Dehydrophenylalanine derivatives as VLA-4 integrin antagonists

Article abstract of DOI:10.1016/S0960-894X(03)00019-2

We describe a series of dehydrophenylalanine derivatives where the Z isomers are potent VLA-4 antagonists but are subject to rapid biliary clearance and the E isomers have poor activity but have a slower rate of clearance. These configurationally constrained molecules have led to the design of a novel class of benzodiazepine VLA-4 antagonists.

Full text of DOI:10.1016/S0960-894X(03)00019-2

Bioorganic & Medicinal Chemistry Letters 13 (2003) 805–808
Dehydrophenylalanine Derivatives as VLA-4 Integrin Antagonists
John R. Porter,* Sarah C. Archibald, Julien A. Brown, Kirstie Childs, David Critchley,
John C. Head, Ted A. H. Parton, Martyn K. Robinson, Anthony Shock,
Richard J. Taylor and Graham J. Warrellow
Department of Medicinal Chemistry, Celltech R&D Ltd, 216 Bath Road, Slough SL1 4EN, UK
Received 10 October 2002; revised 3 January 2003; accepted 7 January 2003
Abstract—We describe a series of dehydrophenylalanine derivatives where the Z isomers are potent VLA-4 antagonists but are
subject to rapid biliary clearance and the E isomers have poor activity but have a slower rate of clearance. These configurationally
constrained molecules have led to the design of a novel class of benzodiazepine VLA-4 antagonists.
# 2003 Elsevier Science Ltd. All rights reserved.
VLA-4 (a4b1 or CD49d/CD29) is a member of the
integrin family1 of glycoprotein cell surface receptors
whose principal function is to mediate cell-cell and cell-
matrix interactions. VLA-4 recognises a number of
ligands2 including vascular cell adhesion molecule
(VCAM),3 expressed on endothelial cells in response to
pro-inflammatory cytokines (such as IL-1, IL-4 and
TNF-a). VLA-4 is present on most leukocytes that play
an essential role in chronic diseases of autoimmune ori-
gin, where it is believed to be involved in the adhesion,
migration and activation of these cell-types at sites of
inflammation.4 Consequently, blocking the action of
VLA-4 would be expected to be of therapeutic benefit in
the treatment of a variety of inflammatory diseases.
Anti-a4 antibodies have shown efficacy in a number of
animal models of chronic inflammatory diseases includ-
ing asthma, rheumatoid arthritis, and multiple sclerosis.5
A humanised monoclonal antibody has shown activity in
Phase II clinical trials for multiple sclerosis and ulcera-
tive colitis,6 and a number of small molecule antagonists
are believed to be in pre-clinical development.7
macokinetic profile in the rat. Further work on similar
phenylalanine based antagonists revealed that the
observed rapid biliary clearance appeared to be a char-
acteristic of these structures, a feature that has also been
reported by a number of other workers in this field.9
In an attempt to gain an understanding of the SAR for
the clearance we set about screening a library of phenyl-
alanine analogues to identify compounds that displayed
acceptable rates of clearance. In order to get a relatively
high throughput screen we chose the isolated perfused
rat liver (IPRL) assay,10 whereby five compounds
(including a reference compound) could be dosed as a
cassette. The elimination of each compound from the
perfusate is expressed in terms of a rate constant, k, and
normalised to the reference compound. The higher the
value of k the more rapidly the compound was cleared.
We calculated that an IPRL k=1 hꢀ1 would equate to
an observed clearance rate of around 10 mL/min/kg in
the rat, a figure that we felt would be acceptable.
An initial library of over 50 compounds was screened in
this assay, only a few of which displayed an acceptably
slow rate of clearance. Of these, the dehydro-
phenylalanine analogues, 1 and 2, attracted our atten-
tion. We were intrigued to notice that although both
compounds displayed relatively poor potency, the
E-isomer, 2, had a significantly slower rate of clearance
when compared to the Z-isomer, 1. Encouraged by
We have recently described the discovery of N-acetyl
d-thiaproline-(O-2,6-dichlorobenzyl)-L-tyrosine, CT5219,
a potent VLA-4 integrin antagonist that showed efficacy
in an animal model of asthma.8 However, we felt unable
to develop this compound as it displayed a poor phar-
these results we prepared
a series of dehydro-
phenylalanine analogues to see if we could improve
their potency as determined by their ability to inhibit
*Corresponding author. Tel.: +44-1753-534655; fax: +44-1753-
536632; e-mail: john.porter@celltechgroup.com
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00019-2

806
J. R. Porter et al. / Bioorg. Med. Chem. Lett. 13 (2003) 805–808
the binding of VLA-4 to VCAM-1 in a protein-based,
ligand binding11 and a cell-based, adhesion assay.12 We
now wish to report our observations on this series of com-
pounds that also lead to some insights on the active con-
formation of these antagonists in the ligand-binding site.
by column chromatography then hydrolysed to the tar-
get acid 8 as before.
We have shown in earlier series8 that the potency of the
antagonists could be significantly enhanced by the
replacement of the benzyl ether with 2,6-dichlorobenzyl
ether or, more preferably, 2,6-dichlorobenzyl amide
analogues. Additionally we had shown that reversal of
the stereochemistry at the thiaproline a-centre gave a
further enhancement to the potency. Gratifyingly,
incorporation of these modifications into the dehydro-
phenylalanine series had a similar effect, Table 1. For
example, the 2,6- dichlorobenzyl amide, Z isomer, 7, is
500-fold more potent in the cell-based assay than the
benzyl ether, Z isomer, 1, whilst the 3,5-dichloropyridyl
amide, Z isomer, 9, is more potent still. Interestingly,
this enhancement in potency is not observed for the E
isomer analogues, where both the ether and amide ana-
logues display poor activity in the cell based assays.
(The apparent improvement in potency for the
The compounds were prepared in a similar fashion to
that depicted in Scheme 1 for the benzamides 7 and 8.
The commercially available phosphonate 3 was depro-
tected by hydrogenation and coupled to the N-acetyl-d-
thiaproline using standard amide coupling conditions to
give the phosphonate 4. Treatment of a solution of the
phosphonate 4 and the requisite benzaldehyde with
DBU gave the dehydroamino ester. NMR analysis of
the crude reaction mixture showed the presence of only
a single double bond isomer whilst nOe experiments13
on the purified products determined that this was the Z
isomer. (At temperatures below 300 K both isomers
exist as a 60:40 mixture of rotamers as a result of
restricted rotation about the thiaproline N-acetyl bond.)
Hydrolysis of the ester under basic conditions gave the
target acid 7. In order to obtain quantities of the E iso-
mer the phosphonate 4 was treated with sodium hydride
in THF, followed by the requisite benzaldehyde to fur-
nish approximately 1:1 mixtures of the E and Z dehydro
aminoesters as determined by NMR analysis of the
crude reaction mixture. The pure E isomer was isolated
Table 1. Potency and rate of clearance for thiaprolines
Compd
R
E/Z Thioproline VLA-4
protein
VLA-4 IPRL
k hꢀ1
cell
IC50 nM IC50 nM
2200
71,000
3.8
1
2
Z
E
l
l
1500
2800
7400
5
20,000
5000
83,000
140
0.6
4.1
0.4
3.9
0.7
4.4
1.2
5
6
Z
E
Z
E
Z
E
d
d
d
d
d
d
7
8
400
2
29,000
25
9
Scheme 1. (i) H2, 10% Pd/C, MeOH; (ii) N-Acetyl-d-thiaproline,
HOBT, EDC, N-methyl morpholine DCM; (iii) DBU, 2,6-dichloro-N-
(4-formylphenyl)benzamide, DCM; (iv) LiOH, THF, H2O; (v) LDA,
THF then 2,6-dichloro-N-(4-formylphenyl)benzamide, THF, DMF.
10
60
200

J. R. Porter et al. / Bioorg. Med. Chem. Lett. 13 (2003) 805–808
807
3,5-dichloropyridyl amide, E isomer, 10, is a result of
the presence of a small amount of the potent Z isomer
as an impurity.) What is perhaps more significant is that
the E isomers all display a markedly slower rate of
clearance than their Z isomer counterparts.
In order to demonstrate that this effect was independent
of the a-amide substituent a series of nicotinamides
were prepared by a similar route to that described in
Scheme 1. Although the data set is more limited a simi-
lar trend was observed, Table 2, with the more potent
3,5-dichloropyridyl amide Z isomer, 13, having a much
faster rate of clearance than the less potent E isomer 14.
A similar effect was observed with the t-butyl amides 15
and 16. Whereas the E isomer, 16, showed poor activity
(VLA-4/VCAMprotein IC
2.7 mM; VLA-4/VCAM
50
cell IC50 80 mM) the Z isomer, 15, was significantly
Scheme 2. (i) HCl, NaNO2, H2O; (ii) NaOAc, NaN3, H2O; (iii) MnO2,
DCM; (iv) N-(2-chloronicotinoyl)-a-phosphonoglycine trimethyl ester,
.
DBU, DCM; (v) PPh3, toluene; (vi) SnCl2 2H2O, MeOH; (vi) 3,5-
dichloroisonicotinoyl chloride, Et3N, DCM; (viii) LiOH, THF, H2O.
more potent (VLA-4/VCAMprotein IC
50
VLA-4/VCAMcell IC 3.7 mM, IPRL k 4.8 hꢀ1).
0.05 mM;
50
potencies,8 the data from the unsaturated systems sug-
gests that the shape of the antagonist can have a pro-
found effect upon both potency and clearance. The
observed difference in clearance is presumably a result
of the difference in binding to plasma protein, for
example the slowly cleared E isomers 6 and 14 are both
greater than 99% bound to plasma whereas the rapidly
cleared Z isomers 9 and 15 are 70 and 87% bound
respectively. We have observed a similar effect in a ser-
ies of squaric acid derivatives.14
Obviously the dehydrophenylalanines are con-
formationally far more rigid than their saturated ana-
logues. Whilst the extra flexibility of the saturated systems
probably accounts for their easier accommodation in the
integrin ligand binding site and hence their greater
We were interested to see whether it would be possible
to effect further conformational restraint in these com-
pounds and retain potency. Consequently, we prepared
the 1,3-benzodiazepine, 23, as shown in Scheme 2. The
azidobenzaldehyde 19, prepared in two steps from the
alcohol 17, was treated with the phosphonoglycine in
the presence of DBU to give the dehydroalanine 20.
Cyclisation to the benzodiazepine 21 by means of an
intramolecular aza-Wittig reaction15 was achieved in
80% yield and was followed by reduction of the nitro
group and coupling with the acid chloride to give the
amide 22. Hydrolysis of the ester gave the target acid 23.
Table 2. Potency and rate of clearance for nicotinamides
Compd
R
E/Z
Z
VLA-4
protein
IC50 nM
VLA-4
cell
IC50 nM
IPRL
k hꢀ1
220
7100
3.7
The benzodiazepine, 23, did show pleasing levels of
potency (VLA-4/VCAMprotein IC
VCAMcell IC
11
12
55 nM; VLA-4/
410 nM) but, as may have been pre-
50
50
dicted from the conformation, it was subject to rapid
clearance having an IPRL clearance rate constant of 5.3
Z
Z
E
140
1.4
66
14,000
70
nd
h
ꢀ1, corresponding to a clearance of 140 mL/min/kg in
the rat (dosing at 10 mg/kg). This compound is, how-
ever, the first of a novel structural class of VLA-4
antagonists.
13
14
3.9
1.5
References and Notes
8500
1. Hynes, R. Cell 1987, 48, 554.
2. Plow, E.; Haas, T.; Zhang, L.; Loftus, J.; Smith, J. J. Biol.
Chem. 2000, 275, 21785.

808
J. R. Porter et al. / Bioorg. Med. Chem. Lett. 13 (2003) 805–808
3. Elices, M. J.; Osborn, L.; Takada, Y.; Crouse, C.;
Luhowskyj, S.; Hemler, M. E.; Lobb, R. R. Cell 1990, 60, 577.
4. Yusuf-Makagiansar, H.; Anderson, M. E.; Yakovleva, T. Y.;
Murray, J. S.; Siahaan, T. J. Med. Res. Rev. 2002, 22, 146.
5. Elices, M. J. In Cell Adhesion Molecules and Matrix Pro-
teins: Role in Health and Diseases; Mousa, S. A., Ed.;
Springer-Verlag and R. G. Landes Co.: Berlin, 1998; pp 133–
147.
pounds were titrated into a solution of 2-domain VCAM-Fc-
Ig in a separate plate and added to the wells. The assay was
carried out in TBS, 1% BSA, 1 mMMnCl 2, 0.1% Tween.
After incubation for 2 h at room temperature the plates were
washed and residual VCAMvisualised with peroxidase cou-
pled anti-human Fc.
12. A Jurkat cell line expressing VLA4 was incubated at 37 ꢁC
for 30 min with human 2-domain VCAM-1-FC immobilised
on a plate with anti-human FC in the presence of the test
compounds. The plates were washed and residual cells were
stained with Rose Bengal.
6. Sorbera, L. A.; Martin, L.; Rabasseda, X. Drugs Future
2000, 25, 917. Elices, M. Curr. Opin. Anti-Inflamm. Immuno-
mod. Invest. Drugs 2000, 2, 228.
7. (a) Porter, J. R. IDrugs 2000, 3, 788. (b) Tilley, J. W.; Sid-
duri, A. Drugs of the Future 2001, 26, 985. (c) Jackson, D. Y.
Curr. Pharm. Des. 2002, 8, 1229.
13. The identity of the geometric isomers was confirmed by
recording NOESY data in phase sensitive mode with a mixing
time of 350 ms. For example with compound 6 the olefinic H
1
8. Archibald, S. C.; Head, J. C.; Gozzard, N.; Howat, D. W.;
Parton, A. H.; Porter, J. R.; Robinson, M. K.; Shock, A.;
Warrellow, G. J.; Abraham, W. M. Bioorg. Med. Chem. Lett.
2000, 10, 997.
resonances in the two conformers are isochronous at 6.60
ppm. Medium strength, positively phased NOESY cross peaks
are observed between this signal and the two NH signals at
9.77 ppm (major) and 10.05 ppm (minor). This identifies 6 as
the E isomer. With compound 5, the olefinic 1H resonances are
resolved at 7.31 ppm (major) and 7.40 ppm (minor). In the
NOESY data no crosspeaks are observed corresponding to
these resonances. Weak, positively phased nOe crosspeaks are
observed between the two NH resonances in each conformer
at 9.39 ppm (major) and 9.66 ppm (minor) and the higher
frequency phenylalanine AA0BB0 multiplet resonance centred
at 7.67 ppm. This identifies 5 as the Z isomer. Parallel data
was recorded for the other isomeric pairs of molecules.
9. See for example (a) Hagmann, W. K.; Durette, P. L.;
Lanza, T.; Kevin, N. J.; de Laszlo, S. E.; Kopka, I. E.; Young,
D.; Magriotis, P. A.; Li, B.; Lin, L. S.; Yang, G.; Kamenecka,
T.; Chang, L. L.; Wilson, J.; MacCoss, M.; Mills, S. G.; Van
Riper, G.; McCauley, E.; Egger, L. A.; Kidambi, U.; Lyons,
K.; Vincent, S.; Stearns, R.; Colletti, A.; Teffera, J.; Tong, S.;
Fenyk-Melody, J.; Owens, K.; Levorse, D.; Kim, P.; Schmidt,
J. A.; Mumford, R. A. Bioorg. Med. Chem. Lett. 2001, 11,
2709. (b) Chen, L.; Trilles, R.; Miklowski, D.; Huang, T.-L.;
Fry, D.; Campbell, R.; Rowan, K.; Schwinge, V.; Tilley, J. W.
Bioorg. Med. Chem. Lett. 2002, 12, 1679.
14. Porter, J. R.; Archibald, S. C.; Childs, K.; Critchley, D.;
Head, J. C.; Linsley, J.; Parton, T. A. H.; Robinson, M. K.;
Shock, A.; Taylor, R. J.; Warrellow, G. J.; Alexander, R.;
Langham, B. Bioorg. Med. Chem. Lett. 2002, 12, 1051.
10. Frink, E.; Kramer, T.; Banchy, S.; Brown, B. Anesth.
Analg. 1990, 71, 484.
11. VLA4 (from HL60 lysate) was immobilised on a plate
with a non-blocking anti-b1 antibody (TS2/16). The test com-
15. Wamhoff, H.; Rechardt, G.; Stolben, S. Advances in
Hetreocyclic Chemistry 1996, 64, 159.
¨