Vitronectin receptor antagonists: Purine-based peptidomimetics

Full text of DOI:10.1002/1521-3773(20000818)39:16<2874::AID-ANIE2874>3.0.CO;2-C

COMMUNICATIONS
Vitronectin Receptor Antagonists:
Purine-Based Peptidomimetics
Anusch Peyman,* Jean-François Gourvest,
Thomas R. Gadek, and Jochen Knolle
Integrins are a widely expressed family of a/b heterodi-
meric cell surface receptors which bind to extracellular matrix
adhesive proteins such as fibrinogen, fibronectin, vitronectin,
laminin, and osteopontin.[1±3] They are composed of dimers of
at least fifteen a subunits and eight b subunits. The b3 class of
the integrin family, aIIbb3 (GPIIb/IIIa or fibrinogen receptor)
and aVb3 (vitronectin receptor), has received special attention
[4, 5]
in recent drug research.
a b3 is prevalent on platelets and
IIb
plays a role in thromboembolic disorders,[5] while aVb3 is the
dominant receptor for mediating the attachment of osteo-
clasts to bone during bone resorption[6, 7] and has been
implicated in tumor progression, angiogenesis,[5, 8, 9] and
restenosis.[10] Many integrins, including aIIbb3 and aVb3,
interact with a common Arg-Gly-Asp (RGD) binding motif
in their target proteins.[11±13] Kessler et al. were able to identify
the structural properties required for the selective inhibition
of either aIIbb3 or aVb3 using stereoisomeric cyclic peptide
libraries.[14±16] In drug development peptidomimetic com-
pounds are often preferred over peptides[17], and several
selective peptidomimetic antagonists of aVb3 have been
reported recently.[18±25] This has led to an increased interest
in finding new building blocks for integrin antagonists that
allow the correct spatial arrangement of pharmacophorically
important groups. Although the purine scaffold has been used
in the synthesis of a variety of chemical libraries, especially in
the design of kinase inhibitors,[26, 27] there are to our knowl-
edge no reports on its use as a central scaffold in peptidomi-
metics. Here we introduce the purine scaffold in the design
and synthesis of selective antagonists of aVb3 and present the
first structure ± activity relationship within this series.
The synthesis of the aVb3 antagonists 5 and 6 is outlined in
Scheme 1. Alkylation of 6-chloropurine 1 with N-Cbz-l-serine
tert-butyl ester under Mitsunobu conditions afforded 2 in
50% yield. Nucleophilic substitution with building blocks
containing one tert-butyloxycarbonyl (Boc) protected and one
unprotected amino group gave aminopurines 3. Only (4-
piperidylmethyl)amine was used without Boc-protection,
reacting selectively (>95%) with the secondary amine. (No
trace of the corresponding product of the primary amine was
found.) The Boc group and the tert-butyl ester were cleaved
simultaneously by treatment with 95% TFA/5% water and
Scheme 1. Synthesis of the aVb3 antagonists 5 and 6. a) PPh3, DEAD, THF,
08C; 50%; b) DMF, DIPEA, 6 h 508C; 60 ± 80%; c) 95% TFA, 2h, RT;
100%; d) 1H-pyrazole-1-carboxamidine, H2O, DIPEA, 48 h, RT; 50 ± 85%;
e) 2-methylsulfanyl-2-imidazoline, DIPEA, 24 h, 508C; 50 ± 70%. DEAD
ˆ diethyl azodicarboxylate, TFA ˆ trifluoroacetic acid, DIPEA ˆ diiso-
propylethylamine.
the resulting molecules 4 were converted into the guanidines 5
by treatment with 1H-pyrazole-1-carboxamidine[28] and DI-
PEA in water with yields ranging from 50 to 85%. Alter-
natively 4 was converted to the (1H-imidazolin-2-yl)amines 6
using 2-methylsulfanyl-2-imidazoline and DIPEA in DMF
with yields of 50 to 70%. The synthesis of antagonist 7, which
contains a 2-aminobenzimidazole group as an arginine
mimetic, is depicted in Scheme 2. N-(1H-benzimidazol-2-
yl)butane-1,4-diamine, prepared by standard methods, was
treated with 2 to give antagonist 7 after cleavage of the tert-
butyl ester. A second series of aVb3 antagonists (10 and 11) is
characterized by a ªreversedº orientation of the central
purine scaffold (amino acid substituent not on the imidazole
but on the pyrimidine ring), and its synthesis is shown in
Scheme 3. Compound 1 was first alkylated with N-Boc-
protected aminoalkyl tosylates to give the chloropurines 8 in
high yields. These compounds reacted with the amino group
of tert-butyl (S)-N2-Cbz-2,3-diaminopropionate and its homo-
logues, and after treatment with TFA, compounds 9 are
obtained. Subsequent conversion to the guanidines 10 and the
[*] Dr. A. Peyman
Aventis Pharma, Medicinal Chemistry G838
65926 Frankfurt (Germany)
Fax : (‡49)69-305-80679
E-mail: anusch.peyman@aventis.com
Dr. J.-F. Gourvest
Â
Aventis Pharma, 102 Route de Noisy, 93235 Romainville Cedex
(France)
Dr. T. R. Gadek
Genentech, 1 DNA Way, South San Francisco, CA 94080 (USA)
Dr. J. Knolle
Axys Inc., 385 Oyster Point Blvd., South San Francisco, CA 94080
(USA)
2874
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1433-7851/00/3916-2874 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2000, 39, No. 16

COMMUNICATIONS
to the same recognition motif RGD and there seem to be only
subtle differences in the ligand structure that determine
selectivity. The IC50 values of compounds 5 ± 7, and 10, 11
(Scheme 4) for the inhibition of binding of fibrinogen to aIIbb3
and of Kistrin (a disintegrin with high affinity to aVb3[7, 29]) to
aVb3 are summarized in Tables 1 (ªnormalº orientation) and 2
(ªreversedº orientation).
Scheme 2. Synthesis of the aVb3 antagonist 7. a) DMF, 08C, 1 h; 95%;
b) H2/Pd-C; 100%; c) HgO, sulfur, 508C; 43%; d) 95% TFA, 2 h, 08C;
100%; e) DMF, DIPEA, 32 h, 508C; 32%; f) 95% TFA, 2 h, 08C; 100%.
Scheme 4. General formulas for the antagonists 5 ± 7, and 10, 11.
The most striking criterion for selectivity is the distance
between the carboxy group and the N-terminal guanidino
group, as can be seen from a comparison of 5a, 5b, and 5c or
in the ªreversedº series by comparing 10a, 10b, and 10c. The
optimum distance for aVb3 binding within these two series is
12 bonds (5b, 10b), while aIIbb3 prefers a substrate where the
two groups are further apart, as in 5a or 10a with a distance of
Table 1. Activity and selectivity of aVb3 antagonists 5 ± 7 (Scheme 4). The IC50
values [mm] denote the concentration required to reduce binding of fibrinogen
(Fg) to aIIbb3 or of Kistrin (K) to aVb3 by 50%.[30]
Configura- Guanidine
Spacer
IC50
IC50
tion ( )
*
K/VnR Fg/IIbIIIa
5a
5b
5c
S
S
S
1.1
0.7
2.2
1.9
> 10
> 10
> 10
6a
7
S
S
0.15
0.16
> 10
Scheme 3. Synthesis of the aVb3 antagonists 10 and 11. a) DMF, K2CO3,
RT; 70 ± 80%; b) DMF/DIPEA (2/1), 708C, 7d; 40 ± 50%; c) 95% TFA,
2 h, 08C; 100%; d) 1H-pyrazol-1-carboxamidine, H2O, DIPEA, 48 h, RT;
50 ± 60%; e) 2-methylsulfanyl-2-imidazoline, DIPEA, 24 h, 508C; 50 ±
60%.
6b
S
0.05
4.0
6c
6d
R
R
0.34
1.0
> 10
> 10
amino-2-imidazolines 11 was accomplished with the same
methods as for the synthesis of 5 and 6.
One of the most interesting challenges in this area is the
design of selective antagonists since both aVb3 and aIIbb3 bind
Angew. Chem. Int. Ed. 2000, 39, No. 16
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1433-7851/00/3916-2875 $ 17.50+.50/0
2875

COMMUNICATIONS
Table 2. Activity and selectivity of aVb3 antagonists 10 and 11 (Scheme 4).
The IC50 values [mm] denote the concentration required to reduce binding
of fibrinogen (Fg) to aIIbb3 or of Kistrin (K) to aVb3 by 50%.[30]
the R configuration and which are approximately sevenfold
less active than the S enantiomers 6b and 6a, respectively.
Unexpectedly, the binding affinity is quite insensitive to the
orientation of the central purine scaffoldÐa comparison of,
for example, 6a (IC50 ˆ 150 nm) and 11a (IC50 ˆ 7 5 nm)
reveals only a small preference for the ªreversedº orientation.
The difference is of the same order of magnitude as the one
induced by changing the position of the purine scaffold within
a given orientation (ªslidomersº); compare, for example, 11a
(IC50 ˆ 7 5 nm) and 11c (IC50 ˆ 190 nm).
Guanidine
m
n
IC50
IC50
K/VnR
Fg/IIbIIIa
10a
1
5
0.65
0.6
10b
10c
10d
11a
1
1
2
1
4
3
3
4
0.175.0
0.54
In conclusion, we have been able to show that the purine
scaffold is a highly versatile building block in the synthesis of
peptidomimetics. To mimic the RGD peptide recognition
sequence of the vitronection receptor, we could easily vary
spacer length, rigidity, and pharmacophoric groups. This led to
a very consistent structure ± activity relationship and resulted
in highly active and selective new antagonists of aVb3. The
ease of derivatization will certainly encourage many future
uses of this scaffold in peptidomimetic chemistry.
3.5
> 10
0.58
0.075
> 10
11b
11c
1
2
5
3
0.21
0.19
> 10
> 10
Received: February 3, 2000 [Z14641]
[1] E. Ruoslahti, M. D. Pierschbacher, Science 1987, 238, 491 ± 497.
[2] Integrins: Molecular and Biological Responses to the Extracellular
Matrix (Eds.: D. A. Cheresh, R. P. Mecham), Academic Press, San
Diego, 1994.
[3] R. O. Hynes, Cell 1992, 69, 11 ± 25.
[4] V. W. Engleman, M. S. Kellogg, T. E. Rogers, Annu. Rep. Med. Chem.
1996, 31, 191 ± 200.
13 bonds. This observation is in good agreement with the
results obtained for other inhibitors[18, 19] and with Kesslerꢁs
observation that in cyclopeptides aVb3-selective molecules are
strongly bent, while aIIbb3-selective molecules bind in a more
extended conformation.[14, 16]
A comparison of 5b, 6a, and 7 illustrates that cyclic
guanidines have a higher affinity for aVb3 than noncyclic ones.
The cyclic guanidino antagonist 6a (IC50 ˆ 150 nm) exhibits a
fivefold affinity over the open guanidino compound 5b
[5] S. A.Mousa, D. A. Cheresh, Drug Discovery Today 1997, 2, 187± 199.
[6] M. H. Helfrich, S. A. Nesbitt, P. T. Lakkakorpi, M. J. Barnes, S. C.
Bodary, G. Shankar, W. T. Mason, D. L. Mendrick, H. K. Vaananen,
M. A. Horton, Bone 1996, 19, 317± 328.
[7] S. B. Rodan, G. A. Rodan, J. Endocrinol. 1997, 154, S47± S56.
[8] P. C. Brooks, Drug News Perspect. 1997, 10, 456 ± 461.
[9] A. Giannis, Biomed. Health Res. 1999, 22, 81 ± 89.
[10] J. Samanen, A. Jonak, D. Rieman, T.-L. Yue, Curr. Pharm. Des. 1997,
3, 545 ± 584.
[11] K. M. Yamada, J. Biol. Chem. 1991, 266, 12809 ± 12812.
[12] E. Ruoslahti, Annu. Rev. Cell Dev. Biol. 1996, 12, 697± 715.
[13] S. E. DꢁSouza, M. H. Ginsberg, E. F. Plow, Trends Biochem. Sci. 1991,
16, 246 ± 250.
[14] R. Haubner, D. Finsinger, H. Kessler, Angew. Chem. 1997, 109, 1440 ±
1456; Angew. Chem. Int. Ed. Engl. 1997, 36, 1374 ± 1389.
[15] M. Pfaff, K. Tangemann, B. Müller, M. Gurrath, G. Müller, H. Kessler,
R. Timpl, J. Engel, J. Biol. Chem. 1994, 269, 20233 ± 20238.
[16] M. A. Dechantsreiter, E. Planker, B. Mathae, E. Lohof, G. Hölze-
mann, A. Jonczyk, S. L. Goodman, H. Kessler, J. Med. Chem. 1999, 42,
3033 ± 3040.
(IC50 ˆ 700 nm). The 2-aminobenzimidazole antagonist
7
shows approximately the same potency as 6a. Interestingly
this increase in affinity is not observed for aIIbb3 binding. A
similar trend can be observed in the ªreversedº series;
however, the increase in affinity from 10b (IC50 ˆ 170 nm) to
the cyclic guanidino compound 11a (IC50 ˆ 7 5 nm) is only by a
factor of 2, but at the same time there is a decrease in affinity
to aIIbb3. This decreased affinity to aIIbb3 for cyclic guanidines
can also be observed in a comparison of the extended
antagonists 10a and 11b. One could speculate that the
antagonists bind to aIIbb3 through an ªend-onº interaction,
while aVb3-binding is achieved through ªside-onº binding on
the guanidine.
[17] A. Giannis, F. Rubsam, Adv. Drug Res. 1997, 29, 1 ± 7 8.
[18] A. L. Rockwell, M. Rafalski, W. J. Pitts, D. G. Batt, J. J. Petraitis, W. F.
DeGrado, S. Mousa, P. K. Jadhav, Bioorg. Med. Chem. Lett. 1999, 9,
937± 942.
Another enhancement in aVb3 binding is obtained through
the rigidification of the spacer from aminobutane in 6a
(IC50 ˆ 150 nm) to methylpiperidine in 6b (IC50 ˆ 50 nm),
keeping the length of the spacer constant. The piperidine
spacer seems to bring the guanidino group into the correct
spatial orientation and, in addition, it has the advantage that
no additional stereocenter is introduced.
The S configuration is clearly preferred for the only
stereocenter that is connected with the lipophilic side chain,
the benzyloxycarbonylamino group, as can be seen from the
IC50 values for 6c (340 nm) and 6d (1000 nm) which both have
[19] K. C. Nicolaou, J. I. Trujillo, B. Jandeleit, K. Chibale, M. Rosenfeld, B.
Diefenbach, D. A. Cheresh, S. L. Goodman, Bioorg. Med. Chem.
1998, 6, 1185 ± 1208.
[20] R. M. Keenan, M. A. Lago, W. H. Miller, F. E. Ali, R. D. Cousins,
L. B. Hall, S.-M. Hwang, D. R. Jakas, C. Kwon, C. Louden, T. T.
Nguyen, E. H. Ohlstein, D. J. Rieman, S. T. Ross, J. M. Samanen, B. R.
Smith, J. Stadel, D. T. Takata, L. Vickery, C. C. K. Yuan, T.-L. Yue,
Bioorg. Med. Chem. Lett. 1998, 8, 3171 ± 3176.
[21] R. M. Keenan, W. H. Miller, M. A. Lago, F. E. Ali, W. E. Bondinell,
J. F. Callahan, R. R. Calvo, R. D. Cousins, S.-M. Hwang, D. R. Jakas,
T. W. Ku, C. Kwon, T. T. Nguyen, V. A. Reader, D. J. Rieman, S. T.
Ross, D. T. Takata, I. N. Uzinskas, C. C. K. Yuan, B. R. Smith, Bioorg.
Med. Chem. Lett. 1998, 8, 3165 ± 3170.
2876
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1433-7851/00/3916-2876 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2000, 39, No. 16

COMMUNICATIONS
acylation catalysts,[2g,h, 3] we have established that they are so
mild that various selective acylation reactions are feasible but
their acidity is not strong enough to perform acylation of
sterically hindered alcohols. In this context, we were intrigued
to employ bismuth triflate Bi(OTf)3[4] since it had proved to be
easy to handle due to its stability in the air and to be acidic
enough to catalyze Friedel ± Crafts,[5] Diels ± Alder,[6] and ene
reactions.[7]
As shown in Table 1, Bi(OTf)3 can promote the acetylation
of primary, secondary, and tertiary alcohols in acetic anhy-
dride at 258C [Eq. (1)]. For the acetylation of 2-phenyl-
ethanol a catalyst loading as low as 0.01 mol% was sufficient
to afford the acetate quantitatively within 10 min (entry 1).
[22] R. M. Keenan, W. H. Miller, C. Kwon, F. E. Ali, J. F. Callahan, R. R.
Calvo, S.-M. Hwang, K. D. Kopple, C. E. Peishoff, J. M. Samanen,
A. S. Wong, C.-K. Yuan, W. F. Huffman, J. Med. Chem. 1997, 40,
2289 ± 2292.
[23] W. J. Hoekstra, B. L. Poulter, Curr. Med. Chem. 1998, 5, 195 ± 204.
[24] W. H. Miller, W. E. Bondinell, R. D. Cousins, K. F. Erhard, D. R.
Jakas, R. M. Keenan, T. W. Ku, K. A. Newlander, S. T. Ross, R. C.
Haltiwanger, J. Bradbeer, F. H. Drake, M. Gowen, S. J. Hoffman, S.-
M. Hwang, I. E. James, M. W. Lark, B. Lechowska, D. J. Rieman, G. B.
Stroup, J. A. Vasko-Moser, D. L. Zembryki, L. M. Azzarano, P. C.
Adams, K. L. Salyers, B. R. Smith, K. W. Ward, K. O. Johanson, W. F.
Huffman, Bioorg. Med. Chem. Lett. 1999, 9, 1807± 1812.
[25] R. M. Keenan, W. H. Miller, L. S. Barton, W. E. Bondinell, R. D.
Cousins, D. F. Eppley, S. M. Hwang, C. Kwon, M. A. Lago, T. T.
Nguyen, B. R. Smith, I. N. Uzinskas, C. C. K. Yuan, Bioorg. Med.
Chem. Lett. 1999, 9, 1801 ± 1806.
[26] Y.-T. Chang, N. S. Gray, G. R. Rosania, D. P. Sutherlin, S. Kwon, T. C.
Norman, R. Sarohia, M. Leost, L. Meijer, P. G. Schultz, Chem. Biol.
1999, 6, 361 ± 375.
[27] N. S. Gray, L. Wodicka, A.-M. W. H. Thunnissen, T. C. Norman, S.
Kwon, F. H. Espinoza, D. O. Morgan, G. Barnes, S. LeClerc, L. Meijer,
S.-H. Kim, D. J. Lockhart, P. G. Schultz, Science 1998, 281, 533 ± 538.
[28] M. S. Bernatowicz, Y. Wu, G. R. Matsueda, J. Org. Chem. 1992, 57,
2497± 2502.
Table 1. Bi(OTf)3-catalyzed acetylation of alcohols [Eq. (1)].[a]
Bi(OTf)3
+
(1)
R
OH
Ac2O
R OAc
Entry Alcohol
Mol% of the catalyst Time [h] Yield [%][b]
1
2
3[c]
4
2-phenylethanol 0.01
2-phenylethanol 0.005
2-phenylethanol 0.5
1-phenylethanol 0.005
0.16798
[29] L. J. Green, V. H. Brierley, M. J. Humphries, Biochem. Soc. Trans.
1995, 23, 505S.
2
4
2
99
92
95
[30] Kistrin, a snake venom protein, is a well known antagonist of aVb3.
We and others found the binding assay to be more stringent with
kistrin than with vitronectin: a) D. R. Clemmons, G. Horvitz, W.
Engleman, T. Nichols, A. Moralez, G. A. Nickols, Endocrinology 1999,
140, 4616 ± 4621; b) K. L. King, J. J. DꢁAnza, S. Bodary, R. Pitti, M.
Siegel, R. A. Lazarus, M. S. Dennis, R. G. Hammonds, Jr, S. C.
Kukreja, J. Bone Miner. Res. 1994, 9, 381 ± 387.
5
6[d]
2-octanol
2-octanol
0.005
0.005
1798
3
98
71-adamantanol
0.01
6
98
[a] Reaction conditions: alcohol (1.0 mmol), acetic anhydride (10 equiv),
258C. [b] Determined by GC. [c] Acetic anhydride (1.5 equiv), CH2Cl2
(wet, 1.0 mL). [d] At 408C.
Even with 0.005 mol% of Bi(OTf)3, this acetylation proceed-
ed quantitatively, but took longer to complete (entry 2). It is
possible to employ a solvent different from the anhydride as
well. For example, 2-phenylethanol in CH2Cl2 was acetylated
by 1.5 equivalents of Ac2O, where CH2Cl2 was used without
any purification (vide infra and entry 3). Secondary alcohols
such as 1-phenylethanol and 2-octanol were also smoothly
converted into the corresponding acetates in the presence of
0.005 mol% of Bi(OTf)3 (entries 4 and 5). Although 2-octanol
is less reactive than 1-phenylethanol, quantitative acetylation
was possible by using longer reaction times or higher reaction
temperatures (entries 5 and 6). Surprisingly, even for the
acetylation of 1-adamantanol a catalyst loading of 0.01 mol%
was enough (entry 7).
After these preliminary examinations, acetylation of func-
tionalized alcohols was scrutinized at 258C (Table 2). Al-
though acid-sensitive geraniol and furfuryl alcohol were
acetylated in 81% and 80% yields, respectively (entries 1
and 3), further improvements of the yields were attained by
adding a donor solvent, such as acetonitrile or THF (entries 2
and 4). Not only primary but also secondary and tertiary
alcohols with functional groups or phenol units underwent
smooth acetylation (entries 5 ± 15). Alcohols having an ester
or amine function did not react more slowly or show
decomposition of the functional groups. More importantly,
neither racemization nor epimerization was detected (en-
tries 7± 9).
Highly Efficient and Versatile Acylation of
Alcohols with Bi(OTf)3 as Catalyst
Akihiro Orita, Chiaki Tanahashi, Atsushi Kakuda, and
Junzo Otera*
The acylation of alcohols is an important transformation in
organic synthesis.[1] Despite a number of precedents, new
efficient methods are still in strong demand. Acid anhydrides
are the most commonly used reagents in the presence of an
acid or base catalyst[2] and the utility of this protocol was
boosted by the discovery of the dimethylaminopyridine
(DMAP) catalyst.[2a] More recently, metal triflates, such as
scandium triflate,[2d] trimethylsilyl triflate,[2e] and indium
triflate,[2f] were found to be effective as well. These catalysts
are very useful, but also suffer from some drawbacks.
Scandium triflate is rather expensive and must be used under
anhydrous conditions. Trimethylsilyl triflate is labile towards
moisture, and its acidity is too strong for acid-sensitive
alcohols as reagents.[2e] In the development of organotin
[*] Prof. Dr. J. Otera, Dr. A. Orita, C. Tanahashi, A. Kakuda
Department of Applied Chemistry, Okayama University of Science
Ridai-Cho, Okayama 700-0005 (Japan)
Notably, tertiary alcohols can be acetylated at ambient
temperatures (entries 10 ± 14). For example, the Bi(OTf)3-
catalyzed acetylation of an a,a-dimethyl-substituted tertiary
Fax : (‡81)86-256-4292
E-mail: otera@high.ous.ac.jp
Angew. Chem. Int. Ed. 2000, 39, No. 16
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1433-7851/00/3916-2877 $ 17.50+.50/0
2877