Synthesis and multiparametric evaluation of thiadiazoles and oxadiazoles as diacylglycerol acyltransferase type 1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2015.11.046

Chemical modulation of a formerly disclosed DGAT-1 inhibitor resulted in the identification of a compound with a suitable profile for preclinical development. Optimisation of solubility is discussed and a PK/PD study is presented.

Full text of DOI:10.1016/j.bmcl.2015.11.046

Bioorganic & Medicinal Chemistry Letters xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and multiparametric evaluation of thiadiazoles and
oxadiazoles as diacylglycerol acyltransferase type 1 inhibitors
a
a
a
a
Patrick Mougenot a, , Claudie Namane , Eykmar Fett , Florence Goumy , Rommel Dadji-Faïhun ,
Gwladys Langot a, Catherine Monseau a, Bénédicte Onofri a, François Pacquet a, Cécile Pascal b,
Olivier Crespin a, Majdi Ben-Hassine a, Jean-Luc Ragot a, Thao Van-Pham a, Christophe Philippo a,
Florence Chatelain-Egger a, Philippe Péron a, Jean-Christophe Le Bail a, Etienne Guillot a,
Philippe Chamiot-Clerc a, Marie-Aude Chabanaud a, Marie-Pierre Pruniaux a, Jérôme Ménegotto c,
⇑
Friedemann Schmidt d, Olivier Venier a, Fabrice Viviani e,y, Eric Nicolai a,
⇑
a Sanofi-aventis R&D, 91385 Chilly-Mazarin, France
b Sanofi-aventis R&D, 69280 Marcy-l’Etoile, France
c Evotec, 31036 Toulouse, France
d Sanofi-aventis Deutschland GmbH, Industriepark Hoechst, 65926 Frankfurt am Main, Germany
e GlaxoSmithKline, 91951 Les Ulis, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Chemical modulation of a formerly disclosed DGAT-1 inhibitor resulted in the identification of a com-
pound with a suitable profile for preclinical development. Optimisation of solubility is discussed and a
PK/PD study is presented.
Received 2 October 2015
Revised 10 November 2015
Accepted 14 November 2015
Available online xxxx
Ó 2015 Elsevier Ltd. All rights reserved.
Keywords:
Diacylglycerol acyltransferase type 1
DGAT-1 inhibitors
Thiadiazole
Oxadiazole
Solubility
Metabolic lability
DGATs (diacylglycerol acyltransferases) are enzymes which
catalyse the final step of the major pathway of triglyceride synthe-
sis. DGAT-11 and DGAT-22 are encoded by two different genes and
are expressed in tissues associated with triglyceride metabolism,
particularly in adipose tissue, liver, mammary glands, small intes-
tine and muscle.1,2 DGAT-1 is anchored in ER membrane and con-
tains approximately 500 very hydrophobic amino acid residues.
Since 2007, several small molecules3 have entered clinical trials
as DGAT-1 inhibitors. Presently, two of them are still in clinical
OH
O
O
F
F3C
N
H
ONa
O
N
N
O
S
N
N
H
N
H
Figure 1. Pradigastat and P-7435.
development; pradigastat (Fig. 1) is in phase III since 2012
for familial hypertriglyceridemia4 and P-7435 is in phase I for
obesity.5
In 2012, we disclosed new and potent thiadiazoles as DGAT-1
inhibitors6 (Fig. 2), which completely blocked TG production
in vivo after an acute lipid challenge. Unfortunately, the low solu-
bility of these compounds prevented further pharmacological stud-
ies as no acceptable vehicle compatible with metabolic readouts
could be identified7 (see Table 1).
Abbreviations: DBU, 1,8-diazabicyclo[5.4.0]undec-7-ene; DCM, dichloro-
methane; DIAD, diisopropylazodicarboxylate; DMF, dimethylformamide; CDI, car-
bonyldiimidazole; met. lia., metabolic liability; MC, methylcellulose; ND, not
determined; PK, pharmacokinetic; PyBroP, bromotrispyrrolidinophosphonium hex-
afluorophosphonate; sol., solubility; T, tween; TEA, triethylamine; TFA, trifluo-
roacetic acid; TG, triglycerides; THF, tetrahydrofuran.
⇑
Corresponding authors.
y
Present address.
http://dx.doi.org/10.1016/j.bmcl.2015.11.046
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Mougenot, P.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.11.046

2
P. Mougenot et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
O
F
O
O
N
S
N
S
N
S
N
N
N
N
H
OH
N
H
OH
N
H
OH
O
O
O
2
1-trans
3
F
Figure 2. Compounds from Ref. 6.
Table 1
Lead compounds 1-trans, 2, and 3
Compd
Enz. DGAT-1 IC50
(l
M)8
Cell. DGAT-1 IC50
(l
M)9
% met. h/m/r10
Caco2 (10ꢀ7 cm/s)11
Sol. pH = 7.4 (l
M)12
1-trans
2
3
0.036
0.009
0.019
0.029
0.048
0.012
0:5:11
1:7:1
36:5:10
112
135
90
<2
<2
<2
It should be noted that the internal Sanofi solubility standard
decided to take it as a model compound to address both the solu-
bility and metabolic stability issues by introduction of polarity at
appropriate sites.
for preclinical candidate selection is set to 100
buffer at pH = 7.4 or in water.
lM in phosphate
Since compound 3 was the most potent cellular inhibitor in this
series but suffered from low solubility in phosphate buffer at pH
7.4 and moderate metabolic stability in human microsomes, we
The metabolic lability of 3 is due to phase I oxidative metabo-
lism with a 63% contribution of cytochrome CYP3A4. Such liability
can be addressed by in silico models derived from 3D structure/
Figure 3. Prediction of sites of metabolism (SOM) with Metasite 5.0, based on Astex CYP3A4 + ketoconazole with reactivity correction for compound 3. Likelihood of
metabolism is indicated in a heatmap representation (small insert). The primary SOM is predicted at the benzylic position of the thiadiazole scaffold (marked as red circle).
Black fields indicate hydrophobic ligand recognition by the cytochrome complex, red fields indicate hydrogen bond acceptor recognition, such as oxygen, and blue fields
represent hydrogen bond donor recognition. Introduction of polarity into a hydrophobic region or introduction of opposite polarity is suggestive for a decrease in the
molecular recognition, potentially leading to a decrease in the rate of metabolism.14
Please cite this article in press as: Mougenot, P.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.11.046

P. Mougenot et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
3
O
N
O
heterocycle
replacement
a, b
c
O,N-introduction
S
O
HO
O
O
O
8
9
O
N
N
solubilizing moieties
salt formation
N
OH
O
ether moieties
N-introduction
O
H
d
S
O
R
O
S
N
S
N
OH
N
H
R
O
N
H
O
O
N
H
N
H
R
12
11
Figure 4. Strategy to increase solubility of the thiadiazole series.
Scheme 1. Synthesis of compound 12. Reagents and conditions: (a) (COCl)2, DCM,
(b) KSCN, CH3CN; (c) RCONHNH2 10, CH3CN, reflux; (d) TFA, four steps 30–65%.
The synthesis15 of compounds 4-cis and 5-cis is depicted in
Scheme 2 starting from the carboxylic acid 6 which is transformed
into the acid chloride with oxalyl chloride. Subsequent coupling
reactions with 2-amino-1,3,4-thiadiazoles led to the amides 7.
We also used PyBroPÒ for the synthesis of amides 7 from car-
boxylic acid 6 and 2-amino-[1,3,4]thiadiazoles. Finally, the corre-
sponding carboxylic acids 4-cis and 5-cis were obtained after
basic hydrolysis.
As seen in Table 3, cis derivatives were at least ten fold more
potent that their trans counterparts which could be explained by
a better alignment when superimposed on 1-trans as shown in
Figure 5.
ligand based methods. Identification of the causal metabolic ‘hot
spots’ in compound 3 would allow the design of improved candi-
dates. We first applied the MetaSite13 prediction tool that includes
3D models of the most important human cytochromes, including
CYP3A4, and allows recognition of labile sites of metabolism
(SOM). Since CYP3A4 was the main contributor to its metabolism
we assessed compound 3 in a CYP3A4 model from a ketoconazole
co-crystal, with built-in corrections for reactivity of the enzyme
and ligand. The predicted primary sites of metabolism were
located on the cyclopentylethyl fragment (Fig. 3).
In accordance with MetaSite predictions, we introduced oxygen
atoms at different positions of the cyclopentylethyl moiety to block
oxidative metabolism and to assess the effect on solubility in
parallel.
Unfortunately, the aqueous solubility at pH = 7.4 of compounds
4-cis and 5-cis was not dramatically improved compared to 1-trans
and 2. On the other hand, these ether analogues were found to be
very potent in enzymatic and cell-based assays. Overall, compound
5-cis displayed a favourable profile with very good potency, good
metabolic stability and intestinal permeability but only a limited
gain in solubility, which made it a good candidate for salt
screening.
In order to assess a possible synergistic effect between the
incremental improvements in solubility seen in 12a and 5-cis, we
introduced a second oxygen in compound 12a, in the hope of fur-
ther improving solubility. The synthesis15 of compound 16 is
depicted in Scheme 3 starting from the carboxylic acid 6. The suc-
cessive reactions with oxalyl chloride and potassium isothio-
cyanate gave the benzoylisothiocyanate 13. Addition of
cyclopentyloxymethylhydrazide followed by trifluoroacetic acid
treatment gave the thiadiazole 15. Finally, basic hydrolysis led to
the carboxylic acid 16 (see Table 4).
As expected, solubility of compound 16 was improved com-
pared to 12a and 5-cis but still remained too low.
Due to the insufficient improvement in solubility by the intro-
duction of oxygen atoms in the above compounds, we decided to
assess the effect of nitrogen insertion.
We first tried the replacement of the phenyl by a pyridyl in
compound 4-cis. The general synthesis15 is depicted in Scheme 4.
Starting from the ethyl hydroxynicotinic ester 19 and cis-tert-butyl
4-hydroxycyclohexanecarboxylic ester 20 in a Mitsunobu reaction
gave, after hydrolysis, the carboxylic acid 22. The subsequent
The synthesis15 of corresponding compounds 12a–c is depicted
in Scheme 1 starting from the carboxylic acid 8. The successive
reactions with oxalyl chloride and potassium isothiocyanate gave
the acylisothiocyanate 9. Addition of the hydrazide 10 followed
by trifluoroacetic acid treatment gave the thiadiazole 12.
As shown in Table 2, the introduction of an ether function on
the cyclopentylethyl group led to an increase in metabolic stability
and solubility but also to a drop in activity compared to compound
3. Compounds 12b and 12c displayed acceptable solubility up to
90 lM at pH = 7.4 but at the expense of potency. On the other
hand, compound 12a retained good potency but its solubility
was too low.
Encouraged by the favourable effect on solubility obtained with
compounds 12a–c, we considered a more aggressive strategy to
address this issue: (i) introduction of heteroatoms at other sites
of the scaffold; (ii) thiadiazole replacement by more soluble five-
membered heterocycles; (iii) introduction of solubilising groups
and (iv) salt formation (Fig. 4).
As had been shown earlier,6 the aliphatic acid appears to be
essential for maintaining the enzymatic activity. Therefore, we first
envisaged the introduction of an oxygen atom between the phenyl
and the cyclohexyl moieties whilst keeping the distance between
the thiadiazole and the carboxylic acid constant as in compounds
4-cis/trans and 5-cis/trans.
Table 2
Modulation of cyclopentyl derivatives 3, 12a–c
O
N
N
N
OH
H
3, 12
S
R
O
Compd
R
Enz. DGAT-1 IC50
(l
M)
Cell. DGAT-1 IC50
(lM)
% met. h/m/r
Caco2 (10ꢀ7 cm/s)
Sol. pH = 7.4 (lM)
3
0.019
0.038
0.012
0.066
36:5:10
3:0:0
90
<2
9
O
12a
103
O
12b
12c
0.144
0.233
ND
ND
9:8:3
ND
72
90
95
O
ND
Please cite this article in press as: Mougenot, P.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.11.046

4
P. Mougenot et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
O
O
O
O
OH
O
O
O
O
O
O
N
S
N
S
a, b
or c
d
N
N
O
6
N
H
N
H
R
R
HO
4-cis, R=H
7
5-cis, R=3,5-difluoro
Scheme 2. Synthesis of compounds 4–5. Reagents and conditions: (a) (COCl)2, DCM; (b) 5-(3,5-difluorobenzyl)-2-amino[1,3,4]thiadiazole, pyridine, CH3CN (2 steps, 86%); (c)
5-benzyl-2-amino-[1,3,4]thiadiazole, PyBroPÒ, 26%; (d) LiOHꢁH2O, THF, H2O, 56–78%.
Table 3
Compounds 4-cis and 5-cis
Compd
Enz. DGAT-1 IC50
(l
M)
Cell. DGAT-1 IC50
(lM)
% met. h/m/r
Caco2 (10ꢀ7 cm/s)
Sol. pH = 7.4 (lM)
4-cis
4-trans
5-cis
0.016
0.280
0.011
0.160
0.028
ND
0.013
0.076
3:27:22
ND
1:42:9
1:9:8
15
ND
29
30
2
ND
4
5-trans
ND
O
OH
O
OH
O
O
N
S
N
N
N
N
H
O
N
O
H
S
R
4-trans
4-cis, R=H
5-cis, R=3,5-difluoro
Figure 5. Compounds 4-cis and 4-trans. Left: The alignment16 of 4-cis with 1-trans (shown in grey) suggests a very good match of the cyclohexyl ring moiety, while the
electrostatic fields of the carboxylates strongly overlap. Right: the configuration of 4-trans leads to a deviation of the carboxylate electrostatic fields, with less overlap with 1-
trans.
coupling reaction with 2-amino-1,3,4-thiadiazoles led to the
amides 23. Finally, treatment with trifluoroacetic acid provided
compound 24.
As shown in Table 5, compound 24a displayed a significant
enhancement of solubility at pH = 7.4, but associated with a drop
in Caco2 permeability. Introduction of the chlorine substituent
led to compound 24b with an improvement of Caco2 permeability
but, unfortunately, lower solubility.
On the other hand, the introduction of a nitrogen atom close to
the thiadiazole ring gave the diaminothiadiazoles 25, synthesized
according to Scheme 2. Compound 25a showed a slight enhance-
ment of solubility at pH = 1 and pH = 7.4. Replacement of the west-
ern cyclopentyl by a phenyl moiety led to diaminothiadiazole 25b,
a potent compound with a low solubility. Both compounds had low
Caco2 permeability (see Table 6).
In view of the limited improvement in solubility through the
introduction of heteroatoms, we tried to change the thiadiazole
core to a more polar five-membered ring such as an oxadiazole.17
We synthesized the 1,3,4- and the two 1,2,4-oxadiazole regioiso-
mers18 in the phenylcyclohexyl and phenoxycyclohexyl series to
explore their solubility according to Scheme 5 below.
The 1,3,4-oxadiazole 28b showed a dramatic increase in solubil-
ity at pH = 7.4, but a significant decrease in potency compared to
the thiadiazole 28a. The 1,2,4-oxadiazole regioisomers 28c and
28d showed a significant increase in solubility at pH = 7.4 com-
pared to the thiadiazole 1-trans but a loss in potency. Finally, in
the phenoxycyclohexyl series, compounds 32a–b showed a very
high solubility, well above our internal standard for clinical candi-
dates, but their low metabolic stability in rodents prevented us
from selecting them for further in vivo studies (see Table 7).
In the absence of a co-crystal structure for DGAT-1, we turned
for further guidance to the ligand-based alignment of 4-cis with a
substrate of DGAT-1, namely coenzyme A. The latter has been
described as a competitive substrate for a synthetic ligand.19 This
finding led us to hypothesise a potential overlap of the ligand bind-
ing sites, and to closely examine shared hydrophilic and hydropho-
bic sites. Several cocrystal structures of acyltransferases have been
published so far, disclosing various coenzyme A fatty acid esters in
their bioactive conformation.20 Interestingly, many of these sub-
strates share an elongated bend structure of the peptidic moiety.
We tested alignments of 4-cis with decanoyl-coenzyme A in bioac-
tive conformation (see Fig. 6). Both compounds apparently share
key pharmacophore recognition features. The carboxylic acid of
4-cis is linked to the proximal phosphate ester, the cyclohexyl moi-
ety of 4-cis to the geminal dimethyl group of coenzyme A, while
the arylamide of 4-cis is matched to the aliphatic amide and the
thiadiazole core of 4-cis to the cleavable thioester of the substrate.
The distal benzyl moiety of 4-cis is matched to the fatty acid chain,
supporting the introduction of aromatic or aliphatic hydrophobic
substituents. If this binding mode is valid, the introduction of polar
Treatment of benzoyl chloride with sodium cyanamide fol-
lowed by hydroxylamine yielded 3-amino-1,2,4-oxadiazole 27.
Coupling with the carboxylic acid 8 and trifluoroacetic acid depro-
tection led to compound 28.
Treatment of benzyl cyanide with hydroxylamine followed by
trichloroacetyl chloride led to the 5-trichloromethyl-1,2,4-oxadia-
zole 30. Nucleophilic substitution with the carboxamide 31 in the
presence of NaH and hydrolysis led to the expected compound 32.
Please cite this article in press as: Mougenot, P.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.11.046

P. Mougenot et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
5
O
O
O
N
O
O
N
H
NH2
O
O
O
O
a, b
S
O
6
c
HO
13
O
O
O
O
O
d
O
O
S
O
O
N
S
N
N
H
N
H
O
O
N
H
N
H
15
14
O
OH
O
e
O
N
S
N
N
H
O
16
Scheme 3. Synthesis of compound 16. Reagents and conditions: (a) (COCl)2, DCM, (b) KSCN, CH3CN, (c) CH3CN, reflux, 89%, three steps, (d) TFA, (e) LiOHꢁH2O, THF, H2O, 87%,
two steps.
Table 4
Compound 16
O
OH
O
O
N
N
N
H
O
S
16
Compd
Enz. DGAT-1 IC50
0.012
(
lM)
Cell. DGAT-1 IC50
0.017
(lM)
% met. h/m/r
5:8:1
Caco2 (10ꢀ7 cm/s)
41
Sol. pH = 7.4 (
lM)
16
18
O
O
N
O
O
N
O
N
O
O
O
O
O
O
a
b
OH
O
O
HO
+
HO
20
N
19
21
22
O
N
OH
O
O
O
O
N
N
d
c
O
N
N
N
H
N
H
O
S
S
R
24a R=H
24b R=Cl
23
R
Scheme 4. Synthesis of compound 24. Reagents and conditions: (a) DIAD, Ph3P, 47%; (b) LiOHꢁH2O, THF, methanol, 98%; (c) 5-(3-benzyl)-[1,3,4]thiadiazol-2-ylamine (or
chloro derivative), PyBroPÒ, DIEA, 63% (49%); (d) TFA, DCM, 80% (90%).
Table 5
Modulation of the R moiety in pyridines 24
O
N
OH
O
O
N
S
N
N
H
24
R
Compd
R
Enz. DGAT-1 IC50
(l
M)
Cell. DGAT-1 IC50
(lM)
% met. h/m/r
Caco2 (10ꢀ7 cm/s)
Sol. pH = 7.4 (lM)
24a
24b
H
Cl
0.045
0.055
0.135
0.085
1:23:37
2:56:27
11
17
89
8
and solubilising groups is to be envisioned only at the eastern part
of the molecule towards the phosphate of coenzyme-A.
This prompted us to try to address the solubility issue with the
preparation of amide derivatives of 1-trans bearing solubilising
moieties.21
the primary amide with hydroxydimethylethyl, dihydroxypropyl
or methoxyethyl amide derivatives led to less active compounds
33b–d without any improvement in solubility. Morpholino or
dimethylamino derivatives 33e and 33f showed slight but not sig-
nificant enhancement of solubility only at pH = 1.
Therefore, although heteroatom introduction and/or thiadiazole
replacement by 1,2,4-oxadiazole lead us to identify several com-
pounds with dramatic improvements in solubility (12b, 12c, 24a,
28b, 28c, 32a and 32b) unfortunately none of these compounds
displayed an optimal overall profile suitable for clinical candidate
nomination.
Scheme 6 depicts the synthesis of compounds 3315 obtained
from the corresponding carboxylic acid 1-trans with a panel of
amines in the presence of PyBroPÒ.
As shown in Table 8, the carboxamide compound 33a retained
the capacity to inhibit DGAT-1 but did not display improved solu-
bility. Further elongation gave compounds 33b–f. Replacement of
Please cite this article in press as: Mougenot, P.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.11.046

6
P. Mougenot et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
Table 6
Modulation in aminothiadiazoles 25
O
OH
O
O
N
S
N
N
H
R
25
N
H
Compd
25a
R
Enz. DGAT-1 IC50
0.022
(lM)
Cell. DGAT-1 IC50
0.009
(l
M)
% met. h/m/r
2:3:4
Caco2 (10ꢀ7 cm/s)
Sol. pH = 1 (
lM)
Sol. pH = 7.4 (lM)
6
6
12
<2
26
<2
25b
0.004
0.003
11:8:4
O
H
N
HO
8
O
Cl
a
O
b
c
N
N
O
N
N
N
NH2
O
O
d
O
N
H
OH
O
N
26
29
27
28
O
O
O
O
O
O
H2N
OH
O
NH2
31
f
g
e
N
O
N
O
N
CCl3
30
N
H
N
h
N
OH
O
32
Scheme 5. Synthesis of compounds 28c–d and 32a–b. Reagents and conditions: (a) NaNHCN, THF, reflux. (b) HONH2ꢁHCl, pyridine, 53%, two steps. (c) Compound 8, CDI, DBU,
DMF, 100 °C, 12%. (d) TFA, DCM, 25%. (e) HONH2ꢁHCl, TEA, ethanol, reflux, 49%. (f) Cl3CCOCl, pyridine, toluene, 85 °C, 90%. (g) Compound 31, NaH, THF, 45%. (h) LiOHꢁH2O, THF,
water, 51%.
Table 7
Five-membered ring modulation
O
O
O
OH
OH
O
O
X
X
X
X
X
X
N
H
N
H
R
R
1, 28
32
X
X
X
Compd
Enz. DGAT-1 IC50
(lM)
Cell. DGAT-1 IC50
(lM)
% met. h/m/r
Caco2 (10ꢀ7 cm/s)
Sol. pH = 1 (
l
M)
Sol. pH = 7.4 (lM)
R
N
N
1-trans
28a
0.036
0.600
1.39
0.029
0.130
ND
0:5:11
ND
112
ND
ND
ND
ND
29
<2
<2
<2
<2
<2
<2
<2
S
F
F
N
N
<2
S
N
N
28b
ND
446
129
41
O
N
O
N
N
28c
0.286
0.253
0.057
ND
ND
N
O
28d
ND
ND
N
O
32a
ND
2:100:53
274
N
F
O
N
32b
0.120
0.310
1:99:61
9
<2
813
N
F
Finally, to combine optimal activity/e-ADMET profile with a sol-
ubility level appropriate for further development, we used com-
pounds 1-trans and 5-cis to perform a salt screening with a panel
of pharmaceutically acceptable bases.22
Four salts (sodium, potassium, calcium and arginine) were iden-
tified and confirmed in crystalline form for compound 1-trans. The
sodium and arginine salts were scaled-up and analysed by differ-
ential scanning calorimetry (DSC) and thermal gravimetric analysis
(TGA). Both salts showed loss of weight or structural modifications
before the melting point which made them unsuitable for further
development.
Compound 5-cis led to seven salts confirmed in crystalline
form; sodium, potassium, calcium, N-methyl glucamine, lysine,
arginine, trometamol, and choline. The potassium salt was aban-
doned due to synthesis reproducibility issues. The arginine,
trometamol, lysine, and sodium salts were scaled-up and analysed
by DSC and TGA.23 Salting out was observed in water with the
sodium salt. The arginine and trometamol salts were polymorphic.
Only the lysine salt displayed an acceptable profile for further
development.
The lysine salt of compounds 5-cis, 5-cisL,24 showed a large
increase in aqueous solubility compared to the carboxylic acid
Please cite this article in press as: Mougenot, P.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.11.046

P. Mougenot et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
7
Figure 6. Ligand-based alignment for compound 4-cis (shown in green) and
decanoyl-coenzyme A in a bioactive conformation (pdb code 4MFK). Negative
charge is represented by red shaded fields, and overlaps to a great extent between
the carboxylic acid of the ligand and the proximal phosphoric acid ester of
decanoyl-coenzyme A.
O
O
a
N
S
N
S
N
N
N
H
N
H
R
OH
33
1-trans
O
O
Figure 7. Lipid challenge assay with compound 5-cisL.
Scheme 6. Synthesis of compounds 34a–f. Reagents and conditions: (a) PyBroPÒ
,
triethylamine, amines, DMF, 7–68%.
compared to 4%) in the bioavailability of the lysine salt form
(see Table 9).
Table 8
Modulation of R moiety in amides 34a–f
Compound 5-cisL showed a good pharmacokinetic profile with a
maximal plasma concentration (Cmax) of 13.2 lM and a half-life of
7.4 h, both compatible with pharmacological studies. It also
O
N
N
N
H
R
S
34
O
showed a large tissue distribution associated with a concentration
of 14.3 lM in the small intestine, the target tissue for the lipid
challenge test (see Table 10).
Compd
R
Enz. DGAT-1
IC50 M)
Sol. pH = 1
M)
Sol. pH = 7.4
(lM)
(
l
(
l
As shown in Figure 7 and 5-cisL, administered at 0.01, 0.03, 0.1,
and 0.3 mg/kg, significantly and dose-dependently decreased
plasma TG (ED50 = 0.08 mg/kg), in an acute lipid challenge test25
1 h after oral administration.
OH
1-trans
0.036
0.048
<2
<2
<2
<2
NH2
33a
OH
OH
HN
NH
33b
0.092
<2
<2
33c
33d
33e
0.081
0.088
0.122
<2
<2
7
<2
<2
<2
In further studies, compound 5-cisL was compared to the free
carboxylic acid form in the acute lipid challenge test. Under
these conditions 5-cisL given orally at 0.3 mg/kg inhibited the
TG rise induced by a corn oil challenge by 153% (p < 0.05),
whereas the carboxylic acid inhibited the TG rise by only 91%
at the same dose. The duration of action was similar for the
two compounds with a maximal effect observed between 1
and 3 h after oral administration (ꢀ76% at 1 mg/kg at 3 h for
the carboxylic acid and ꢀ101% at 0.3 mg/kg at 3 h for the
lysine salt).
OH
O
NH
NH
O
N
NH
33f
0.279
6
<2
N
Table 9
Comparison between carboxylic acid form and lysine salt of compound 5-cis
Compd Sol. water (
l
M) Sol. pH = 7.4 (
l
M) Bioavailability in % (in rats)
In conclusion, among the different strategies used to improve
the solubility of this chemical series, the best results were
obtained by thiadiazole replacement leading to the oxadiazole
32a which unfortunately displayed insufficient metabolic
stability in rodents for in vivo evaluation. Finally, a salt screen
of best thiadiazole leads 1-trans and 5-cis provided the lysine
salt 5-cisL as a potential development candidate displaying good
aqueous solubility in water and a favourable PK/PD profile.
Evaluation of this compound in an acute lipid challenge test
showed a drop in TG at a low administered dose. On the basis
of its overall profile, compound 5-cisL was selected for
preclinical development and results will be reported elsewhere
in due course.
5-cis
5-cisL
<21
4
8
4
40
864*
*
pH measured = 9.0.
Table 10
Pharmacokinetic and distribution profile for compound 5-cisL in rats after a single
oral administration (10 mg/kg in MC/T)
Distribution Cmax (ng/ml/lM)
Plasma
Liver
Small intestine
6980/14.3
Heart
Brain
6440/13.2
121,000/247.9
2040/4.2
75.8/0.16
Plasma half life: 7.4 h; Tmax: 4 h.
Acknowledgment
form. To confirm the interest of this lysine salt form, we performed
a head to head bioavailability comparison in rats with the free
carboxylic acid and found a very significant improvement (40%
The authors would like to thank Fiona Ducrey for guidance and
proof reading this Letter.
Please cite this article in press as: Mougenot, P.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.11.046

8
P. Mougenot et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
flow of 1.5 ml/min, coupled to a flow scintillation analyzer (Perkin Elmer
References and notes
radiomatic 625T) to separate and quantify the relative amount of all 14C-
acylglycerides formed.
1. (a) Cases, S.; Smith, S.; Zheng, Y.-W.; Myers, H. M.; Lear, R. S.; Sande, E.; Novak,
S.; Collins, C.; Welch, C. B.; Lusis, A. J.; Erikson, S. K.; Farese, R. V., Jr. Proc. Natl.
Acad. Sci. U.S.A. 1998, 95, 13018; (b) Sturley, S. L.; Oelkers, P. U.S. Pat. 6,100,077.
2. Cases, S.; Stone, S. J.; Zhou, P.; Yen, E.; Tow, B.; Lardizabal, K. D.; Voelker, T.;
Farese, R. V., Jr. J. Biol. Chem. 2001, 276, 38870.
3. DeVita, R. J.; Pinto, S. J. Med. Chem. 2013, 56, 9820.
4. Notte, G. T. Ann. Rep. Med. Chem. 2013, 48, 451.
5. (a) Trialtrove database: P-7435 in Phase
10. Metabolic lability: compound is incubated 20 min with human or mouse
hepatic microsomal fraction. Metabolisation (Met.)% result corresponds to
disappearance of tested compound at the end of the experiment.
11. Caco2 permeability: (a) Grès, M.-C.; Julian, B.; Bourrié, M.; Meunier, V.; Roques,
C.; Berger, M.; Boulenc, X.; Berger, Y.; Fabre, G. Pharm. Res. 1998, 15, 726; (b)
Grandi, D. D.; Press, B. Curr. Drug Metab. 2008, 9, 893.
12. Solubility: 2 mg of compound is incubated in 1 mL of the media for 20 h with a
shaker. The media is then filtered. The solubility was determined with an UPLC
I completed with the status
discontinued. Consulted in November 2015; (b) Cortellis database: P-7435 in
Phase I with the status discontinued. The structure has been disclosed but not
in a peer-review source. Consulted in November 2015.
system using a Acquity HSS C18 (50 * 2.1 mm; 1.8 lm) column with a mobile
phase 2–100% of ACN/TFA 0.035% in H2O/TFA 0.05% in 3.5 min. Solubility at
pH = 1 or pH = 7.4 indicated an HCl 0.1 N solution or a phosphate buffer 50 mM
pH 7.4, respectively.
6. Mougenot, P.; Namane, C.; Fett, E.; Camy, F.; Dadji-Faïhun, R.; Langot, G.;
Monseau, C.; Onofri, B.; Pacquet, F.; Pascal, C.; Crespin, O.; Ben-Hassine, M.;
Ragot, J.-L.; Van-Pham, T.; Philippo, C.; Chatelain-Egger, F.; Péron, P.; Le Bail, J.-
C.; Guillot, E.; Chamiot-Clerc, P.; Chabanaud, M.-A.; Pruniaux, M.-P.; Schmidt,
F.; Venier, O.; Nicolaï, E.; Viviani, F. Bioorg. Med. Chem. Lett. 2012, 22, 2497.
7. Zimmermann, S.; Zarse, K.; Schulz, T. J.; Siems, K.; Müller-Kuhrt, L.; Birringer,
M.; Ristow, M. Horm. Metab. Res. 2008, 40, 29.
13. Cruciani, G.; Carosati, E.; De Boeck, B.; Ethirajulu, K.; Mackie, C.; Howe, T.;
Vianello, R. J. Med. Chem. 2005, 48, 6970.
14. Baringhaus, K.-H.; Hessler, G.; Matter, H.; Schmidt, F. In Chemoinformatics for
Drug Discovery; Jürgen Bajorath, Ed.; Wiley, 2013.
15. Fett, E.; Mougenot, P.; Namane, C.; Nicolaï, E.; Philippo, C. WO2010/086551.
16. For generating
a
high-quality 3D alignment we employed our internal
combinatorial analysis of ROCS 3D
8. Enzymatic assay: (a) compound activity (0.000038–10 lM, final concentration)
application MARS, which performs
a
was measured in a phase-partition based assay using a recombinant human
DGAT-1 enzyme (Kristie, Anal. Biochem. 2006, 358). Briefly, the assay was
shape-based alignments. The computational procedure has been reported in a
previous publication.6 The best ranked alignment—as determined by the
maximum sum of scores—is being reported.
performed in 96-well Isoplate, in a 50
250 mM sucrose, 150 1,2-di-(cis-9-octadecenoyl)-sn-glycerol, 40
3H-octanoyl-CoA (1 Ci/ml), 10 mM MgCl2 and 0.25 of microsomal
ll final volume, with 100 mM Hepes,
lM
lM
17. Boström, J.; Hogner, A.; Llinàs, A.; Wellner, E.; Plowright, A. T. J. Med. Chem.
2012, 55, 1817.
18. Fett, E.; Mougenot, P.; Namane, C.; Nicolaï, E.; Philippo, C. WO2012/011081.
19. Cao, J.; Zhou, Y.; Peng, H.; Huang, X.; Stahler, S.; Suri, V.; Qadri, A.; Gareski, T.;
Jones, J.; Hahm, S.; Perreault, M.; McKew, J.; Shi, M.; Xu, X.; Tobin, J. F.; Gimeno,
R. E. J. Biol. Chem. 2011, 286, 41838.
20. The protein data bank PDB: http://www.rcsb.org.
21. Wermuth, C. G. The Practice of Medicinal Chemistry, 3rd ed.; Academic Press,
2008. Chapter 38, p 767.
l
lg
proteins from Sf9 insect cells overexpressing human DGAT1. The reaction
was initiated by the addition of DGAT1 microsomes and carried out for 60 min
at 25 °C, then stopped with 10
l
l of acetic acid (3.3% final concentration). After
l of BetaPlate Scint
45 min of further incubation at room temperature, 60
l
cocktail (Perkin Elmer) were added in order to partition lipids from the
aqueous phase. Radioactivity of the upper organic phase was determined using
a MicroBeta counter (Perkin Elmer). As confirmation of the screening, IC50s
were assessed through dose–response in monoplicate. Each run was
22. Panel of bases: sodium hydroxide, potassium hydroxide, calcium hydroxide, N-
validated with
a
reference compound JT-553 described in litterature8b,8c
methylglucamine, L-lysine, L-arginine, trometamol, betaine, ammonia, choline.
(IC50—28 nM 3.1, n > 100); (b) Fox, B. M.; Furukawa, N.; Hao, X.; Iio, K.;
Inaba, T.; Jackson, S. M.; Kayser, F.; Labelle, M.; Li, K.; Matsui, T.; Mcminn, D. L.;
Ogawa, N.; Rubenstein, S. M.; Sagawa, S.; Sugimoto, K.; Suzuki, M.; Tanaka, M.;
Ye, G.; Yoshida, A.; Zhang, J. WO2004/047755; (c) Dow, R. L.; Andrews, M.;
Aspnes, G. E.; Balan, G.; Gibbs, E. M.; Guzman-Perez, A.; Karki, K.; LaPerle, J. L.;
Li, J.-C.; Litchfield, J.; Munchhof, M. J.; Perreault, C.; Patel, L. Bioorg. Med. Chem.
Lett. 2011, 21, 6122.
23. DSC: TA INSTRUMENTS Q200 under nitrogen at 50 ml/mn flow and a rate of
10 °C/min. TGA: TA INSTRUMENTS: TGA Q500 under nitrogen with a rate of
10 °C/min.
24. Procedure: A solution of
L
-lysine (0.34 g, 2.32 mmol) in 21 mL of a 4:1 mixture
(1 g, 2.11 mmol) in
MeOH/water was added to the carboxylic acid
5
tetrahydrofuran (70 mL). The mixture was heated at 80 °C within 2 h. After
cooling, the precipitate was filtered, washed successively with EtOH/water and
twice with EtOH. After drying in vacuum for 48 h, 0.842 g of a white-off solid
was obtained.
9. Cellular assay: compound activity was measured in vitro using Chang liver
cells. Briefly, the day before the assay, cells were seeded in 24-well plates at
1.8 ꢂ 105 cells per well. The cells were starved overnight in a DMEM 4.5 g/L
glucose medium supplemented with 2% Oleic Acid–Albumin complex.
Thereafter, compounds were dispensed into respective wells and the plates
were incubated 30 min at 37 °C, 5% CO2. Then, the substrate (Glycerol [14C]
25. Male C57Bl/6J mice (22–24 g), previously fed ad libitum with a standard
laboratory diet and tap water, are fasted 17 h before the treatment. On the day
of experiment, mice are treated orally (10 mL/kg) with vehicle (water) or
compound at 0.01, 0.03, 0.1 and 0.3 mg/kg prepared in vehicle (n = 10 mice per
group) eighteen to one hour before the lipid challenge. At T0, mice are treated
with 5 mL/kg of corn oil or with 5 mL/kg of water. One hour after corn oil
treatment, blood collection is performed (retro-orbital sinus puncture) under
isoflurane anesthesia for assessment of plasma triglycerides. ED50 values are
calculated using BioSt@t-Speed-LTS 2.0. Statistical analysis are performed
using SASÒ V8.2 with the Everstat interface: One-way-Anova followed by a
Kruskal Wallis test. The significance level is set to 0.05.
0.4 lCi/ml final) was added into each well and the plates were incubated 6 h at
37 °C, 5% CO2. Incubation medium was removed and the cells were washed
twice with PBS. After trypsination, cells were centrifuged for 5 min at 1300ꢂg
and the supernatant was discarded. Lipids were extracted from cell pellets by
adding 400
samples were sonicated and filtered (0.45
HPLC (Waters 2695 with a SunFireÒ column C18 3
mobile phase 5% (H2O + 0.1% TFA), 70% methanol, 25% dichloromethane with a
l
l of a methanol/dichloromethane/TFA (50:50:0.1%) solution. The
m). The samples were analysed by
m 4.6 ꢂ 75 mm), with a
l
l
Please cite this article in press as: Mougenot, P.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.11.046