Biaryl analogues of teriflunomide as potent DHODH inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.10.052

The structure-activity relationships of a novel series of biaryl dihydroorotate dehydrogenase (DHODH) inhibitors related to teriflunomide are disclosed. These biaryl derivatives were the result of structure-based design and proved to be potent DHODH inhibitors which in addition showed good antiproliferative activities on peripheral blood mononuclear cells and good efficacies in vivo in the rat adjuvant-induced-arthritis model.

Full text of DOI:10.1016/j.bmcl.2011.10.052

Bioorganic & Medicinal Chemistry Letters 21 (2011) 7268–7272
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Biaryl analogues of teriflunomide as potent DHODH inhibitors
Montse Erra a, , Imma Moreno a, Jordi Sanahuja a, Miriam Andrés a, Raquel F. Reinoso b, Estrella Lozoya a,
⇑
Pilar Pizcueta c, Núria Godessart a, Julio Cesar Castro-Palomino d
a Almirall, R&D Centre, Laureà Miró 408-410, Sant Feliu de Llobregat, 08980 Barcelona, Spain
b Esteve, Av. Mare de Déu de Montserrat 221, 08041 Barcelona, Spain
c CNIO, Melchor Fernández Almagro 3, 28029 Madrid, Spain
d PaloBiofarma, Baldiri i Reixac 15-21, 08028 Barcelona, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
The structure–activity relationships of a novel series of biaryl dihydroorotate dehydrogenase (DHODH)
inhibitors related to teriflunomide are disclosed. These biaryl derivatives were the result of structure-
based design and proved to be potent DHODH inhibitors which in addition showed good antiproliferative
activities on peripheral blood mononuclear cells and good efficacies in vivo in the rat adjuvant-induced-
arthritis model.
Received 2 August 2011
Revised 14 October 2011
Accepted 14 October 2011
Available online 20 October 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
DHODH inhibitors
Teriflunomide
Autoimmune diseases
RA
MS
Dihydroorotate dehydrogenase (DHODH) is a mitochondrial
enzyme that catalyzes the fourth step in the pyrimidine biosyn-
thetic pathway, namely the conversion of dihydroorotate to oro-
tate.1 Inhibitors of DHODH exhibit anti-inflammatory and
immunosuppressant properties and have demonstrated to be effi-
cacious in the treatment of autoimmune diseases such as multiple
sclerosis2 and rheumatoid arthritis.3
Leflunomide 1 and brequinar 2 (Fig. 1) are two examples of low-
molecular weight inhibitors of DHODH that have been in clinical
development.2–5 Leflunomide (AravaÒ, Sanofi-Aventis) is the only
member of its class in the market. It is approved for the treatment
of rheumatoid arthritis and psoriatic arthritis.6 Teriflunomide 3 is
the active metabolite of leflunomide4 and is currently in Phase III
clinical trials for multiple sclerosis.7
teriflunomide.9 Analyses of this crystal (Fig. 2) provided a good
starting point for structure-based design of more potent inhibitors.
The crystal revealed that trifluoromethyl phenyl group present
in 3 only partially filled the binding cavity that is mainly lined by
hydrophobic amino acid residues (Fig. 3).
Initially, methyl group of teriflunomide 3 was changed by ethyl
and a two-fold increase in potency was observed. Then, in com-
pound 4 (Fig. 4), where a bulkier hydrophobic group was intro-
duced to better fill the cavity, an improvement in in vitro
hDHODH activity by an order of magnitude was obtained (Table
1). In addition, the iv half-life of 4 was significantly reduced com-
pared to 3, but was still perceived as too long having in mind the
different behaviour observed with teriflunomide between rat
(14 h) and man (2 weeks) (Table 1).
However, teriflunomide has a low potency against human
DHODH3 and a long half-life of approximately 2 weeks in plasma,8
which could represent a serious obstacle for patients who need to
withdraw the treatment in case of toxicity or pregnancy. Herein,
we report our efforts to increase inhibitory potency against human
DHODH enzyme and to adjust pharmacokinetic properties by
reducing half-life.
The use of soft metabolic centres was then investigated as a
mean to modulate and reduce t1/2. The idea of introducing a methyl
ester led to compound 5. This molecule proved to be 15-fold more
O
OH
Our strategy to develop DHODH inhibitors was based on the re-
cently published co-crystal structure of the human enzyme with
F3C
F3C
F
O
O
CN
N
H
N
H
F
N
N
O
HO
2
1
3
⇑
Corresponding author. Tel.: +34 933 128714; fax: +34 932 913420.
E-mail address: montse.erra@almirall.com (M. Erra).
Figure 1. Reference compounds.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.10.052

M. Erra et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7268–7272
7269
Figure 2. Crystal structure of teriflunomide in the lipophilic pocket of human DHODH (PDB code: 1D3H). The surface of the binding site is shown in green spheres. Dashed
lines correspond to Hydrogen bond interactions with aminoacids Arg136 and Tyr356.
Figure 3. Surface representation of teriflunomide’s trifluoromethylphenyl moiety inside the hydrophobic channel of human DHODH. The pictures show the space not filled
by teriflunomide binding.
Table 1
CF3
O S O
O
Biological activities and pharmacokinetic profiles in rat
F3C
O
O
Compound
hDHODH IC50 (nM)
PK iv (1 mg/kg) t1/2 (h)
N
N
N
H
N
H
3
4
5
1083
150
100
14.1
8
2
HO
HO
3
4
CF3
O S O
O
O
O
O
N
binding to proteins. The triflate group in 5 was put to good syn-
thetic use, however, as a cross-coupling partner, which allowed a
rapid assessment of SAR. Thus, introduction of a phenyl ring10
was effected using a Suzuki reaction and variation of the substitu-
ents on this ring were explored with respect to both enzymatic and
cellular immunosuppressive activity (Table 2).
N
H
HO
5
Figure 4. First teriflunomide analogues.
A flexible synthetic route (Scheme 1) was employed in the syn-
thesis of these series of compounds. Thus, synthesis of intermedi-
ate 7 was carried out by methylation of 2-hydroxy-5-nitrobenzoic
acid 6, subsequent reduction of the nitro group to give the corre-
sponding aniline and coupling of this aniline with cyanoacetic acid.
Intermediate 7 was treated with a triflating agent to give the
active than 3 in in vitro hDHODH activity and a greatly modified iv
PK profile in rat (Table 1).
A concern with compound 5, however, was that it contained a
potentially displaceable triflate in the b-position of an
a,b-unsatu-
rated carboxylate which could be prone to promote covalent

7270
M. Erra et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7268–7272
Table 2
Table 3
Biological activities of biaryl derivatives
In vitro profiles of selected derivatives
O
O
Compound
hDHODH IC50
(l
M)
Proliferation hPBMC IC50 (lM)
R
3
1.08
0.11
0.03
0.04
0.03
0.03
0.14
0.08
46
2
34
13
37
6
O
OH
10
12
13
14
16
21
24
N
H
N
R
Compound
IC50 hDHODH (nM)
19
18.5
OSO2CF3
Phenyl
4-Fluorophenyl
2-Chlorophenyl
3-Chlorophenyl
4-Chlorophenyl
3-Trifluoromethoxyphenyl
3-Fluorophenyl
2,4-Difluorophenyl
3-Pyridine
4-Fluoro-2-methylphenyl
6-Fluoropyridin-3-yl
3,4-Difluorophenyl
3,4,5-Trifluorophenyl
4-Fluoro-2-methoxyphenyl
3-Ethoxyphenyl
5
9
93
73
110
52
31
45
27
198
31
40
49
150
97
137
66
59
76
26
66
34
31
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Table 4
In vitro and in vivo profiles of selected derivatives
Compound hDHODH
IC50 (nM)
rDHODH
IC50 (nM)
R/H
PK iv rat
AIAb ED50
(mg/Kg)
Met%a (1 mg/kg) t1/2
(h)
3
10
1083
110
26
350
0/3
14
2.88
49% @
10 mg/kg
19/17 0.9
a
Assay conditions: 1 mg/mL of microsomal protein in a phosphate buffer con-
taining the compd at 5 M at 37 °C. The reaction is initiated by the addition of the
NADPH generating system. Results:
1,3-Benzodioxol-5-yl
3-Methoxyphenyl
l
% of compound degradation after 30 min
2,3-Dihydro-1,4-benzodioxin-6-yl
3-Difluoromethoxyphenyl
3-Cyclobutoxyphenyl
incubation.
b
Results are expressed as the inhibition of the area under the curve of the right
hind paw swelling during the whole treatment period (10 days), using 6–7 rats per
group. The arthritis was induced by sub-plantar injection of complete Freund’s
adjuvant in the left hind paw. Vehicle or compounds were dosed once daily in 0.5%
methocel starting from day 10 post-induction and for 10 consecutive days.
F3C
O
O
O
S
O
O
OH O
OH O
us, in effect, to reduce the half-life (Table 4). The corresponding
acid was found to be the main metabolite in human microsomes
and it was 100-fold less active against hDHODH than the ester.
In an attempt to identify the specific non-covalent binding
interactions which were responsible for the high binding affinity,
co-crystal structures with the human enzyme were determined
for selected analogs. Herein we describe the structure of DHODH
complex with the inhibitor 10. The structure was solved at a reso-
lution of 2.18 Å, and revealed the detailed binding mode of the li-
gand bound to the putative ubiquinone binding site.
O
OH
O
ii
i
HN
NO2
HN
O
8
7
6
CN
iii
CN
R
F3C
O
O
O
S
O
O
O
O
iv
The ligand forms two specific hydrogen bonds to DHODH,
namely to the side chain residues Arg136 and Tyr356. Further-
more, binding of ligand is enhanced by a variety of hydrophobic
HN
O
HN
O
HO
HO
CN
CN
9-28
interactions involving residues of the N-terminal helices
a1 and
4
a2—namely Tyr38, Met43, Leu46, Pro52, His56, Ala59, Phe62,
Leu68 as well as Phe98, Met111, Thr360 and Pro364 (Fig. 5).
These hydrophobic interactions are consistent with the good
activity of the lipophilic biphenyl derivatives (compounds
9–28)—in particular, compounds 12, 13, 14 16, 17, 25, 27 and 28
with an IC50 <50 nM.
Scheme 1. Reagents and conditions: (i) (a) Me2SO4, acetone, reflux, 89%, (b) H2, Ni-
Ra, THF/MeOH, 40 psi, 85%, (c) CNCH2COOH, EDC.HCl, rt, 89%; (ii) PhN(OTf)2, Et3N,
THF, reflux, 58%; (iii) propionic anhydride, NaH, THF, rt, 78%; (iv) RB(OH)2,
Pd(PPH3)4, K2CO3, dioxane, 100 °C, 52–75%.
The co-crystal structure reveals that the newly introduced phe-
nyl group fills the above hydrophobic binding cavity in a more
complementary fashion than the CF3 of 3 (Fig. 6).
cyanoacetamide 8, which was then acylated to afford the b-
hydroxyenamide 4 using sodium hydride and propionic anhydride.
Finally, compounds 9–28 were synthesized by standard Suzuki
coupling of compound 4 and the corresponding boronic acid or
boronic acid pinacol ester employing tetrakis(triphenylphos-
phine)palladium(0) as a catalyst. Most boronic acids were either
commercially available or prepared by standard literature
methods.
Inhibition of DHODH at a cellular level translates into the inhi-
bition of proliferation of mitogen-stimulated human peripheral
blood mononuclear cells (hPBMCs),11 which can be measured by
incorporation of tritiated thymidine following 72 h of incubation
of cells with phytohemagglutinin (PHA). Data for few selected
compounds are presented in Table 3. Thus, enzymatic inhibition
of hDHODH was translated into a good cellular effect, in particular
compounds 10 and 16 which displayed potent antiproliferative
activities on PBMCs.
This hydrophobic biphenyl tail10 gave good binding affinity for
human DHODH enzyme (IC50 <0.2 lM). As shown in Table 2, sev-
eral compounds—for example, the 3-chloro (12, 31 nM) and 3-tri-
fluoromethoxy (14; 27 nM) and closely related analogs—were
particularly potent. Moreover, the attractive PK properties of the
series were maintained by inclusion of the methyl ester, allowing
The in vivo anti-inflammatory effect of compound 10 was fur-
ther completed using the functional model of adjuvant induced
arthritis (AIA) in rats12 as shown in Table 4.

M. Erra et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7268–7272
7271
Figure 5. Superimposition of the co-crystallized complexes of teriflunomide (blue) and compound 10 (PDB code: 3U2O) (orange) at the binding site of human DHODH
enzyme. The surface of the binding site is shown in green spheres. Dashed lines correspond to Hydrogen bond interactions with aminoacids Arg136 and Tyr356. The
fluorophenyl group of compound 10 fills better the hydrophobic cavity of the enzyme and makes several interactions along the channel defined by residues: Tyr38, Met43,
Leu46, Pro52, His56, Ala59, Phe62, Leu68, Phe98, Met111, Thr360 and Pro364.
Figure 6. Surface representation of the hydrophobic channel of human DHODH. The picture shows the additional space filled by compound 10 (orange) in comparison with
teriflunomide (blue) inside the hydrophobic cavity of the enzyme. Compound 10 makes more hydrophobic interactions inside the hydrophobic channel.
Compound 10 was shown to be active in the AIA model,
although with a lower potency than teriflunomide. These results
are not surprising in the light of the reduced half-life and the lower
potency as rat DHODH inhibitor of compound 10 versus terifluno-
mide. Given that compound is 10 times more potent in the human
enzyme, we would expect a similar efficacy as teriflunomide in hu-
mans and with a more acceptable pharmacokinetic profile.
A new family of potent and orally bioavailable hDHODH inhib-
itors has been designed, synthesized and characterized. Compound
10 is more than 20-fold more potent than teriflunomide as an anti-
proliferative in PBMCs and has an attractive PK profile in rat which
is likely to obviate the liabilities (a half life of 2 weeks in man) of
teriflunomide. Compound 10 has also shown efficacy in the exper-
imental model of arthritis, AIA. Thus, compound 10 has been
identified as an attractive starting point for further optimization.
Further studies have addressed replacements for the b-hydroxye-
namide moiety, to minimize the possible biphenylaniline metabo-
lite that could be formed during Phase I metabolism. This work will
be reported in due course.
Acknowledgment
The authors wish to thank Proteros Biostructures GmbH for X-
ray structure determination of human DHODH in complex with
compound 10 (PDB code: 3U2O).
References and notes
1. Loffler, M.; Jöckel, J.; Schuster, G.; Becker, C. Mol. Cell Biochem. 1997, 174, 125.
2. O’Connor, P. W.; Li, D.; Freedman, M. S.; Bar-Or, A.; Rice, G. P. A.; Confraveux, C.;
Patty, D. W.; Steweart, J. A.; Scheyer, R. Neurology 2006, 66, 894.
3. Fox, R. I. J. Rheumatol. Suppl. 1998, 53, 20.
4. Breedveld, F. C.; Dayer, J. M. Ann. Rheum. Dis. 2000, 59, 841.

7272
M. Erra et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7268–7272
5. Makowka, L.; Sher, L. S.; Cramer, D. V. Immunol. Rev. 1993, 136, 51.
6. Behrens, F.; Koehm, M.; Burkhardt, H. Curr. Opin. Rheumatol. 2011, 23, 282.
7. O’ Connor, P.; Wiolinsky, J. S.; Confraveux, C.; Comi, G.; Kappos, L.; Olsson, T. P.;
Benzerdjeb, H.; Truffinet, P.; Wang, L.; Miller, A.; Freedman, M. S. N. Engl. J. Med.
2011, 365, 14.
9. Liu et al Structure 2000, 8, 25.
10. Davies, M.; Heikkilä, T.; McConkey, G. A.; Fishwick, C. W.; Parsons, M. R.;
Johnson, A. P. J. Med. Chem. 2009, 52, 2683.
11. Fitzpatrick, L. R.; Deml, L.; Hoffmann, C.; Small, J. S.; Groeppel, M.; Hamm, S.;
Lemstra, S.; Leban, J.; Ammendola, A. Inflamm. Bowel Dis. 2010, 16, 1763.
12. Pearson, C. M. Proc. Soc. Exp. Biol. (New York) 1956, 91, 95.
8. Rozman, B. Clin. Pharmacokinet. 2002, 41, 421.