Design and Synthesis of 4,5-disubstituted-thiophene-2-amidines as potent urokinase inhibitors

Article abstract of DOI:10.1016/S0960-894X(01)00787-9

A study of the S1 binding of lead 5-methylthiothiophene amidine 3, an inhibitor of urokinase-type plasminogen activator, was undertaken by the introduction of a variety of substituents at the thiophene 5-position. The 5-alkyl substituted and unsubstituted thiophenes were prepared using organolithium chemistry. Heteroatom substituents were introduced at the 5-position using a novel displacement reaction of 5-methylsulfonylthiophenes and the corresponding oxygen or sulfur anions. Small alkyl group substitution at the 5-position provided inhibitors equipotent with 3 but possessing improved solubility.

Full text of DOI:10.1016/S0960-894X(01)00787-9

Bioorganic & Medicinal Chemistry Letters 12 (2002) 491–495
Design and Synthesis of 4,5-Disubstituted-thiophene-2-amidines
as Potent Urokinase Inhibitors
M. Jonathan Rudolph,* Carl R. Illig, Nalin L. Subasinghe, Kenneth J. Wilson,
James B. Hoffman, Troy Randle, David Green, Chris J. Molloy, Richard M. Soll,
Frank Lewandowski, Marie Zhang, Roger Bone, John C. Spurlino, Ingrid C. Deckman,
Carl Manthey, Celia Sharp, Diane Maguire, Bruce L. Grasberger,
Renee L. DesJarlais and Zhao Zhou
3-Dimensional Pharmaceuticals Inc., 665 Stockton Drive, Exton, PA 19341, USA
Received 24 July 2001; accepted 19 November 2001
Abstract—A study of the S1 binding of lead 5-methylthiothiophene amidine 3, an inhibitor of urokinase-type plasminogen acti-
vator, was undertaken by the introduction of a variety of substituents at the thiophene 5-position. The 5-alkyl substituted and
unsubstituted thiophenes were prepared using organolithium chemistry. Heteroatom substituents were introduced at the 5-position
using a novel displacement reaction of 5-methylsulfonylthiophenes and the corresponding oxygen or sulfur anions. Small alkyl
group substitution at the 5-position provided inhibitors equipotent with 3 but possessing improved solubility. # 2002 Elsevier
Science Ltd. All rights reserved.
Urokinase-type plasminogen activator (uPA) is a serine
protease that is involved in many cellular processes
including angiogenesis, tissue remodeling, tumor inva-
sion, and cancer metastasis.1,2 The ability of tumor cells
to invade and metastasize can be downregulated by uPA
inhibitors, making this an attractive target for drug dis-
covery.3,4 Several novel inhibitors of this enzyme have
been reported, all having a common guanidine or ami-
dine functionality. Included are: cyclohexylthio-phene-
2-carboxamidines,5 naphthylamidines,6 isoquino-linyl-
By screening amidine libraries, 5-methylthiothiophene-
2-carboxamidine 2 was identified as an inhibitor with a
Ki of 6 mM. A molecular modeling comparison of tem-
plate 2 to inhibitor 1 indicated that substitution offthe
4-position of the thiophene should lead to increased
potency.6 This ultimately led to the new lead inhibitor 3
with a Ki of 101 nM.14 In addition to its modest
potency, 3 possessed marginal solubility. It was decided
to further explore groups that would be tolerated at the
thiophene 5-position in the S1 pocket and develop an
SAR (Fig. 1).
guanidines,7
and
benzo[b]thiophene-2-carboxami-
dines,8,9 for example 1 (Eisai, B428, Ki=0.32 mM). The
inhibitor 1 has been used in recent X-ray crystal and
NMR structure studies of uPA.10,11 The structural
studies show the effects of varying templates and the
trajectory into the proximal pocket.12 This trypsin-like
serine protease prefers a positively charged P1 that
makes a salt bridge with Asp189 in the S1 pocket.13 Our
primary goal was to optimize S1 binding of any emerg-
ing leads prior to optimizing binding in the proximal
pocket.
*Corresponding author. Tel.: +1-610-458-5264 fax: +1-610-458-
8249; e-mail: rudolph@3dp.com
Figure 1. Lead thiophene amidine compounds.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00787-9

492
M. J. Rudolph et al. / Bioorg. Med. Chem. Lett. 12 (2002) 491–495
Scheme 1. (a) NaClO2, 20% NaH2PO4, t-BuOH, 2-methyl-2-butene, 0 ꢀC to rt (93%); (b) (i) SOCl2, CH3OH, ꢁ20 ꢀC to reflux (95%) for 16;
or (i) CH2Cl2, oxalyl chloride, then i-PrOH, Na2CO3 (32%) for 20; or 2 M TMSCHN2 in hexanes (100%) for 20; or (i) CH2Cl2, oxalyl chloride, then
i-PrOH, CH2Cl2, pyridine (48%) for 23; (ii) CuCN, DMF, reflux (58% for R1=H, 63% for R1=CH3, 53% for R1=CH2CH3, 68% for R1=CH3,
R2=CH3; (iii) H2S, TEA, MeOH (82% for 17, 56% for 21, 74% for 24, 48% (mixture with 60% starting nitrile) for 25); (c) (i) a-bromoacetophe-
none, acetone, reflux (90% for 17, 66% for 21, 49% for 24, 53% for 25); (ii) Al(CH3)3, NH4Cl, toluene, reflux (45% for 6, 100% for 7, 67% for 8,
9% for 9); (d) (i) LDA, ꢁ78 ꢀC, (ii) CH3I (59%); (e) (i) n-butyllithium, ꢁ78 ꢀC; (ii) CO2; (g) (iii) 6 N HCl (aq) (94%); (f) (i) n-BuLi (2 equiv); (ii)
CH3CH2I (85%).
Scheme 2. (a) m-CPBA, CH2Cl2, rt, 2 h (95%); (b) NaOCH3, CH3OH, reflux, 15 min (73%); (c) p-methoxybenzyl thiol, TEA, CH3OH, reflux, 15
min (95%); (d) H2S, TEA, MeOH (59% for R=OCH3; 48% for R=p-CH3O-benzyl); (e) 2-bromobenzophenone, acetone, reflux (10% for
R=OCH3; 49% for R=p-CH3O-benzyl); (f) Al(CH3)3, NH4Cl, toluene, reflux (69% for 10; 91% for 11).
Modeling suggested that since the orientation of the
thiophene in S1, and therefore the resulting trajectory
offthe thiophene 4-position, would be expected to
change for different 5-position substituents, this opti-
mization had to be undertaken first. The introduction
of 5-alkyl substituents would require their incorporation
early in the synthesis. For 5-heteroatom substitutions,
the strategy was to develop chemistry to carry out sin-
gle-step modifications on the intact inhibitor.15
one equivalent of LDA resulted in a stable 5-lithio-2,4-
dibromothiophene intermediate via a ‘base-catalyzed
halogen dance’ mechanism which was trapped by iodo-
methane and gave 2,4-dibromo-5-methylthiophene 19.23
Selective lithiation of 19, and treatment of the inter-
mediate with excess CO2 gas gave the acid 20.24 Esteri-
fication with oxalyl chloride and isopropanol, then
treatment with copper cyanide resulted in the nitrile,
which was converted to thioamide 21 as before. The
preparation of the more methanol-insoluble isopropyl
esters was necessary since the methyl esters did not pre-
cipitate during the reaction, leading to an inseparable
mixture of unreacted nitrile and thioamide. Precipita-
tion of the thioamide product from solution under these
conditions was necessary to push this reaction to
completion. Hantzch condensation and amidinylation
gave 7.
The inhibitors (Table 1) were synthesized in a straight-
forward manner and the synthesis of 4-aryl-5-methyl-
thiothiophene compounds 3, 4, and 5 were described
previously.16
The 5-unsubstituted thiophene
6 was synthesized
(Scheme 1) starting with 4-bromothiophene-2-aldehyde
15. Oxidation to the acid via the method of Kraus17
gave acid 16. Esterification18 and treatment with cop-
per(I) cyanide gave the nitrile,19 which was then con-
verted to thioamide 17 with hydrogen sulfide.20 Hantsch
condensation of 17 with 2-bromo-acetophenone21 and
treatment of the resulting ester with trimethylaluminum
and ammonium chloride22 gave amidine 6.
The 5-ethylthiophene 8 was synthesized (Scheme 1) start-
ing with commercially available 4,5-dibromothiophene-2-
carboxylic acid25 22. Selective lithiation with n-butyl
lithium and trapping with iodoethane gave 23. Sub-
sequent transformations were carried out as above to
give 8.
The 5-methylthiophene 7 was prepared (Scheme 1)
starting with 2,5-dibromothiophene 18. Treatment with
The 3,4-dimethoxyphenyl-5-methylthiophene 9 was
prepared (Scheme 1) in a manner similar to 7, but using

M. J. Rudolph et al. / Bioorg. Med. Chem. Lett. 12 (2002) 491–495
493
30,40-dimethoxy-a-bromoacetophenone as the con-
densation partner during the Hantzch synthesis.16
For the introduction of heteroatoms in the 5-position, a
different approach was taken. We reasoned that a suffi-
ciently electron-deficient thiophene might allow direct
displacement of a sulfone moiety which would be easily
derived from a readily available sulfide. Thus, synthesis
(Scheme 2) of 5-methoxythiophene 10 started with nitri-
loester 26.26 Oxidation with m-chloroperbenzoic acid gave
the sulfone 27.18 In practice, the sulfone was readily dis-
placed by refluxing with sodium methoxide in methanol
to give 5-methoxythiophene nitrile 28.27 The 5-methoxy
derivative 10 was then prepared as described above.
Scheme 3. (a) m-CPBA, CH2Cl2, 1.5 h, rt (77%); (b) Al(CH3)3,
NH4Cl, toluene, reflux (31% for 12, 53% for 13); (c) benzyl mercap-
tan, TEA, CH3OH, reflux, 15 min (85%); (d) 30% H2O2, 1,1,1,3,3,3-
hexafluoro-2-propanol, CH2Cl2, 45 h, rt (90%).
The synthesis (Scheme 2) of 5-(4-methoxybenzyl-
mercapto)thiophene 11 was also carried out by the dis-
placement of sulfone 27 with the triethylammonium salt
of 4-methoxybenzylmercaptan in refluxing methanol
and gave 29.27 Following the subsequent transforma-
tions for 7 gave 11.
Table 1. SAR of the 5-substituted thiophene
The synthesis (Scheme 3) of the 5-sulfoxide 12 started
with commercially available26 5-methylthio ester 30.
Selective mono-oxidation to sulfoxide 31 was carried
out with 30% hydrogen peroxide and 1,1,1,3,3,3-hexa-
fluoro-2-propanol in dichloromethane.28 This ester 31
was amidinylated as above to give 12.22
Compd
X
Aryl
Ki (mM)
3
—SCH3
0.101
The 5-methylsulfonylthiophene 13 (Scheme 3) was syn-
thesized by peroxidation of 5-methylthiothiophene ester
30 with mCPBA, which gave 5-methylsulfonylthiophene
ester 31,18 then amidinylated to give 13.22
4
—SCH3
—SCH3
—H
0.058
0.102
5.49
5
The 5-benzylmercaptothiophene 14 (Scheme 3) was
synthesized by the same sulfone displacement used to
make 28, but by carrying out this transformation on
preformed amidine 13. Thus, 13 was treated with the
triethylammonium salt of benzyl mercaptan in refluxing
methanol to yield 14.27 Interestingly, subjecting sulfonyl
ester 32 to the same conditions resulted in unchanged
starting material. The absence of the electron-with-
drawing nitrile group in 32 was compensated for by the
amidine group in 13. Further exploitation of this useful
transformation is in progress.
6
7
—CH3
0.103
0.138
0.169
0.531
1.52
8
—CH2CH3
—CH3
9
All compounds were assayed against human kidney cell
urokinase,29 and the SAR results are shown in Table 1.
Estimated solubilities of the most potent inhibitors
were measured.30 During the course of this project,
three different substitutions were employed for the aryl
component. Whereas the p-chlorophenyl derivative 5
(Ki 102 nM) is essentially equipotent with the unsub-
stituted parent phenyl derivative 3, the 3,4-dimethoxy-
10
11
12
13
14
—OCH3
phenyl derivative
4 is almost twice as potent,
1.42
presumably as a result of filling the proximal pocket to a
greater degree. Modeling suggested that one method to
improve potency would be to optimize the interaction of
the amidine with the protein. Replacement of the
methylthio group with smaller groups was investigated
to see if the thiophene ring would slide deeper into the
S1 pocket bringing the amidine closer to Asp189 to
form a better salt bridge. However, unsubstituted thio-
2.5
2.5

494
M. J. Rudolph et al. / Bioorg. Med. Chem. Lett. 12 (2002) 491–495
phene 6 lost significant potency. Likewise, 10 with the
smaller 5-methoxy group should allow formation of a
better salt bridge relative to parent 5-methylthio-sub-
stituted 3 but, in fact, 10 was 5-fold less potent than 3.
Modeling studies show that a perpendicular geometry
for the 5-position substituent fits better into the S1
pocket. The S–Me with a lower barrier to rotation out
of the plane of the ring can achieve the perpendicular
geometry more easily than the O–Me. Replacement of
the heteroatom in the 5-substitutent of 3 with a carbon
atom resulted in 5-ethyl-substituted thiophene 8. Inter-
estingly, this compound was nearly equipotent with 3,
but showed a 9-fold increase in water solubility (2.7
mM aqueous solubility for 8 vs 0.3 mM for 3). Com-
pound 7, with the smaller 5-methyl group was also
equipotent with 3 and showed a significant 30-fold
increase in solubility (9.0 mM aqueous solubility). It is
interesting to note that the 5-methyl version containing
3,4-dimethoxyphenyl 9 was now 1.5 times less potent
than the corresponding 5-methylthio analogue 4. This
may indicate that the phenyl rings in 7 and 9 may be
going offat different vectors due to subtle positioning
differences of the thiophene ring within the S1 pocket.
Mono- or dioxidation of the methylthio in the 5-
methylthiothiophene 5 was explored to attenuate the
pKa of the amidine and to potentially pickup additional
hydrogen-bonding interactions with nearby Ser195 at
the edge of the S1 pocket. However, 12 and 13 were less
potent inhibitors relative to 5, probably because these
groups are sterically too large. Substitution of the 5-
methylthio group with larger arylmethylthio groups in 11
and 14 also resulted in less potent inhibitors due to their
steric size.
5. Tanaka, A.; Mizuno, H.; Sukarai, M. International Patent
WO 98/11089 1998. Chem. Abstr. 1998, 128, 230238.
6. Geyer, A. G.; McClellan, W. J.; Rockaway, T. W.; Stewart,
K. D.; Weitzberg, M.; Wendt, M. D. International Patent WO
99/05096 1999. Chem. Abstr. 1999, 130, 153476.
7. Barber, C. G.; Fish, P. V.; Dickinson, R. P. International
Patent WO 99/20608 1999. Chem. Abstr. 1999, 130, 311705.
8. Bridges, A. J.; Lee, A.; Schwartz, C. E.; Towle, M. J.; Lit-
tlefield, B. A. Bioorg. Med. Chem. 1993, 1, 403.
9. Towle, M. J.; Lee, A.; Maduakor, E. C.; Schwartz, E.;
Bridges, A. J.; Littlefield, B. A. Cancer Res. 1993, 53, 2553.
10. Nienaber, V. L.; Davidson, D.; Edalji, R.; Giranda, V. L.;
Klinghofer, V.; Henkin, J.; Magdalinos, P.; Mantei, R.; Mer-
rick, S.; Severin, J. M.; Smith, R. A.; Stewart, K.; Walter, K.;
Wang, J.; Wendt, M.; Weitzberg, M.; Zhao, X.; Rockway, T.
Structure 2000, 8, 553.
11. Hajduk, P. J.; Boyd, S.; Nettesheim, D.; Nienaber, V.;
Severin, J.; Smith, R.; Davidson, D.; Rockaway, T.; Fesik,
S. W. J. Med. Chem. 2000, 43, 3862.
12. This proximal pocket is defined by a disulfide linkage at
the back between Cys191 and Cys220 and has been referred to
as the S1b site by the authors of refs 10 and 11.
13. Leung, D.; Abbenante, G.; Fairlie, D. P. J. Med. Chem.
2000, 43, 305.
14. Subasinghe, N.; Hoffman, J.; Rudolph, J.; Soll, R.; Wil-
son, K.; Randle, T.; Green, D.; Lewandowski, F.; Zhang, M.;
Bone, R.; Spurlino, J.; Deckman, I.; Manthey, C.; Zhou, Z.;
Sharp, C.; Kratz, D.; Grasberger, B.; DesJarlais, R.; Molloy,
C.; Illig, C. Structure-Based Drug Design of Novel Urokinase
Plasminogen Activator Inhibitors, 1999 American Association
of Cancer Research/National Cancer Institute/European
Organization for Research and Treatment of Cancer Interna-
tional Conference, Washington, DC, Nov 16–19, 1999; Amer-
ican Association of Cancer Research: Philadelphia, PA, 1999;
Section 5, Drug Design, Metabolism and Pharmacogenetics
556.
15. Rudolph, M. J.; Subasinghe, N.; Illig, C. R.; Wilson, K.
J.; Hoffman, J. B.; Randall, T.; Green, D.; Molloy, C. J.; Soll,
R.; Lewandowski, F.; Zhang, M.; Bone, R.; Spurlino, J.;
Deckman, I.; Manthey, C.; Zhou, Z.; Sharp, C.; Kratz, D.;
Grasberger, B.; DesJarlais, R. Structure-Based Design and
Synthesis of 4,5-Disubstituted-Thiophene-2-Amidines as Potent
Urokinase Inhibitors, 219th National Meeting of the American
Chemical Society, San Francisco, CA, March 26–30, 2000;
American Chemical Society: Washington, DC, 2000; MEDI
233.
To conclude, although the modifications explored in
this study did not result in the expected improvements
in potency, equipotent analogues 7 and 8 were found
that demonstrated markedly (up to 30-fold) improved
solubility. The ease with which useful substitutions can
be made at the 5-position of electron-deficient 5-
methylsulfonylthiophenes was also shown. Further
modifications to other regions of these compounds are
being pursued and will be disclosed in due course.
16. Illig, C. R.; Subasinghe, N. L.; Hoffman, J. B.; Wilson, K.
J.; Rudolph, M. J. World Patent Application PCT/US99/
02784 1999; Chem. Abstr. 511156.
17. Kraus, G. A.; Taschner, H. J. J. Org. Chem. 1980, 45,
1175.
Acknowledgements
18. Kleeman, H.-W.; Lang, H.-J.; Schwark, J.-R.; Weichert,
A.; Scholz, W.; Albus, U. European Patent 0676395 A2 1995;.
Chem. Abstr. 1995, 124, 86809.
19. Janda, M.; Valenta, M.; Holy, P. Coll. Czech. Chem.
Commun. 1974, 39, 959.
The authors are grateful to Dr. Norman Hubert for
many helpful discussions and Rolanda Reed for techni-
cal assistance.
20. Paventi, M.; Edward, J. T. Can. J. Chem. 1987, 65, 282.
21. Singh, S. P.; Sehgal, S. Indian J. Chem. 1988, 27B, 941.
22. Garigipati, R. S. Tetrahedron Lett. 1990, 31, 1969.
23. Kano, S.; Yuasa, Y.; Yokomatsu, T.; Shibuya, S. Hetero-
cycles 1983, 20, 2035.
24. Crayston, J. A.; Iraqui, A.; Mallon, P.; Walton, J. C. J.
Chem. Soc., Perkin Trans. 2 1993, 29, 1589.
25. Lancaster Synthesis: Windham, NH, USA.
References and Notes
1. Tkachuk, V.; Stepanova, V.; Little, P. J.; Bobik, A. Clin.
Exp. Pharmacol. Physiol. 1996, 23, 759.
2. Andreasen, P. A.; Kjøller, L.; Christensen, L.; Duffy, M.
Int. J. Cancer 1997, 72, 1.
26. Maybridge Chemical Co.: Cornwall, UK.
27. Stayanovich, F. M.; Fedorov, B. P. Khim. Getero. Soedin.
1967, 3, 650.
28. Ravikumar, K. S.; Begue, J.-P.; Bonnet-Delpon, D.
Tetrahedron Lett. 1998, 39, 3141.
3. Evans, C. P.; Elfman, F.; Parangi, S.; Conn, M.; Cunha,
G.; Shuman, M. A. Cancer Res. 1997, 57, 3594.
4. Crowley, C. W.; Cohen, R. C.; Lucas, B. K.; Liu, G.; Shu-
man, M. A.; Levinson, A. D. Proc. Natl. Acad. Sci. U.S.A.
1993, 90, 5021.

M. J. Rudolph et al. / Bioorg. Med. Chem. Lett. 12 (2002) 491–495
495
29. Human kidney cell urokinase (Sigma U5004) was pur-
chased from Sigma Chemical Company (St. Louis, MO,
USA). All assays were based on the ability of the test com-
pound to inhibit the enzyme catalyzed hydrolysis of a peptide
p-nitroanilide substrate. In a typical Ki determination, into
each well of a 96-well plate is pipetted 280 mL of N-CBz-Val-
Gly-Arg-p-nitroanilide solution, 10 mL of the test compound
solution, and the plate allowed to thermally equilibrate at
37 ꢀC in a Molecular Devices plate reader for >15 min.
Reactions were initiated by the addition of a 10 mL aliquot of
enzyme and the absorbance increase at 405 nm is recorded for
15 min. Data corresponding to less than 10% of the total
substrate hydrolysis were used in the calculations. The ratio of
rate of change in absorbance as a function of time for a sample
containing no test compound is divided by the velocity of a
sample containing test compound, and is plotted as a function
of test compound concentration. The data are fit into a linear
regression, and the value of the slope of the line calculated.
The inverse of the slope is the experimentally determined Ki
value.
30. Concentrated solutions (10 mM) of the inhibitors were
prepared, shaken and incubated at 37 ꢀC for 24 h, then filtered
through a 0.22 m pore size, nylon membrane, polypropylene
syringe filter (Phenomenex, Torrance, CA). Analysis of the
˚
solutions by HPLC was carried out using a SPER HTS, 60 A,
5 m, 50ꢂ4.6 mm column (Princeton Chromatography, Prince-
ton, NJ, USA), running a gradient of 5–100% acetonitrile in
0.1% TFA (aq) over 7.5 min, with UV detection at 215 nm.
The rank order of solubility was determined according to the
area under the curve and accounting for any dilution factor.