L. Silpa et al. / Bioorg. Med. Chem. Lett. 26 (2016) 114–120
115
to facilitate screening. To perform this assay, an Eimeria tenella
strain expressing the yellow fluorescent protein (YFP+) was used,
previously produced in our laboratory. The effectiveness of the
compounds was assessed on the basis of invasion of epithelial
MDBK cells by the sporozoites. Calculating the number of infected
cells in relation to the number of total cells determined the per-
centage of invasion. The percentage was greatly reduced in the
presence of an effective compound. Compounds that reduced inva-
sion levels to 20% compared to control wells were considered to be
potential leads. Five of the compounds tested possessing different
cores were identified as having modest activity, with IC50 ranging
1. Moreover, as the CF3 moiety is known to influence metabolic sta-
bility, binding activity and lipophilicity,19,20 we assumed that the
introduction of such a group would be likely to have a valuable
effect and improve biological activity. We report here the synthesis
of several pyrido[1,2-a]pyrimidin-2-ones 3 using various synthe-
sized (2a–n)21 or commercially available (2o–z) 2-aminopyridines
as starting material (Table 1).
Through these cyclization reactions, we investigated the reac-
tivity of 2-aminopyridines substituted with different R1, R2, R3
and R4 groups when exposed to ethyl 4,4,4-trifluorobutynoate 1.
Mono- and polyhalogenated products were easily synthesized
(entries 1–14), and the corresponding products 3a–3n were
obtained in good to excellent yields, ranging from 41% to 95%.
However, in the case of 2-aminopyridines substituted at R4 with
a halogen atom such as 2o–2q (entries 15–17), the reaction did
not proceed and only the starting substrate was recovered. The
strongly deactivating nitro substituent, led to modest yields
(entries 18 and 19). Similarly, reactions in the presence of 2-
aminopyridines bearing ester (entries 20–22), aldehyde (entry
23) and nitrile functions (entry 24) or a trifluoromethyl group
(entry 25) afforded products 3t–y isolated at acceptable yields
(45–75%). Finally, no cyclization product was obtained in the case
of 2-aminonicotinic acid 2z, (entry 26).
Other structural modifications of the substituents on the pyri-
dine ring were then envisaged, starting from halogenated trifluo-
romethylpyrido[1,2-a]pyrimidin-2-ones 3a–h. We planned to
introduce various aromatic rings by exposing these cores to
Suzuki–Miyaura cross-coupling conditions. Unfortunately only
the starting material was recovered when testing several coupling
conditions, despite the use of high temperatures and long reaction
times. The synthesis of analogs such as 5a–g was finally possible
starting from the corresponding arylated 2-aminopyridines, and
the methodology described by Liu et al.22 allowed us to prepare
4a–g. This efficient method was performed with ethylene glycol
using Pd(PPh3)4 as catalyst in the absence of ligand in short reac-
tion times under air. The reactions of 2a, 2c and 2h in the presence
of aryl boronic acids all provided 4a–g in excellent yields. These
coupling reactions were followed by the same cyclization step with
a twofold excess of 1. They afforded good overall yields of arylated
substituted pyrido[1,2-a]pyrimidin-2-ones 5a–g (Scheme 3).
The series of synthesized molecules with varying substituents
was subjected to in vitro assays to evaluate the activity of the com-
pounds. Each well seeded with MDBK cells was treated with a
defined concentration of each compound ranging from 0 to
from 10 to 40
lM (Scheme 1). The dose-dependent efficacy of
the biologically active 3g was good (IC50 = 15
l
M), with an absence
of cytotoxicity toward the MDBK cells. Moreover, 3g was synthe-
sized in a very straightforward manner and was found to be a good
candidate for SAR analysis. It is of note that our lead compound
possessed structural similarities (halogen atoms, heteroaromatic
ring) to Halofuginone (StenorolÒ), an antiprotozoal drug originally
synthesized in the late 1960s. Halofuginone was first used as an
antimalarial and then as an anticoccidial feed additive for broilers.
The antiprotozoal activity of Halofuginone in vitro is highly effi-
cient, with IC50 values ranging from 17 nM to 0.24 nM, according
to parasite/host cell models.13–16 Following identification of this
hit, we synthesized several analogs with the aim of establishing
preliminary structure–activity relationships allowing optimization
of the hit.
As described above, compound 3g was shown to impede the
invasion of MDBK cells by E. tenella sporozoites and was subse-
quently chosen as our lead. Preparation of this compound was
inspired by the work of Harriman et al.17, and consisted of the con-
densation of the corresponding 2-aminopyridine 2g on ethyl 4,4,4-
trifluorobut-2-ynoate 1 (Scheme 2). It is of note that our group has
previously studied the reactivity of this alkyne with other
a-
aminoazaheterocycles, such as substituted 3-aminopyridazines7
and 2-amino-1,3,4-thiadiazoles.18 All the corresponding heterocy-
cles were included in the screening, but they did not afford signif-
icant biological activity.
In order to set up a SAR starting from our lead 3g, we modified
the substituent of the pyridine core, but left the trifluo-
romethylpyrimidin-2-one core unmodified. The condensation step
of such 2-aminopyridine is only possible with activated alkyne as
50 lM. In order to study the influence of substituents on the ability
Library of 87 compounds
of the compounds to inhibit the invasion of the host cell by sporo-
zoites, the steric hindrance and electronic factors of the chosen
substituents were taken into account. Preliminary SAR was imple-
mented by observing the influence of these parameters according
to their position on the pyridine ring. All results are reported in
Table 2.
First, the C(7)-substituted derivatives were explored. C(7)-sub-
stituents with minor steric hindrance, electron-withdrawing prop-
erties or bulky phenyl, p-methoxyphenyl or p-fluorophenyl
screening
I
O
N
OH
N
O
Cl
Br
N
N
O
HN
Br
CF3
Halofuginone
3g
(entries 1–6) substituents had IC50 values higher than 50
ever, in the presence of the cyano substituent (entry 3) the IC50 was
37.7 M.
lM. How-
IC50 = 15 µM
IC50 = 0.24-17 nM
l
Scheme 1. Hit identification and validation.
Having revealed restricted tolerance for variations when using C
(7)-substituents, we then focused on compounds affording C(9)-
substituents. Unfortunately, regiochemistry was not found to
enhance activity. Indeed, former C(7)-substituents were used at C
(9) and all had IC50 values greater than 50 lM (entries 7–9). Simi-
larly, electron-deficient cores bearing small C(9)-bromine, C(9)-
N
O
NH2
R
R
F3C
COOEt
+
N
N
nitro or C(9)-methyl ester (entries 10–12) substituents were ren-
CF3
1
2
3
dered ineffective (IC50 > 50
lM). However, when testing a carbox-
Scheme 2. Strategy for the pharmacomodulation of 3.
aldehyde substituent derivative (entry 13) activity was slightly