4
Y.-J. Hu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Table 6
Physical and DMPK properties of selected compounds
Compd
2
37
40
48
hTRPA1 IC50
rTRPA1 IC50
(
l
M)
M)
0.41
3.0
>10
0.51
2.6
33
10
100
100
24
11
18
>20
44 (1A2)
+18
0.32
1.7
>10
0.49
3.4
77
100
100
100
45
6
33
0.31
0.33
1.8
(l
1.2
hTRPV1 (
lM)
ndi
ndi
LE16
0.52
3.4
0.40
4.3
ClogP
Solubilitya
(l
M)
24
8
Aq. stabilityb, pH 1 (% remaining)
Aq. stabilityb, pH 7 (% remaining)
Aq. Stabilityb, pH 10 (% remaining)
ndi
ndi
ndi
ndi
ndi
ndi
hCLintc
(l
L/min/mg)
74
9
12
>20
110
2
6
1.9 (1A2)
No TDI
+2
Human PPBd (% free)
Caco-2e (1E-6 cm/s)
Competitive CYP inhibitionf, IC50
P450 TDIg,% shift (isoform)
GSH adduct formation in HLMh
(l
M)
>20
42 (1A2) 23 (3A4)
+16, +18
57 (1A2) 31 (2D6)
+16
a
b
c
Thermodynamic solubility measured in a pH 7.4 buffer using HPLC with UV detection for quantification.
Determined by incubation in buffer at 70 °C for 24 h with monitoring by HPLC with UV detection to determine loss of parent over 24 h.
Metabolic stability performed in human liver microsomes (1 mg/mL) in presence of NADPH using test compound at 1 M.
l
d
e
Protein binding of test compound at 10
lM in human plasma, as determined by equilibrium dialysis.
Transwell assay. Apparent permeability (apical to basolateral) across a Caco-2 cell monolayer measured at pH 7.4 using test compound at 10
lM applied on the apical
side.
f
Competitive CYP450 (1A2, 2C9, 2C19, 2D6, 3A4) inhibition screening assay in human liver microsomes, using standard probe substrates and detected by LC–MS/MS.
g
CYP450 (1A2, 2C9, 2D6, 3A4) time-dependent inhibition IC50 shift screening assay in human liver microsomes at 10
in this assay if % shift is >20%; data for compounds that meet this criterion are shown.
lM test compound. TDI risk is considered significant
h
GSH adduct formation in human liver microsomes supplemented with 5 mM GSH, as detected by MS/MS using precursor ion scan. Result shows the molecular weight
added to the parent-glutathione adduct of the primary metabolite.
i
Not determined.
Also, if such a mechanism was important for the antagonist activ-
N
N
O
O
ity, then we would expect that compounds like 35 would be com-
pletely inactive due to lack of a nitrogen adjacent to the phenol
linkage. And again, compound 49 has the same electrophilic site
as 1 yet is inactive.
N
F
N
O
F
O
F
Selected compounds in this series have been further character-
ized for their in vitro DMPK properties (Table 6). These compounds
have moderate to high levels of plasma protein binding and fast
rates of human microsomal clearance. While most compounds
are cell permeable (except for 48), all compounds show potential
for metabolic activation since they form detectable GSH adducts
upon incubation in human liver microsomes. Furthermore, most
compounds show P450 time-dependent inhibition. Further optimi-
zation of this series is needed to attain more favorable DMPK
properties.
TRPA1 remains a very important therapeutic target for pain and
other indications. Most of the chemotype classes that have been
described so far are highly lipophilic and have low solubility. In
addition, many are particularly selective for the human form and
much weaker against mouse and rat, thus complicating in vivo
translational studies. The series of antagonists described here have
high ligand efficiency, improved physical properties, and retain
(albeit weakened) activity at the rat channel. Consequently, they
may serve as leads for further optimization.
F
48
IC50 0.33 µM
49
IC50 >32 µM
The most potent compounds to emerge were further evaluated
and the results are summarized in Table 6. Their antagonist poten-
cies range from IC50 0.31 to 0.41 lM, and all have particularly high
ligand efficiency (P0.40), moderate solubility (except for 48), and
low lipophilicity (except for 48). To assess selectivity we tested
many compounds against TRPV1; all compounds tested across this
series appeared inactive against TRPV1 (IC50 >10 lM). Species-
specific differences are well known for TRPA1; antagonists that
are potent at the human channel are typically much weaker at
the mouse and/or rat homologs. This is also true for the series pre-
sented here, where we see a 4–7 fold reduction in potency against
rat.
A concern for this series is that antagonist activity could poten-
tially be due to pseudo-reversible or irreversible mechanisms. It is
known that most TRPA1 agonists are highly electrophilic and react
with a thiol cysteine in the channel.8,14 Thus it is possible that the
requisite para-fluorophenyl ring serves to react with nucleophilic
groups on the channel. However, several lines of evidence suggest
that this is not the case here. First, the compounds are very stable
in aqueous solution both under neutral and basic conditions.
Second, compound 3 which lacks fluorine substitution) maintains
Supplementary data
Supplementary data (representative experimental conditions
for the synthesis of compounds (2, 37, 40 and 48) associated with
significant, albeit weaker, antagonist potency (IC50 3.2
lM). Fur-
thermore, compound 49, the tri-F analog of 1 had no detectable
References and notes
activity (IC50 >32
reactivity.
lM) despite also being subject to nucelophilic
We also considered that nucleophiles could potentially react at
the heteroaryl core (with liberation of the phenol); however, we do
not detect any released phenol on incubation with glutathione.