)-4-methyl-2-oxo-5,6-
dihydropyridin-1(2H)-ylamino)methyl)benzoic acid (2). TFA (5 mL, 66 mmol) was added
to 6 (86 mg, 0.147 mmol) in CH Cl (5 mL) and the solution was stirred at rt for 40 min. The
2
2
solvents were removed in vacuo and to give pure 2 as a TFA salt (86 mg, 97 %). The product
1
was used directly without further purification. H NMR (500 MHz, DMSO) δ 1.81 (s, 3H,
H11), 2.30 (t, J = 7.1 Hz, 2H, H2), 3.19 (s, 2H, H7), 3.30 (t, J = 7.2 Hz, 2H, H1), 3.96 (s, 2H,
H14), 4.7ꢀ6.3 (br, 1H, H22), 4.11 (q, J = 5.7 Hz, 2H, H32), 4.29 (d, J = 6.0 Hz, 2H, H24),
7.46ꢀ7.40 (om, 3H, H28/29/26), 7.44 (t, J = 7.7 Hz, 1H, H19), 7.57 (dt, J = 1.2, 7.6 Hz, 1H,
H20), 7.84 (dt, J = 1.4, 7.7 Hz, 1H, H18), 7.93 (s, 1H, H16), 8.15 (br s, 3H, H33/H10), 8.39
13
(t, J = 6.0 Hz, 1H, H9). C NMR (126 MHz, DMSO) 170.5, 167.3, 164.7, 158.0, 147.5,
140.7, 138.7, 133.4, 131.5, 130.7, 130.4, 129.8, 128.5, 128.2, 127.0, 122.5, 53.0, 46.8, 33.4,
+
30.3, 20.2. HRMS (ESI) m/z calcd for C H N O Cl [M + H] 471.1799 ; found 471.1807.
24
27
4
4
25-Chloro-18-methyl-13,14,17,20,22,23-hexahydro-1H,7H-15,19-methano-8,12-
(metheno)-4,6,2,14,15,22-benzodioxatetraazacyclopentacosine-3,7,21,28(2H,16H)-
tetrone (3). Cesium carbonate (14 mg, 40 ꢁmol) was added to a solution of 6 (40 mg, 70
ꢁmol) in methanol (0.5 mL) and was stirred at rt for 1 h. The solvent was removed in vacuo
to generate the cesium salt as a light green solid. The solid was redissolved in dry DMF (0.6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
mL) and the solution was added to a solution of 8 (23 mg, 70 ꢁmol) dissolved in dry DMF
(0.6 mL) at 0 °C during 30 min. The reaction was monitored with LCꢀMS at pH 3 and was
stirred at 0 °C for an additional 4 h. The solvent was removed in vacuo and the residue was
redissolved in EtOAc (5 mL) and was then sequentially washed with H O (5 mL), brine (5
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
mL) and concentrated in vacuo to give a residue containing 49 % alkylated intermediate 25.
Crude 25 was dissolved in CH
and the solution was stirred at rt for 6 h. The reaction was monitored with LCꢀMS at pH 3
which indicated complete NꢀBoc deprotection. The solution was diluted with CH Cl (80 mL)
2 2
Cl (3 mL), after which TFA (0.15 mL, 2.0 mmol) was added
2
2
to obtain a concentration of 0.7 mM of the deprotected intermediate. It was then washed with
aqueous 10 % K CO (5 mL), filtered with a phase separator and slowly stirred at rt for 72 h.
2
3
The reaction was monitored with LCꢀMS at pH 10 which indicated that macrocycle 3 had
been formed. The solvent was removed in vacuo and the residue was purified with
2 4
preparative HPLC, (10 to 70 % acetonitrile in H OꢀNH OAc) to give macrocycle 3 (6 mg, 16
1
% from 6) as a white solid. H NMR (600 MHz, CDCl , contains set of two rotamers) δ 2.02
3
(
s, 1.5H, H6), 2.07 (s, 1.5H, H6), 2.08 – 2.16 (m, 1H, H7), 2.26 (d, J = 7.0 Hz, 1H, H7), 2.94
s, 1H, H8), 3.16 (d, J = 7.2 Hz, 1H, H8), 3.28 (s, 1H, H9), 3.33 (s, 1H, H9), 3.87 (s, 1H,
(
H14), 3.89 (s, 1H, H14), 4.24 (d, J = 6.6 Hz, 1H, H30), 4.36 (d, J = 6.1 Hz, 1H, H22), 4.38
d, J = 5.7 Hz, 1H, H30), 4.42 (bs, 1H, H22), 5.93 (s, 1H, H34), 5.96 (s, 1H, H34), 6.21 –
(
6.35 (m, 1H, H31), 6.93 – 6.97 (m, 0.5H, H26), 6.99 (s, 0.5H, H24), 7.06 – 7.1 (m, 0.5H,
H26), 7.21 (s, 0.5H, H16), 7.22 – 7.29 (om, 2H, H16/24/27), 7.34 – 7.46 (m, 1H, H19), 7.54
(d, J = 7.5 Hz, 0.5H, H20), 7.59 (d, J = 7.6 Hz, 0.5H, H20), 7.64 – 7.68 (m, 0.5H, H11), 7.74
1
3
(
d, J = 7.5 Hz, 0.5H, H18), 7.92 (d, J = 7.7 Hz, 0.5H, H18), 8.00 – 8.06 (m, 0.5H, H11); C
, contains set of two rotamers) δ 171.1 (C10), 171.0 (C10), 166.5
C2), 166.0 (C2), 165.1 (C21), 164.8 (C21), 154.4 (C36), 154.0 (C36), 151.0 (C5), 149.0
NMR (151 MHz, CDCl
3
(
(
C5), 139.0 (C23), 138.6 (C23), 138.0 (C15), 137.8 (C15), 134.6 (C20/C28), 134.5 (C20),
ACS Paragon Plus Environment

Page 25 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
134.2 (C25), 133.8 (C25), 133.3 (C28), 132.0 (C27), 131.9 (C27), 131.0 (C16), 129.9 (C16),
1
29.3(C17), 129.2 (C18), 128.9 (C18), 128.8 (C19), 128.7(C17), 128.6 (C24), 128.3(C24),
27.4 (C26), 127.2 (C26), 122.2 (C4), 121.3 (C4), 81.8 (C34), 81.2 (C34), 53.9 (C14), 53.7
1
(C14), 47.6 (C8), 47.5 (C8), 42.6 (C30), 42.0 (C30), 39.6 (C22), 38.9 (C22), 35.3 (C9), 35.3
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(C9), 30.6 (C7), 30.4 (C7), 20.8 (C6), 20.6 (C6). HRMS (ESI) m/z calcd for C H N O Cl
26
28
4
6
+
[
M + H] 527.1697; found 527.1574.
25-Chloro-5,18-dimethyl-13,14,17,20,22,23-hexahydro-1H,7H-15,19-methano-8,12-
(
metheno)-4,6,2,14,15,22-benzodioxatetraazacyclopentacosine-3,7,21,28(2H,16H)-
tetrone (4). Cesium carbonate (34 mg, 0.11 mmol) was added to a solution of 6 (100 mg,
.18 mmol) in methanol (1 mL) and was stirred at rt for 1 h. The solvent was removed in
0
vacuo to generate the cesium salt as a light green solid. The solid was redissolved in dry
DMF (1.2 mL) and the solution was added to a solution of 9 (81 mg of 75% w/w, 0.18 mmol)
dissolved in dry DMF (1.2 mL) at 0 °C during 30 min. The reaction was monitored with LCꢀ
MS at pH 3 and was stirred at 0 °C for an additional 4 h. The solvent was removed in vacuo
and the residue was redissolved in EtOAc (10 mL) and was then sequentially washed with
H
2
O (10 mL), brine (10 mL) and concentrated in vacuo to give a residue containing 25 %
alkylated intermediate 26. Crude 26 was dissolved in CH Cl (4 mL), after which TFA (0.15
2
2
mL, 2.0 mmol) was added and the solution was stirred at rt for 6 h. The reaction was
monitored with LCꢀMS at pH 3 which indicated complete NꢀBoc deprotection. The solution
was diluted with CH
2
Cl
2
(150 mL) to obtain a concentration of 0.7 mM of the deprotected
CO (10 mL), filtered with a phase
intermediate. It was then washed with aqueous 10 % K
2
3
separator and slowly stirred at rt for 72 h. The reaction was monitored with LCꢀMS at pH 10,
+
which indicated that macrocycle 4 (observed [M + H] : 497 – loss of CO ) had been formed.
2
The solvent was removed in vacuo and the residue was purified with preparative HPLC, (10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to 70 % acetonitrile in H
2
OꢀNH
4
OAc) to give macrocycle 4 (8 mg, 8 % from 6) as a white
, contains set of two rotamers) δ 1.65 (d, J = 5.4 Hz, 1.5H,
1
solid. H NMR (500 MHz, CDCl
3
H38) 1.75 (d, J = 5.4 Hz, 1.5H, H38), 1.96 – 2.03 (om, 1.5H, H7/8), 2.05 (s, 1.5H, H6), 2.08
s, 1.5H, H6), 2.53 (s, 1H, H7), 2.74 (s, 0.5H, H8), 3.02 – 3.15 (m, 1H, H8/9), 3.22 – 3.36
(om, 2H, H8/9), 3.38 – 3.45 (m, 0.5H, H9), 3.62 (d, J = 12.7 Hz, 0.5H, H14), 3.80 – 3.98 (m,
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2.5H, H14/22), 4.10 – 4.19 (m, 0.5H, H30), 4.23 (s, 0.5H, H30), 4.44 (dd, J = 8.3, 15.0 Hz,
0.5H, H30), 4.57 (dd, J = 8.5, 14.7 Hz, 0.5H, H30), 4.82 (dd, J = 9.0, 14.9 Hz, 0.5H, H22),
5.01 – 5.10 (m, 0.5H, H22), 6.13 – 6.17 (m, 0.5H, H31), 6.19 (s, 0.5H, H31), 6.79 (d, J = 5.5
Hz, 0.5H, H34), 6.87 (s, 0.5H, H16), 6.94 – 7.01 (om, 1.5H, H26/34), 7.14 (d, J = 6.1 Hz,
0.5H, H26), 7.24 (d, J = 9.0 Hz, 0.5H, H16), 7.27 – 7.35 (om, 1.5H, H24/27), 7.39 – 7.48 (m,
1
.5H, H19/27), 7.58 (d, J = 7.8 Hz, 0.5H, H20), 7.64 (d, J = 7.0 Hz, 0.5H, H20), 7.75 (s,
.5H, H11), 7.88 (d, J = 7.8 Hz, 0.5H, H18), 7.91 (d, J = 7.5 Hz, 0.5H, H18), 8.23 (s, 0.5H,
0
13
H11); C NMR (151 MHz, CDCl
3
, contains set of two rotamers) δ 171.2 (C10), 171.1 (C10),
166.5 (C2), 166.3 (C2), 164.7(C21), 154.3 (C36), 153.4 (C36), 151.8 (C5), 149.4 (C5), 138.7
(C23), 138.7 (C23), 138.1 (C15), 137.8 (C15), 135.1 (C28), 135.1 (C28), 135.0 (C20), 134.5
(C20), 134.1 (C17), 134.0 (C17), 132.9 (C27), 132.5 (C27), 130.6 (C16), 129.9 (C16), 129.7
(C25), 129.5 (C25), 129.4 (C18), 129.1 (C18), 128.9 (C19), 128.8 (C19), 127.6 (C26), 127.4
(C26), 122.1 (C4), 120.6 (C4), 91.4 (C34), 90.9 (C34), 54.7 (C14), 53.9 (C14), 48.2 (C8),
42.3 (C30), 41.9 (C30), 39.4 (C22), 39.0 (C22), 35.5 (C9), 35.4 (C9), 30.6 (C7), 30.3 (C7),
0.8 (C6), 20.7 (C6), 19.7 (C38), 19.5 (C38). HRMS (ESI) m/z calcd for C H N O Cl [M +
2
27
30
4
6
+
H] 541.1854; found 541.1841.
Methyl 3-(((3-(2-((5-chloro-2-(((methoxycarbonyl)amino)methyl)benzyl)amino)-2-
oxoethyl)-4-methyl-2-oxo-5,6-dihydropyridin-1(2H)-yl)amino)methyl)benzoate (5). TFA
(80 ꢁL, 1.0 mmol) was added to a solution of 7 (35 mg, 60 ꢁmol) in CH Cl (1.5 mL) and the
2 2
ACS Paragon Plus Environment

Page 27 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reaction was stirred at rt for 6h. The reaction was diluted with CH
2
Cl
2
(5 mL) and
O (2 x 10 mL),
), filtered and concentrated in vacuo to give the deprotected
amine and was redissolved in CH Cl (0.8 mL). To this solution was added methyl
sequentially washed with aqueous solution of 10 % K
brine (2 x 10 mL), dried (Na SO
2 3 2
CO (2 x 10 mL), H
2
4
2
2
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
carbonochloridate (4.7 µL, 60 ꢁmol) dissolved in CH Cl (1 mL) dropwise at 0 °C for 20
2
2
min. The reaction was carefully monitored with LCꢀMS at pH 10 and the reaction was
quenched with an aqueous 5% HCl solution (15 mL) and sequentially washed with saturated
aqueous NaHCO
3 2 2 4
(2 x 10 mL), H O (2 x 10 mL), brine (2 x 10 mL), dried (Na SO ), filtered
and concentrated in vacuo to give a 2:1 mixture of monoꢀacylated 5 and biꢀacylated
byproduct. The product was purified with preparative HPLC at pH10 (15 to 70 % acetonitrile
1
2
in H OꢀNH
3
OAc) to give 5 (23 mg, 71 %) as a white solid. H NMR (500 MHz, CDCl
3
) δ
2
.01 (s, 3H, H11), 2.34 (t, J = 7.2 Hz, 2H, H2), 3.29 (s, 2H, H7), 3.37 (t, J = 7.2 Hz, 2H, H1),
.63 (s, 3H, H37), 3.92 (s, 3H, H38), 4.00 (s, 2H, H14), 4.34 (d, J = 5.7 Hz, 2H, H32), 4.36
3
(d, J = 6.0 Hz, 2H, H24), 5.87 (br, 1H), 7.13 – 7.16 (m, 1H, H29), 7.19 (dd, J = 2.1, 8.1 Hz,
1H, H28), 7.30 (d, J = 8.1 Hz, 1H, H26), 7.36 (br, 1H), 7.39 (t, J = 7.7 Hz, 1H, H19), 7.51 –
13
7
.56 (m, 1H, H20), 7.92 – 7.97 (m, 1H, H18), 8.04 (s, 1H, H16); C NMR (126 MHz,
CDCl
3
) δ 167.1, 166.3, 157.3, 148.9, 138.5, 138.0, 135.1, 134.1, 133.7, 131.2, 130.7, 130.1,
129.2, 128.7, 128.4, 128.1, 127.7, 54.4, 52.8, 51.9, 48.2, 41.4, 40.3, 35.6, 31.8, 31.1. HRMS
+
(ESI) m/z calcd for C H N O Cl [M + H] 543.2010; found 543.2025.
27
32
4
6
3
-(((3-(2-((2-(((Tert-butoxycarbonyl)amino)methyl)-5-chlorobenzyl)amino)-2-oxoethyl)-
-methyl-2-oxo-5,6-dihydropyridin-1(2H)-yl)amino)methyl)benzoic acid (6). Lithium
4
hydroxide (9 mg, 0.35 mmol) and a catalytic amount of methanol (1.2 µL, 30 ꢁmol) were
added to a solution of 7 (170 mg, 0.29 mmol) in 1,4ꢀdioxane and water (1:1, 1 mL). The
mixture was stirred at rt for 8 h. It was then diluted with EtOAc (5 mL), washed with aqueous
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 48
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2 4
HCl (pH 5, 3 x 5 mL), dried (Na SO ), filtered with a phase separator and concentrated in
vacuo to give carboxylic acid 6 (157 mg, 95 %) as a white solid. The product was used
1
directly in the next step without further purification. H NMR (500 MHz, MeODꢀd4) δ 1.44
(s, 9H, H38/39/40), 1.93 (s, 3H, H11), 2.37 (t, J = 7.2 Hz, 2H, H2), 3.34 (s, 2H, H7), 3.37 (t,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
J = 7.2 Hz, 2H, H1), 4.02 (s, 2H, H14), 4.26 (s, 2H, H32), 4.39 (s, 2H, H24), 7.20 – 7.24 (m,
1H, H28), 7.25 – 7.28 (m, 1H, H29), 7.28 – 7.30 (m, 1H, H26), 7.35 (t, J = 7.6 Hz, 1H, H19),
13
7
.47 – 7.53 (m, 1H, H20), 7.86 – 7.90 (m, 1H, H18), 7.98 – 8.02 (m, 1H, H16); C NMR
(126 MHz, MeODꢀd4) 173.6, 167.0, 158.4, 150.7, 145.6, 139.6, 138.9, 137.0, 134.1, 133.2,
131.5, 130.7, 129.7, 129.1, 128.9, 128.2, 123.6, 80.3, 55.1, 48.3, 41.9, 41.1, 38.5, 34.8, 31.7,
+
28.8, 20.6. HRMS (ESI) m/z calcd for C H N O Cl [M + H] 571.2323; found 571.2338.
29
36
4
6
3
-(((3-(2-((2-(((Tert-butoxycarbonyl)amino)methyl)-5-chlorobenzyl)amino)-2-oxoethyl)-
-methyl-2-oxo-5,6-dihydropyridin-1(2H)-yl)amino)methyl)benzoate (7). HATU (1.10 g,
4
2.91 mmol), DIPEA (1.01 mL, 5.82 mmol) and benzylic amine 12 (578 mg, 2.13 mmol) were
sequentially added to a solution of carboxylic acid 20 (645 mg, 1.94 mmol) in CH Cl and
2
2
NMP (1:1; 6 mL). The solution was stirred at 40 °C for 6 h then diluted with EtOAc (15 mL)
and sequentially washed with saturated aqueous NaHCO (20 mL), brine (20 mL), saturated
aqueous NH Cl (20 mL) and brine (20 mL), dried (Na SO ), filtered with a phase separator
3
4
2
4
and concentrated in vacuo. The residue was purified with preparative HPLC at pH 10 (15 to
1
8
0 % acetonitrile in H OꢀNH ) to give 7 (810 mg, 71 %) as a white solid. H NMR (500
2
3
MHz, MeODꢀd4) δ 1.44 (s, 9H, H38/39/40), 1.93 (s, 3H, H11), 2.38 (t, J = 7.2 Hz, 2H, H2),
.34 (s, 2H, H7), 3.40 (t, J = 7.2 Hz, 2H, H1), 3.90 (s, 3H, H41), 4.03 (s, 2H, H14), 4.25 (s,
3
2H, H32), 4.39 (s, 2H, H24), 7.22 (dd, J = 1.7, 8.2 Hz, 1H, H28), 7.27 (d, J = 8.1 Hz, 2H,
H26/29), 7.42 (t, J = 7.7 Hz, 1H, H19), 7.63 (d, J = 7.6 Hz, 1H, H20), 7.91 (d, J = 7.8 Hz,
13
1
H, H18), 8.05 (s, 1H, H16); C NMR δ (126 MHz, MeODꢀd4) 173.3, 168.2, 166.9, 158.1,
C
ACS Paragon Plus Environment

Page 29 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
150.4, 139.6, 139.5, 136.8, 135.1, 133.9, 131.3, 131.2, 130.6, 129.6, 129.5, 128.7, 128.1,
123.6, 80.2, 54.6, 52.7, 48.1, 41.9, 41.1, 34.9, 31.6, 28.8, 20.6. HRMS (ESI) m/z calcd for
+
C
30
H
38
N
4
O
6
Cl [M + H] 585.2480; found 585.2497.
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
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
0
1-Iodomethyl (4-nitrophenyl) carbonate (8). Prepared as previously reported with minor
1
modifications. The H NMR is in agreement with reported data and the product was used
8 6 5
directly used without further purifications, see S.I. HRMS (FI) m/z calcd for C H NO I [M +
+
H] 322.9291; found 322.9296.
1-Iodoethyl (4-nitrophenyl) carbonate (9). Sodium iodide (1.19 g, 7.94 mmol) was added
1
to 24 (650 mg, 2.65 mmol) in dry acetone (4 mL). The reaction was monitored with H NMR
spectroscopy and was heated at 50 °C for 48 h in a capped vial. The reaction was stopped
after 75% conversion to prevent the competing elimination of ethene. The solvent was
removed in vacuo and redissolved in tertꢀbutyl methyl ether (7.5 mL), sequentially washed
with aqueous 10 % Na SO (3 x 10 mL), H O (3 x 15 mL), brine (3 x 15 mL), dried
2
3
2
(Na
2
SO
4
), filtered and concentrated in vacuo to give an inseparable mixture of 9 (560 mg, 47
1
%
) and starting material (3:1) as a yellow oil. The yield is based on integration of the H
3
1
NMR spectrum, which is in agreement with reported data, see S.I. The product was used
directly without further purification. All attempts to obtain HRMS for this compound failed.
2
-(Trimethylsilyl)ethyl 2-(1-amino-4-methyl-2-oxo-1,2,5,6-tetrahydropyridin-3-
yl)acetate (11). Nanopowder zinc (927 mg, 14.2 mmol) and concentrated acetic acid (10 mL,
168 mmol) were added to a solution of Nꢀnitroso compound 17 (1410 mg, 4.70 mmol) in dry
methanol (10 mL) at 0 °C. The mixture was stirred at 0 °C for 20 min, after which the greenꢀ
yellow colour had disappeared. The mixture was then allowed to attain rt, filtered through
ACS Paragon Plus Environment