Heterocycle-substituted proline dipeptides as potent VLA-4 antagonists

Article abstract of DOI:10.1016/j.bmcl.2009.12.009

A variety of N-linked tertiary amines and heteroarylamines were examined at the 4-position of sulfonylated proline dipeptides in order to improve VLA-4 receptor off-rates and overcome the issue of CYP3A4 time-dependent inhibition of ester prodrugs. A tight-binding inhibitor 5j with a long off-rate provided sustained receptor occupancy despite poor oral pharmacokinetics.

Full text of DOI:10.1016/j.bmcl.2009.12.009

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1173–1176
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Heterocycle-substituted proline dipeptides as potent VLA-4 antagonists
Thomas S. Reger a, , , Jasmine Zunic a,à, Nicholas Stock a,à, Bowei Wang a,§, Nicholas D. Smith a,–
,
*
Benito Munoz a,||, Mitchell D. Green b,–, Michael F. Gardner b, , Joyce P. James b, , Weichao Chen b,àà
,
Kenneth Alves c, Qian Si c, Kelly M. Treonze c, Russell B. Lingham c, Richard A. Mumford c
a Department of Medicinal Chemistry, Merck Research Laboratories, San Diego, CA 92121, USA
b Department of Drug Metabolism and Pharmacokinetics, Merck Research Laboratories, San Diego, CA 92121, USA
c Department of Immunology and Rheumatology Research, Merck Research Laboratories, Rahway, NJ 07065, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A variety of N-linked tertiary amines and heteroarylamines were examined at the 4-position of sulfony-
lated proline dipeptides in order to improve VLA-4 receptor off-rates and overcome the issue of CYP3A4
time-dependent inhibition of ester prodrugs. A tight-binding inhibitor 5j with a long off-rate provided
sustained receptor occupancy despite poor oral pharmacokinetics.
Received 3 November 2009
Revised 1 December 2009
Accepted 2 December 2009
Available online 4 December 2009
Ó 2009 Elsevier Ltd. All rights reserved.
TheadhesionmoleculeVLA-4(
a
4b1; CD49d/CD29;‘verylateanti-
VLA-4.8 Extensive SAR led to identification of the 3-cya-
nobenzenesulfonyl proline group coupled to a N-isonicotinoyl-( )-
4-aminophenylalanine unit as an optimal scaffold for potency
against VLA-4.8 Additionally, the presence of a basic amine at the
4-position of the proline ring is crucial for minimizing plasma pro-
tein binding.
gen-4’) is a member of the integrin family and is expressed on all cir-
culating leukocytes except platelets.1 VLA-4 binds to vascular cell
adhesion molecule-1 (VCAM-1) which is expressed on activated
endothelial cells and is upregulated in response to inflammatory
cytokines. The specific interaction between VLA-4 on leukocytes
and VCAM-1 on the vascular endothelium may be required for the
activation, migration, proliferation, and differentiation of leukocytes
during normal and pathophysiological processes.2 Inhibition of
VLA-4 may therefore be effective in preventing recruitment and
infiltration of cell types required for a prolonged inflammatory
response. Indeed, monoclonal antibodies and blocking peptides
L
against the
models of asthma,3 rheumatoid arthritis,4 and multiple sclerosis.5
The humanized 4 monoclonal antibody natalizumab (Tysabri, Bio-
a4-integrin have been shown to be effective in animal
a
gen Idec/Elan) was recently approved for the treatment of multiple
sclerosis and inflammatory bowel disease.6
Previous communications from our laboratories have detailed
the development of small molecule inhibitors of VLA-4.7 Compound
1 is representative of a class of sulfonylated dipeptides that exhibit
strong potency against both the activated and resting states of
In general, this class of inhibitors suffers from poor oral bio-
availability and high plasma clearance in preclinical species. De-
spite the poor pharmacokinetics of 1, it was anticipated that its
slow off-rate (90% bound at 1 h) from the VLA-4 receptor might
lead to a sustained receptor occupancy (RO) in vivo and the poten-
tial for a prolonged pharmacological effect. Indeed, after a 5 mpk
oral dose of 1 to rats, 67% RO was observed at 12 h. Furthermore,
administration of the ethyl ester prodrug of 1 under the same
parameters gave higher RO (77%) of the acid at 12 h.8 The ester,
however, was found to be a time-dependent inhibitor of CYP3A4
which may pose a risk for clinical drug–drug interactions.9
* Corresponding author. Tel.: +1 215 652 2438; fax: +1 215 652 3971.
E-mail address: thomas_reger@merck.com (T.S. Reger).
Present address: Merck Research Laboratories, West Point, PA, 19446, USA.
Present address: Amira Pharmaceuticals, 9535 Waples St., Suite 100, San Diego,
à
CA 92121, USA.
§
Present address: Merck Research Laboratories, Rahway, NJ 07065, USA.
Present address: Aragon Pharmaceuticals, 4215 Sorrento Valley Blvd., Suite 215,
–
Oxidative dealkylation of the secondary amine on proline was
thought to be a key step in a pathway leading to the species
responsible for CYP3A4 time-dependent inhibition (TDI).10 Our
strategy to minimize this event was to replace the secondary alkyl
amine of 1 with either a tertiary amine or an N-linked aromatic
San Diego, CA 92121, USA.
||
Present address: Broad Institute, 7 Cambridge Center, Cambridge, MA 02142, USA.
Present address: Ambit Biosciences, 4215 Sorrento Valley Blvd., San Diego, CA
92121, USA.
àà
Present address: Arena Pharmaceuticals, 6166 Nancy Ridge Dr., San Diego, CA
92121, USA.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.12.009

1174
T. S. Reger et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1173–1176
with heteroarylamines or secondary amines to provide products
3.12 Hydrolysis of the methyl ester was carried out with LiOH
and the resulting acid was coupled to N-isonicotinoyl-( )-4-amin-
L
ophenylalanine ethyl ester 4 to give the sulfonylated dipeptide
ethyl esters. Conversion to the carboxylic acids was readily
achieved through hydrolysis with lithium hydroxide.
Acids 5a–s were tested for potency in an in vitro binding assay
against the resting state of human VLA-4.13,14 The effect of plasma
protein binding was evaluated by repeating the assay in the pres-
ence of 90% human plasma. Finally, the corresponding ethyl esters
were counter-screened for TDI of CYP3A4.
Table 1 summarizes the results for heteroaryl-substituted ana-
logues. Tetrazole 5d and benzimidazole 5g displayed the strongest
binding affinity to VLA-4 with IC50 values of 80 pM. With the
exception of indazole 5e, all compounds in this series had reduced
TDI relative to secondary amine-substituted proline derivatives.
We next examined cycloalkylamine substitution on proline and
the effect of ring size on inhibition of VLA-4. As illustrated in Table
2, compounds 5h–l exhibited subnanomolar inhibition of VLA-4
and, with the exception of eight-membered ring analogue 5l, had
minimal potency shift in the presence of 90% plasma. The greater
hydrophobicity of the cyclooctylamine group in 5l likely results
in higher plasma protein binding and the observed potency shift.
Among this group, piperidine 5j had the best overall profile with
no plasma shift in the binding assay and the lowest TDI value for
a corresponding ethyl ester.
Scheme 1. Reagents and conditions: (a) 3-cyanobenzenesulfonyl chloride, NEt3,
CH2Cl2; (b) (i) Tf2O, iPr2NEt, CH2Cl2, À78 °C, (ii) amine (R), À30 °C to rt; (c) (i) LiOH,
MeCN/H2O, (ii) HCl; (d) 4, HATU, NMM, DMF; (e) LiOH, MeCN/H2O.
heterocycle, thus slowing dealkylation and reducing TDI. Herein,
we describe our efforts to identify potent VLA-4 antagonists with
slow receptor off-rates and high receptor occupancy, and whose
ethyl ester prodrugs have low risk of TDI.11
Our general synthetic approach to compounds 5a-s is outlined
in Scheme 1. After N-sulfonylation of cis (4S)-hydroxyproline
methyl ester, the alcohol was converted to the triflate and reacted
Building on this result, a number of additional N-linked six-
membered ring heterocycles were attached to the proline core.
Morpholine 5m and thiomorpholine 5n along with methylpiperi-
dine compounds 5o–s were all very potent in the VLA-4 binding
Table 1
Inhibition of VLA-4 by heteroaryl-substituted proline analogues
Compd
5a
R
VLA-4a (IC50, nM)
0.22
90% Plasmab (IC50, nM)
0.48
TDIc (minÀ1
)
0.009
5b
5c
5d
5e
0.81
0.55
0.08
0.25
100% 100 nM
93% 62.5 nM
100% 100 nM
2.01
0.012
0.007
0.008
0.038
5f
0.35
0.08
4.75
0.50
0.016
0.019
5g
a
b
c
Competitive binding assay against the resting state of VLA-4.
See Ref. 8 for a description of this assay.
Values for CYP3A4 time-dependent inhibition (kobs, 10 lM) were determined from the ethyl ester.

T. S. Reger et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1173–1176
1175
Table 2
Effect of ring size and substitution on inhibition of VLA-4
Compd
5h
R
VLA-4a,b (IC50, nM)
0.34
90% Plasmab (IC50, nM)
TDIc (minÀ1
)
0.42
0.22
0.022
0.024
5i
5j
0.16
0.11
0.12
0.45
0.016
0.021
5k
51
0.42
0.25
0.81
0.023
5m
5n
0.13
0.13
0.11
0.15
0.23
0.20
0.020
0.002
0.002
5od
5pd
5q
0.13
0.16
0.14
0.17
0.17
0.31
0.019
0.009
0.019
5rd
5s
0.16
0.41
0.011
a
b
c
Competitive binding assay against the resting state of VLA-4.
See Ref. 8 for a description of this assay.
Values for CYP3A4 time-dependent inhibition (kobs, 10 lM) were determined from the ethyl ester.
d
Mixture of diastereomers
assay with minimal shifts in the presence of 90% plasma. Addition-
ally, the ethyl esters of all the compounds in this group had favor-
able TDI rates of 0.020 minÀ1 or lower. Thiomorpholine 5n and 2-
methylpiperidine 5o had particularly low TDI values, suggesting
that oxidative dealkylation to a primary amine may be mitigated
with these substructures.
A receptor off-rate assay was utilized to further differentiate be-
tween these numerous potent, non-plasma shifted compounds.14
These results are presented as the percentage of compound bound
to receptor after 3 h and are shown in Table 3. Compounds 5j, 5p,
and 5s emerged as the tightest binding antagonists as they re-
mained 85% bound to receptor after 3 h.
The subnanomolar potency, absence of plasma shift, and slow
off-rate of 5j, in addition to reduced TDI risk for its ethyl ester
(5j–Et), highlights the desirable profile of this acid/ester pair. We
utilized a rat receptor occupancy model to evaluate the potential
for target engagement in vivo.14 As shown in Figure 1, a 5 mpk oral
dose of acid 5j resulted in receptor occupancy of nearly 80% at 12 h
post-dose, decreasing to 20% at 24 h. The same dose of prodrug 5j–
Et significantly improved receptor occupancy to >90% at 12 h and
remained at 60% after 24 h.
Plasma levels of 5j were determined in systemic circulation and
in the portal vein after oral dosing of 5j or 5j–Et. Table 4 shows the
systemic and portal vein Cmax of acid 5j after a 5 mpk oral dose of 5j

1176
T. S. Reger et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1173–1176
Table 3
Acknowledgment
Off-rate determination for selected compounds
Compd
Off-ratea (%)
We thank Dr. William Hagmann for critical reading of this
Letter.
5d
5g
5i
10
35
65
85
60
15
34
38
85
67
72
85
References and notes
5j
5k
5m
5n
5o
5p
5q
5r
1. (a) Hemler, M. E.; Elices, M. J.; Parker, C.; Takada, Y. Immunol. Rev. 1990, 114,
45; (b) Elices, M. J. Curr. Opin. Anti-inflam. Immun. 1999, 1, 15.
2. (a) Gearing, A. J. H.; Newman, W. Immunol. Today 1993, 14, 506; (b) Hemler, M.
E. Annu. Rev. Immunol. 1990, 8, 365.
3. (a) Abraham, W. M.; Ahmed, A.; Sielczak, M. W.; Narita, M.; Arrhenius, T.;
Elices, M. J. Am. J. Resp. Crit. Care Med. 1997, 156, 696; (b) Koo, G. C.; Shah, K.;
Ding, G. J.; Xiao, J.; Wnek, R.; Doherty, G.; Tong, X. C.; Pepinsky, R. B.; Lin, K. C.;
Hagmann, W. K.; Kawka, D.; Singer, I. I. Am. J. Respir. Crit. Care Med. 2003, 167,
1400.
5s
a
4. Seiffge, D. J. Rheum. 1996, 23, 2086.
Bound at 3 h; see Ref. 8 for a description of this assay
5. (a) Pedolsky, D.; Lobb, R. R.; King, N.; Benjamin, C.; Pepinsky, R. B.; Sehgal, P.;
deBeaumont, M. J. Clin. Invest. 1993, 92, 372; (b) Leone, D. R.; Giza, K.; Gill, A.;
Dolinski, B. M.; Yang, W.; Perper, S.; Scott, D. M.; Lee, W. C.; Cornebise, M.;
Wortham, K.; Nickerson-Nutter, C.; Chen, L. L.; LePage, D.; Spell, J. C.; Whalley,
E. T.; Petter, R. C.; Adams, S. P.; Lobb, R. R.; Pepinsky, R. B. J. Pharmacol. Exp. Ther.
2003, 305, 1150.
6. (a) Miller, D. H.; Khan, O. A.; Sheremata, W. A.; Blumhardt, L. D.; Rice, G. P.;
Libonati, M. A.; Willmer-Hulme, A. J.; Dalton, C. M.; Miszkiel, K. A.; O’Connor, P.
W. N. Engl. J. Med. 2003, 348, 15; (b) Sandborn, W. J.; Colombel, J. F.; Enns, R.;
Feagan, B. G.; Hanauer, S. B.; Lawrance, I. C.; Panaccione, R.; Sanders, M.;
Schreiber, S.; Targan, S.; van Deventer, S.; Goldblum, R.; Despain, D.; Hogge, G.
S.; Rutgeerts, P. N. N. Engl. J. Med. 2005, 353, 1912.
100
80
60
40
7. (a) Doherty, G. A.; Yang, G. X.; Borges, E.; Tong, S.; McCauley, E. D.; Treonze, K.
M.; Van Riper, G.; Pacholok, S.; Si, Q.; Koo, G. C.; Shah, K.; Mumford, R. A.;
Hagmann, W. K. Bioorg. Med. Chem. Lett. 2003, 13, 1891; (b) Hagmann, W. K.
Curr. Top. Med. Chem. 2004, 4, 1461; (c) Venkatraman, S.; Lebsack, A. D.; Alves,
K.; Gardner, M. F.; James, J.; Lingham, R. B.; Maniar, S.; Mumford, R. A.; Si, Q.;
Stock, N.; Treonze, K. M.; Wang, B.; Zunic, J.; Munoz, B. Bioorg. Med. Chem. Lett.
2009, 19, 5803.
8. Lin, L. S.; Lanza, T., Jr.; Jewell, J. P.; Liu, P.; Jones, C.; Kieczykowski, G. R.;
Treonze, K.; Si, Q.; Manior, S.; Koo, G.; Tong, X.; Wang, J.; Schuelke, A.;
Pivnichny, J.; Wang, R.; Raab, C.; Vincent, S.; Davies, P.; MacCoss, M.; Mumford,
R. A.; Hagmann, W. K. J. Med. Chem. 2009, 52, 3449.
20
5 mpk 5j-Et
5 mpk 5j
0
0
3
6
9
12
24
Time (h)
Figure 1. Rat receptor occupancy of 5j from dosing of acid 5j or ethyl ester prodrug
5j–Et
9. Riley, R. J.; Grime, K.; Weaver, R. Expert Opin. Drug Metab. Toxicol. 2007, 3, 51.
10. Bensoussan, C.; Delaforge, M.; Mansuy, D. Biochem. Pharmacol. 1995, 49, 591.
11. We targeted kobs (10
prodrugs. By comparison, the TDI rate for the ethyl ester of 1 was 0.040 minÀ1
l
M) values of less than 0.020 minÀ1 for TDI by ethyl ester
Table 4
Rat pharmacokinetics of 5j
.
12. For discussions on the regioselectivity of 1,2,4-triazole and tetrazole alkylation,
see: (a) Bulger, P. G.; Cottrell, I. F.; Cowden, C. J.; Davies, A. J.; Dolling, U.-H.
Tetrahedron Lett. 2000, 41, 1297; (b) Ulhaq, S.; Chinje, E. C.; Naylor, M. A.; Jaffar,
M.; Stratford, I. J.; Threadgill, M. D. Bioorg. Med. Chem. Lett. 1998, 6, 2139.
13. We have previously demonstrated that this class of compounds generally has a
higher affinity for the activated state of VLA-4. Thus, potency was examined
only against the resting state of VLA-4 to best differentiate compounds.
14. The [125I]-tracer shown below was utilized to assess binding potency, receptor
off-rate, and rat receptor occupancy. A detailed description of each of these
experiments was recently provided in the Supplementary data of Ref. 8.
Compound dosed (5 mpk po)
Systemic Cmax (nM)
Portal vein Cmax (nM)
Acid 5j
Ester 5j–Et
3
12
245
397
or prodrug 5j–Et in rats. Levels of 5j are substantially higher in the
portal vein than in systemic circulation. This is critical as we be-
lieve that the slow off-rate and initial exposure of 5j to lympho-
cytes bearing VLA-4 in the portal vein combine to provide the
observed high levels of receptor occupancy in rats.
In summary, tertiary heterocycle-substituted proline dipeptides
were shown to be potent antagonists of VLA-4. Acid 5j had the
optimal profile, displaying excellent potency and a slow off-rate
from human VLA-4. The ester prodrug 5j–Et had reduced TDI of
CYP3A4 and gave sustained (12–24 h) receptor occupancy when
dosed to rats. We have recently characterized the intrinsic and
pharmacologic properties of this compound against equine VLA-
4.15 Results described therein suggest that 5j may have utility for
the treatment of recurring airway obstruction in horses.
15. Treonze, K. M.; Alves, K.; Fischer, P.; Hagmann, W. K.; Hora, D.; Kulick, A.;
Vakerich, K.; Smith, N. D.; Lingham, R. B.; Maniar, S.; Reger, T. S.; Zunic, J.;
Munoz, B.; Prasit, P.; Nicholson, D.; Si, Q.; Judd, K.; Nicolich, S.; Kellerhouse, P.;
Thompson, D.; Mumford, R. A. Vet. Immun. Immunopath. 2009, 130, 79.