N-Aroyl-L-phenylalanine derivatives as VCAM/VLA-4 antagonists.

Article abstract of DOI:10.1016/S0960-894X(02)00387-6

A series of N-benzoyl-4-[(2,6-dichlorobenzoyl)amino]-L-phenylalanine derivatives was prepared in order to optimize the substitution on the N-benzoyl moiety for VCAM/VLA-4 antagonist activity. Disubstitution in the 2- and 6-positions is favored and a range

Full text of DOI:10.1016/S0960-894X(02)00387-6

Bioorganic & Medicinal Chemistry Letters 12 (2002) 2479–2482
N-Aroyl-L-Phenylalanine Derivatives as VCAM/VLA-4
Antagonists
Achyutharao Sidduri,* Jefferson W. Tilley, Jian Ping Lou, Li Chen, Gerry Kaplan,
FrankMennona, Robert Campbell, Robert Guthrie, Tai-Nan Huang, Karen Rowan,
Virginia Schwinge and Louis M. Renzetti
Roche Research Center, Hoffmann-La Roche Inc., 340 Kingsland Street, Nutley, NJ07110, USA
Received 19 February 2002; accepted 22 April 2002
Abstract—A series of N-benzoyl-4-[(2,6-dichlorobenzoyl)amino]-l-phenylalanine derivatives was prepared in order to optimize the
substitution on the N-benzoyl moiety for VCAM/VLA-4 antagonist activity. Disubstitution in the 2- and 6-positions is favored and
a range of small alkyl and halogen are tolerated. A model of the bioactive conformation of these compounds is proposed. # 2002
Elsevier Science Ltd. All rights reserved.
The integrin VLA-4 (a4b1) is expressed on a variety of
leucocytes including B-cells, T-cells, basophils and eosin-
ophils and is involved in the recruitment, activation and
survival of these cell types.1 Data supporting a role for
VLA-4 in a number of inflammatory diseases including
asthma, rheumatoid arthritis, inflammatory bowel dis-
eases, multiple sclerosis, atherosclerosis, and diabetes
have emerged and is summarized in recent reviews.2 The
variety of diseases potentially impacted has prompted
an intense search for effective inhibitors of the VCAM/
VLA-4 interaction.
(t1/2=15 min) prompting us to explore a library
approach to new N-acyl derivatives capable of con-
formationally mimicking the N-benzylpyroglutamyl
moiety of 1.
We have previously reported identification of the potent
phenylalanine derivative 1 and crystal structure data
supporting our proposal that N-benzylpyroglutamyl
phenylalanine derivatives such as 1 mimic essential
pharmacophoric features of a NMR derived three
dimensional model of the cyclic peptide 2.3 Key com-
mon elements include the proline ring and the car-
boxylic acid with the phenylalanine aromatic ring of 1
occupying the same region of space as the disulfide
moiety of 2.4,5
This workidentified the ortho-substituted N-benzoyl
analogue 3 and the N-[(1-phenyl)cycloalkyl]carbonyl
derivative
4 as promising candidates for further
exploration.6 Herein, we describe the details of our
structure–activity studies of N-benzoyl phenylalanines
related to 3 and the X-ray crystal structure of the aniline
Pharmacokinetic studies of 1 in mice employing an
ELISA based binding assay to determine blood levels
indicate that this compound is cleared extremely rapidly
*Corresponding author. Tel.: +1-973-235-7143; fax: +1-973-235-
7122; e-mail: achyutharao.sidduri@roche.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00387-6

2480
A. Sidduri et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2479–2482
28, an intermediate in the scale up of one of the more
potent members of this series, 8l. This structure indi-
cates that the molecule is able to adopt a compact, low
energy, conformation that closely mimics the key
pharmacophoric elements of the cyclic peptide 2.
compete for bound recombinant human VCAM in the
presence of serial dilutions of test compound. VLA-4
bound to VCAM-1 was detected by a complex of anti-
b1 antibody and HRP-conjugated anti-mouse IgG:
chromogenic substrate (K-Blue).8
The new analogues were prepared as shown in Scheme 1
from Boc-4-nitrophenylalanine, 5. ortho-Substituted
benzoic acid derivatives that were not commercially
available were prepared by conventional methods as
shown in Scheme 2. For example, the ortho-di-
substituted aryl bromides, iodides or triflates 9 were
carbonylated in the presence of carbon monoxide,
Pd(OAc)2 and dppp in acetonitrile–water to obtain the
corresponding acids 10. Other approaches involved the
treatment of the substituted 2-methoxyphenyloxazoline
derivative, 11 with an alkyl Grignard reagent followed
by hydrolysis of the oxazoline ring of 12 to give the
desired acid 13.7 The benzoic acids required for the
preparation of compounds 18 and 19 were prepared as
shown in Scheme 3; (full experimental details are
available in ref 8).
They were further evaluated in a cell based assay for
their ability to blockthe interaction between fluoro-
scently labeled Ramos cells, which express VLA-4, with
VCAM coated microtiter plates. The ELISA assay does
not discriminate well among compounds with
IC50s<1.5 nM, we rely on the more stringent cell based
Compounds were assayed for VLA-4 antagonist activity
using a solid-phase, dual antibody ELISA in which
VLA-4 derived from Ramos cells was allowed to
Scheme 3. (a) 3-Hydroxybenzylamine, HBTU, DIPEA, DMF; (b)
0.5 N NaOH, H2O; (c) NaNO2, concd HCl; (d) SO2, HOAc, CuCl2; (e)
3-Acetoxybenzylamine, NEt3, CH2Cl2; (f) LiOH, 1:1 MeOH/THF.
Table 1. VCAM/VLA-4 binding inhibition of N-aroyl-4-[((2,6-
dichlorophenyl)carbonyl)amino]-l-phenylalanine derivatives
Compd
R
R1
ELISA IC50 (nM) Ramos IC50 (nM)
8a
8b
8c
8d
8e
8f
8g
3
8h
8i
Br
OCH3
NHCH3
CF3
SCH3
SO2CH3
C(CH3)3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
i-C3H7
F
H
H
H
H
H
0.49
8.6
1.9
16
0.40
3.3
1.9
1.3
0.30
0.42
1.2
0.44
0.30
0.20
45
6800
1000
650
100
160
74
84
10
17
23
85
12
9
460
67
85
49
1300
1600
460
450
19
33
350
9.5
1300
Scheme 1. (a) MeI, NaHCO3, DMF, rt, 15 h, 98%; (b) Zn dust,
NH4Cl, MeOH, H2O, 50 ꢀC, 2 h, 98%; (c) 2,6-dichlorobenzoyl chlo-
ride, DIPEA, CH2Cl2, rt, 15 h, 99%; (d) 4.0 N HCl in dioxane, rt, 1 h,
97%; (e) aromatic acid chloride, DIPEA, CH2Cl2, rt, 15 h or aromatic
acid, HBTU, DIPEA, DMF, rt, 15–48 h; (f) 1.0 N NaOH, EtOH,
45–50 ꢀC, 4–15 h.
H
H
6-CH3
6-C2H5
6-n-C3H7
6-i-C3H7
6-F
6-Cl
6-Br
6-i-C3H7
6-F
6-CF3
6-C(O)CH3
3-CH3
3-Cl
5-CH3
5-CF3
5-OCH3
4-OH
4-OCH3
4-SO2CH3
4-Br
8j
8k
8l
8m
8n
8o
8p
8q
8r
8s
8t
8u
8v
8w
8x
8y
8z
11
1.2
0.40
1.6
1.4
2.2
F
Cl
Cl
Cl
Cl
Cl
Br
Cl
11
8.8
0.25
0.46
1.9
0.25
1.60
Cl
Cl
Cl
Scheme 2. (a) Pd(OAc)2, dppp, CO, CH3CN, H2O, NEt3, 83 ꢀC, 15 h;
(b) R1MgCl, THF, À45 ꢀC to rt, 15 h; (c) MeI, rt, 15 h then NaOH,
MeOH, reflux, 15 h.

A. Sidduri et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2479–2482
2481
assay using Ramos cells to rankorder potency. Of the
mono ortho-substituted analogues, only the 2-bromo-,
2-methylthio- and 2-tertbutyl- derivatives, 8a, 8e and 8g
were potent in both assays. In general, the best activity
was seen among the 2,6-disubstituted analogues 3 and
8h–8q. Among these, there is a wide tolerance for sub-
stitution by alkyl, from methyl to isopropyl and halo-
gen, from fluorine to bromine. Only the 2,6-di-isopropyl
derivative 8n seems to exceed the favorable steric
limitations of the binding site.
inhibitors whereas the 2-chloro-4-methoxy- (8x) and
2-chloro-4-bromo- (8z) analogues are comparatively
weakinhibitors, it is unlikely that the differences in
potency are due to either steric or electronic effects. We
thus conclude that in the bound conformation, the
4-position is unlikely to be in close contact with VLA-4
and that the differences we observe in potency may be
due to solvation effects (Table 1).
2,3-Disubstitution was unfavorable as evidenced by 8r
and 8s whereas the activity of a limited number of 2,5-
disubstituted examples is similar to that expected for the
corresponding mono-ortho substituted analogues, sug-
gesting that the substituents in the 5-position contribute
little to binding.
The role of groups in the 4-position is somewhat enig-
matic, particularly in light of the potent activity asso-
ciated with 18 (IC50s ELISA 0.50 nM, Ramos 110 nM)
and 19 (IC50s ELISA 0.53 nM, Ramos 76 nM), pre-
pared as part of our LFA-ICAM inhibition program.9
Given that the 2-chloro-4-hydroxy- (8w) and 2-chloro-4-
methyl sulfonyl- (8y) derivatives are also potent
We also were interested in exploring ortho-substituted
heteroaromatic rings in this series. The compounds in
Table 2 were prepared from commercially available
heterocycles using the general procedures indicated in
Scheme 1. Whereas the 3-bromothiophen-2-yl analogue
20 was considerably less potent than the corresponding
phenyl derivative 8a, the 2,4-dimethyl-3-pyridinyl-
compound 23 was slightly more potent than the corre-
sponding phenyl derivative 3. Substitution of the 4-methyl
in 23 with a trifluoromethyl group (24) led to a decrease in
activity. Of the other compounds in Table 2, only the
thiazole 21 had good potency in both assay systems.
Table 2. VCAM/VLA-4 binding inhibition of N-heteroaroyl-4-[((2,6-
dichlorophenyl)carbonyl)amino]-l-phenylalanine derivatives
As stated in the introduction, we hypothesize that acyl
and aroyl phenylalanines related to 1, 3 and 4 are cap-
able of conformationally mimicking the cyclic peptidic
VCAM-VLA-4 inhibitor 2. In further support of this
notion, we obtained a crystal structure of the inter-
mediate 28, prepared during a scale up of 8l. There are
two molecules per unit cell which differ only by a 180ꢀ
rotation of the 2-chloro-6-methylaryl ring about the
amide bond. One of these is shown in Figure 1. As
expected for a 2,6-disubstituted aroyl species, the plane
of the aromatic ring is nearly orthogonal to that of the
amide bond (87ꢀ). The X-ray structure reveals a
compact, gauche (À) conformation of the phenylalanine
with one of the ortho-substituents making a hydrophobic
Compd
R
ELISA IC50 (nM)
1.5
Ramos IC50 (nM)
890
20
21
22
23
0.56
2.0
69
180
23
0.89
24
—
200
25
26
27
2.0
25
190
1440
580
Figure 1. (a) Space filling representation of one conformer of the X-
ray crystal structure of 28. The second conformer differs only by a
180ꢀ rotation of the disubstituted aromatic ring with respect to the
amide bond; (b) overlay of two NMR derived structures of the cyclic
core of peptide 2 with 28.
15

2482
A. Sidduri et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2479–2482
contact with the phenylalanine aromatic ring.10 An over-
lay of one of these structures with the NMR derived con-
formational model of the cyclic peptide 2 is also shown in
Figure 1 and is fully consistent with the concept that both
classes of inhibitor bind to VLA-4 in a similar manner.
2. (a) Tilley, J. W.; Sidduri, A. Drugs Future 2001, 26, 985. (b)
Porter, J. R. Idrugs 2000, 3, 788. (c) Adams, S. P.; Lobb, R. R.
Prog. Respir. Res., Basel, Karger 2001, 31, 302. (d) Boer, J.;
Gottschling, D.; Schuster, A.; Semmrich, M.; Holzmann, B.;
Kessler, H. J. Med. Chem. 2001, 44, 2586.
3. Fotouhi, N.; Joshi, P.; Fry, D.; Cook, C.; Tilley, J. W.;
Kaplan, G.; Hanglow, A.; Rowan, R.; Schwinge, V.;
Wolitzky, B. Bioog. Med. Chem. Lett. 2000, 10, 1171.
4. Chen, L.; Tilley, J. W.; Guthrie, R. W.; Mennona, F.;
Huang, T.-N.; Kaplan, G.; Trilles, R.; Miklowski, D.; Huby,
N.; Schwinge, V.; Wolitzky, B.; Rowan, K. Bioorg. Med.
Chem. Lett. 2000, 10, 729.
5. Chen, L.; Tilley, J. W.; Trilles, R. V.; Yun, W.; Fry, D.;
Cook, C.; Rowen, K.; Schwinge, V.; Campbell, R. Bioorg.
Med. Chem. Lett. 2002, 12, 137.
6. Chen, L.; Trilles, R.; Huang, T.-N.; Miklowski, D.; Huang,
T.-N.; Campbell, R.; Rowen, K.; Schwinge, V.; Tilley, J. W.
Bioorg. Med. Chem. Lett. 2002, 12, 1679.
7. Meyers, A. I.; Gabel, R. M.; Mihelick, E. D. J. Org. Chem.
1978, 43, 1372.
8. Chen, L.; Guthrie, R. W.; Huang, T.-N.; Hull, K. G.; Sid-
duri, A.; Tilley, J. W. WO 9910312. Chem. Abstr. 1999, 130,
196.
9. Burdick, D. J.; McDowell, R. S.; Stanley, M. S.; Marsters,
J. C.; Paris, K. J.; Oare, D. A.; Reynolds, M. E.; Ladner, C.;
Lee, W. P.; Gribling, P.; Dennis, M. S.; Skelton, N. J.; Tumas,
D. B.; Clark, K. R.; Keating, S. M.; Beresini, M. H.; Tilley, J.
W.; Presta, L. G.; Bodary, S. C.; Gadek, T. R. Science 2002,
295, 1086.
10. X-ray report: compound 28 was crystallized from ethyl
acetate, mp: 100–102 ꢀC, crystal size (mm): 0.17Â0.22Â0.86,
space group: P212121; Rw: 0.0604. The atomic coordinates
were provided to CCD with deposition number CCDC185241.
11. Sidduri, A.; Tilley, J. W.; Hull, K.; Lou, J. P.; Kaplan, G.;
Sheffron, A.; Chen, L.; Campbell, R.; Guthrie, R.; Huang, T.-N.;
Huby, N.; Rowan, K.; Schwinge, V.; Renzetti, L. M. Bioorg.
Med. Chem. Lett. 2002, 12, 2475.
In the present work, 2,6-disubstitution is generally pre-
ferred although certain 2,4-disubstituted analogues are
also highly potent. The requirement for at least one
ortho-substituent suggests that a conformation in which
the aromatic ring is rotated out of the plane of the
amide bond is preferred. The lackof electronic effects in
the present series and the comparable activity of
cycloalkyl derivatives related to 411 suggest that one of
the groups in the ortho position makes a hydrophobic
contact with VLA-4 and that this contact is important
for efficient binding. Future papers in this series will
document the pharmacokinetics and the in vivo phar-
macology of members of this series.
References and Notes
1. Elices, M. J. In Cell Adhes. Mol. Matrix Proteins; Mouse,
S. A., Ed.; Springer: Berlin, 1999, p 133.