1H-Imidazo[4,5-c]pyridine-4-carbonitrile as cathepsin S inhibitors: Separation of desired cellular activity from undesired tissue accumulation through optimization of basic nitrogen pka

Article abstract of DOI:10.1016/j.bmcl.2010.12.065

Based on the theoretical understanding of the in vivo lysosomotropism, by adjusting the pk_a of basic nitrogen containing cathepsin S inhibitors, a set of compounds with pk_a 6-8 were identified to have excellent cell based Lip10 activ

Full text of DOI:10.1016/j.bmcl.2010.12.065

Bioorganic & Medicinal Chemistry Letters 21 (2011) 932–935
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
1H-Imidazo[4,5-c]pyridine-4-carbonitrile as cathepsin S inhibitors: Separation
of desired cellular activity from undesired tissue accumulation through
optimization of basic nitrogen pka
⇑
Wullie Arbuckle, Mark Baugh, Simone Belshaw, D. Jonathan Bennett, John Bruin, Jiaqiang Cai ,
Kenneth S. Cameron, Chris Claxton, Maureen Dempster, Kathryn Everett, Xavier Fradera, William Hamilton,
Philip S. Jones, Emma Kinghorn, Clive Long, Iain Martin, John Robinson, Paul Westwood
Merck Research laboratories, MSD, Lanarkshire ML1 5SH, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Based on the theoretical understanding of the in vivo lysosomotropism, by adjusting the pka of basic
nitrogen containing cathepsin S inhibitors, a set of compounds with pka 6–8 were identified to have
excellent cell based Lip10 activity, yet avoiding undesired sequestration in spleen.
Ó 2011 Elsevier Ltd. All rights reserved.
Received 14 October 2010
Revised 14 December 2010
Accepted 15 December 2010
Available online 19 December 2010
Keywords:
Cathepsin s inhibitor
Lysosomotropism
Tissue sequestration
pka
There are eleven members of the cathepsin family cysteine pro-
teases identified in the human genome (cathepsins B, C, H, F, K, L,
O, S, V, W and X).1 In contrast to most of the cathepsins which are
only function in an acidic pH environment, cathepsin S is active in
both acidic and neutral conditions, and it has a dual functionality,
that is, extracellular matrix degradation and intracellular invariant
chain processing.2 As we are interested in cathepsin S as a target to
treat autoimmune disorders by modulating cathepsin S activity in-
side the low pH lysosomes, good intracellular activity measured by
the accumulation of the natural substrate of cathepsin S, Lip10 in
human JY cell3 is critically important. In our previous publica-
tions3–5 we have demonstrated the importance of basic nitrogen
for the cell based Lip10 activity through so-called lysosomotro-
pism6,7 as shown by our previous examples in Figure 1. Although
these basic nitrogen containing compounds, such as compounds
2–4, have good activity in the cell based Lip10 assays, they were
found to have very high plasma clearance and high distribution
volume. These compounds are highly sequestered in spleen, liver
and lung with spleen level as much as 50 times higher than in plas-
ma for compound 3. In response to concerns over safety relating to
tissue accumulation, we made a strategic decision to rule out these
highly sequestered compounds for further development.
The lysosomotropism can be quantitatively described6–8 appendix
by Eq. 1 for monobasic compounds where F is the ratio of total
compound concentration (including all protein, lipid and tissue
bound compound treated as homogeneous and other unbound
compound in solution) between inside the lysosome and its
surrounding medium; pHM is the pH of the surrounding medium
(taken as 7.4) and pHL is pH inside the lysosome (taken as
4.8);6,7 ffM is the free fraction in the medium and ffL is the free
CN
CN
N
N
N
N
N
N
F3C
F3C
O
N
N
O
O
2 (pka>9)
1 (neutral)
hCat S IC50: 41nM
lip10 IC50: 84nM
hCat S IC50: 7.8nM
lip10 IC50: 1420nM
CN
CN
N
N
N
N
N
F3C
F3C
N
HN
N
O
O
N
4 (pka>9)
3 (pka 8-9 and 5-7)
hCat S IC50: 8.3nM
lip10 IC50: 9nM
hCat S IC50: 7.2nM
lip10 IC50: 63nM
⇑
Corresponding author. Tel.: +44 01419467424; fax: +44 01698736187.
E-mail address: jiaqiang.cai@yahoo.co.uk (J. Cai).
Figure 1.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.12.065

W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 932–935
933
fraction in the lysosome. When only unbound compound concen-
tration (both protonated and unprotonated) is considered, the
compound concentration ratio (f) between lysosome and the med-
ium can be simplified as Eq. 2. When f is plotted against pka, a titra-
tion curve as shown by Figure 2 can be obtained. Based on this
equation, the theoretical maximum concentration ratio for a
monobasic nitrogen containing compound in solution is ca. 400.
As lysosomes occupy only a very small portion of the tissue
(0.004–0.0068 cm3/g estimated for liver tissue6), the maximum
lysosomotropism should only account for less than three-fold extra
compound level (i.e., 400 ꢀ 0.0068 = 2.76). The observed high
sequestration ratio of 25–50 is probably due to other tissue bind-
ings, most likely phospholipid binding in the described cases. As
we are targeting lysosome cathepsin S, the lysosomotropism
should be considered as a positive factor. This lysosomotropism
also seems essential for achieving desired cellular functional activ-
ity against Lip10 degradation. It therefore becomes a formidable
challenge to obtain a compound which possesses desired cellular
Lip10 activity without undesired organ accumulation.
For three reasons compound 4 was selected for exploring the
pka and in vitro Lip10 relationship. The first is that compound 4
is the most active from the series and it has excellent microsome
stability across all species examined (SD rat, C57 mouse and
human). Secondly, structural biology work indicated that an
attachment to the piperidine nitrogen is exposed to solvent outside
the S2 pocket and should have relatively little impact on the
human cathepsin S inhibitory activity, facilitating future data
interpretation. Thirdly, a wide range of pka for the piperidine nitro-
gen could be easily achievable by judicious choice of substituents
with different electronic properties.
Compounds for these studies were synthesized according to
Scheme 1 by simple alkylation or acylation of the piperidine nitro-
gen of compound 4. Results of these studies are shown in Table 2.
The two most basic compounds 4 and 5a share a similar activity
against purified human cathepsin S enzyme and also in cell based
Lip10 assays. Attaching
a N,N-dimethylaminocarbonylmethyl
group to the piperidine nitrogen (5b) reduces pka to 7.5. Compound
5b is not only highly active against purified human cathepsin S, it
also has high activity in the cellular assay with a Lip10 IC50 of
22 nM. In order to reduce the pka further to the region of 6–7, by
using our in-house pka calculation tools,10 it was found that a het-
eroarylmethyl group could affect the piperidine nitrogen pka
significantly. By arranging the heteroatoms O, N and S inside the
five-membered heteroaryl ring, it looked possible to achieve the
required pka range based on the calculations. A selected set of four
compounds 5c–5f were synthesized and the results are shown in
10pH
10pK þ 10pH
ffM
ffL
M
a
L
F ¼
ꢀ
ꢀ
ð1Þ
10pH
10pK þ 10pH
L
a
M
10pH
10pH
10pK þ 10pH
107:4 10pK þ 104:8
M
a
L
a
f ¼
ꢀ
¼
ꢀ
ð2Þ
10pK þ 10pH
104:8 10pK þ 107:4
L
a
M
a
In this paper we will report our effort in designing such com-
pounds by selecting a narrow pka window from the theoretical
understanding of the lysosomotropism.
CN
CN
N
N
N
N
N
Expansion of the Figure 2 graph between pka 5.5–8.5 is shown
in Table 1; compounds with pka 6–8 should have ꢁ16 to ꢁ300-fold
(f) higher level in lysosome than other neutral compartments. In
theory it should be possible to design compounds with pka in this
range to have reasonably good cellular Lip10 activity. Compounds
with pka in the lower end of 6–8 should be expected to have rela-
tively low phospholipid binding as judged by their in vivo distribu-
tion volume (Vss) and should then have low sequestration level in
organs such as spleen, liver and lung. On top of what was discussed
above, these low pka (6–8) compounds will mostly be positively
charged in the low pH (4.5–5) lysosomal environment with a low
log D and should have lower binding affinities towards proteins
and lipids than the corresponding uncharged neutral spices, and
as such a higher free fraction (ffL) level for target engagement.
a
N
F3C
O
5a-5h
R
F3C
O
N
HN
4
Scheme 1. Reagents and conditions: (a) RCl, DIPEA, NMP, rt, 16 h.
Table 2
Inhibitory activity against both purified cathepsin S enzymea and natural substrate in
cell based assaysa and measured pka
b
c
Compds
R–N
IC50 (nM)
Lip10 (nM)
pka
hCatS
41
7.2
hCatK
2
3
4
776
331
178
na
84
63
9
>9 (c)
8–9 (c)
>9 (c)
>9 (c)
HN
MeN
O
8.3
7.9
500
400
300
5a
10
N
5b
5c
5d
8.3
3.3
407
288
708
22
12
39
7.5 (m)
7.9 (m)
6.3 (m)
N
N
O
N
f
f
N
O
200
100
0
N
7.9
1.6d
14.1
18.6
N
N
N
O
N
5e
5f
813
123
294
5.6 (m)
5.3 (m)
N
N O
O
0
5
10
15
pka
1148
N
O
N
O
Figure 2. f plotted against pka.
5g
5h
13
27
1905
2344
235
353
Neutral
Neutral
S
a
b
c
Table 1
relationship between f and pka
For assay conditions see Ref. 3.
All compounds have IC50 of >10
pKa of conjugated base, values were determined using a Sirius GLpKa potenti-
l
M for human cathepsins B, L and V.
pka
5.5
6
6
6.5
45
7
113
7.5
221
8
318
8.5
368
ometric titrator9; m = measured; c = calculated.10
f
16.2
d
IC50 for mouse cathepsin s.

934
W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 932–935
Table 3
containing cathepsin S inhibitors, a set of compounds with pka 6–8
were identified to have excellent cell based Lip10 activity, yet
avoiding undesired accumulation in spleen.
Total compound level in plasma and spleen measured at 2 h time point after 30 mg/kg
po dosing (C57BI/6 mice)
Compds
Total plasma (nM)
Total spleen (nM)
Spleen/plasma ratio
2
3
94 (4 h)
90
2433 (4 h)
4570
26
51
Appendix
5b
5c
5d
5e
5f
1031
3400
6610
6369
5664
11,738
19,567
12,502
18,165
4087
11.4
5.8
1.9
2.9
0.7
Consider lysosome and its surrounding medium as a two com-
partment system as shown in Figure 3, where:
[M]: free unprotonated compound concentration in the medium,
[MH+]: unbound protonated compound concentration in the
medium,
Compounds were dosed as HCl salt. 18–25 g C57Bl/6 mice were dosed by gavage
with 30 mg/kg compound (0.5% gelatin, 5% mannitol in water). Blood and spleen
were collected at 2 h time point after dosing and compound level measured.
[Mbound]: all other bound compound concentration in the med-
ium considered as homogeneous,
[HM+]: medium proton concentration,
Table 2. The pka of these compounds were measured and the values
are in the expected range from 5 to 8. It is interesting to see that
there is a relationship between pka and cathepsin S inhibitory po-
tency, the higher the pka, the higher the cathepsin S inhibitory po-
tency (or lower IC50). There seems to be very little difference
between low pka compound (5f) and the neutral analogs 5g and
5h11 in terms of both biochemical and cellular potency. Although
the cellular activities (Lip10) share a parallel relationship with
pka as IC50s do, the gap between Lip10 and IC50 becomes bigger
when pka is lower, for example, Lip10/IC50 ratio is nearly 16 for
compound 5f and it is only 3.6 for the higher pka compound 5c.
These differences, although relatively small, could be easily ex-
plained by the accumulation factor f. Compounds 5b–5f, together
with compounds 2 and 3 were assessed in an in vivo spleen
sequestration assay using C57BI/6 mice. These sequestration re-
sults are shown in Table 3. As the data shows, all the compounds
with pka lower than 7 do not have significant sequestration in
the mouse spleen. For compounds with pka 7–8, such as 5b and
5c, the ratio between spleen and plasma is also quite low (<12)
as compared with the higher pka analogs 2 and 3.
[m]: free unprotonated compound concentration in the
lysosome,
[mH+]: unbound protonated compound concentration in the
lysosome,
[mbound]: all other bound compound concentration in the lyso-
some considered as homogeneous, and
[HL+]: lysosome proton concentration.
As lysosmal membrane should prevent all species moving in or
out lysosomes except free uncharged molecules and at the steady-
state, this uncharged molecule will have equal concentration
across the two compartments, that is, [M] = [m]
Total medium concentration Mt:
½Mtꢂ ¼ ½Mꢂ þ ½MHþꢂ þ ½Mbound
ꢂ
¼ ½Mꢂ þ ½MHþꢂ þ ½Mtꢂð1 ꢃ ½ffMꢂÞ
½Mꢂ þ ½MHþꢂ
½Mtꢂ ¼
½ffMꢂ
Total lysosome concentration [mt]:
½mꢂ þ ½mHþꢂ
Compounds 5b and 5d were further evaluated for full pharma-
cokinetic properties in male C57 mice, Sprague Dawley rats and
Cynomolgus monkeys. The results are shown in Table 4. Although
their rat and monkey oral bioavailabilities are less than 10%, both
compounds showed excellent mouse PK profiles with oral bioava-
ilabilities at 62% and 112% for 5b and 5d, respectively.
Considering compound 5d is also highly active against mouse
cathepsin S (1.6 nM, Table 2), it was further assessed for its
in vivo cathepsin S inhibition and its efficacy in collagen induced
arthritis model (CIA) dosed in a therapeutically relevant setting.
The full results of these in vivo studies will be published in a sep-
arate account.
½mtꢂ ¼
½ffLꢂ
½mtꢂ ½mꢂ þ ½mHþꢂ
½ffMꢂ
½ffLꢂ
F ¼
¼
ꢀ
½Mꢂ þ ½MHþꢂ
½Mtꢂ
½Mꢂ½HþMꢂ
½mꢂ½HþL
With ½MHþꢂ ¼
, ½mHþꢂ ¼
ꢂ, and ½Mꢂ ¼ ½mꢂ at the
Ka
Ka
equilibrium, then:
½mtꢂ
½Mtꢂ
Ka þ ½HþL ꢂ
½ffMꢂ
F ¼
¼
ꢀ
Ka þ ½HþMꢂ
½ffLꢂ
To replace Ka with more widely used pKa and [H+] with pH with
In summary, based on the theoretical understanding of the
in vivo lysosomotropism and by adjusting the pka of basic nitrogen
a ; ½HþL ꢂ ¼
and ½HþMꢂ ¼
then
1
1
1
Ka
¼
10pK
10pH
10pH
L
M
10pH
10pH
10pK þ 10pH
ffM
ffL
a
M
L
F ¼
and
f ¼
ꢀ
ꢀ
Table 4
10pK þ 10pH
a
L
M
Pharmacokinetic parameters (iv 2 mg/kg; po 10 mg/kg, gelatin manitol)
Compds
Cl (mL/min/kg)
Vss (L/kg)
Cmax (ng/mL)
F (%)
10pH
10pH
10pK þ 10pH
a
M
L
5b
C57
SD
C57
Cyno
SD
79.4
93.5
13.9
10
6.17
9.8
1.48
0.6
295.5
13.37
3160
234
62
½mtꢂ½ffLꢂ
½Mtꢂ½ffMꢂ
¼
ꢀ
8.93
112
9
10pK þ 10pH
a
L
M
5d
Based on this quantitative model, high pka compounds with
high phospholipid binding could be highly sequestered in high
phospholipids containing organs, such as spleen and lung where
free fraction level (ff) will be low.
73.8
2.43
9.64
3.6
Lysosome:
Medium:
[m], [mH+],
[M], [MH+],
References and notes
[HL+], [mbound
]
[HM+], [Mbound
]
1. Gupta, S.; Singh, R. K.; Dastidar, S.; Ray, A. Expert Opin. Ther. Targets 2008, 12,
291.
2. Leroy, V.; Thurairatnam, S. Expert Opin. Ther. Pat. 2004, 14, 301.
Figure 3.

W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 932–935
935
3. Cai, J.; Baugh, M.; Black, D.; Long, C.; Bennett, J.; Dempster, M.; Fradera, X.;
Gillespie, J.; Andrews, F.; Boucharens, S.; Bruin, J.; Cameron, K. S.; Cumming, I.;
Hamilton, W.; Jones, P. S.; Kaptein, A.; Kinghorn, E.; Maidment, M.; Martin, I.;
Mitchell, A.; Rankovic, Z.; Robinson, J.; Scullion, P.; Uitdehaag, J. C. M.; Vink, P.;
Westwood, P.; van Zeeland, M.; van Berkom, L.; Bastiani, M.; Meulemans, T.
Bioorg. Med. Chem. Lett. 2010, 20, 4350.
4. Cai, J.; Bennett, J.; Rankovic, Z.; Dempster, M.; Fradera, X.; Gillespie, J.;
Cumming, I.; Finlay, W.; Baugh, M.; Boucharens, S.; Bruin, J.; Cameron, K.;
Hamilton, W.; Kerr, J.; Kinghorn, E.; McGarry, G.; Robinson, J.; Scullion, P.;
Uitdehaag, J. C. M.; van Zeeland, M.; Potin, D.; Saniere, L.; Fouquet, A.;
Chevallier, F.; Deronzier, H.; Dorleans, C.; Nicolai, E. Bioorg. Med. Chem. Lett.
2010, 20, 4447.
W.; Kinghorn, E.; Long, C.; Uitdehaag, J. C. M. Bioorg. Med. Chem. Lett. 2010, 20,
4507.
6. de Duve, C.; de Barsy, T.; Poole, B.; Trouet, A.; Tulkens, P.; van Hoof, F. Biochem.
Pharmacol. 1974, 23, 2495.
7. MacIntyre, A.; Cutler, D. J. Biopharm. Drug Dispos. 1988, 9, 513.
8. Black, W. C.; Percival, M. D. ChemBioChem 2006, 7, 1525.
9. Takacs-Novac, K.; Box, K. J.; Avdeef, A. Int. J. Pharm. 1997, 151, 235.
10. ACD PhysChem Batch, version 4.76; Advanced Chemistry Development,
Toronto, ON, Canada.
11. 1H-imidazo[4,5-c]pyridine-4-carbonitrile core has
a calculated pka of 1.8.
Compounds with this pka should be unprotonated inside lysosomes, it is then
regarded as neutral.
5. Cai, J.; Fradera, X.; van Zeeland, M.; Dempster, M.; Cameron, K.; Bennett, J.;
Robinson, J.; Popplestone, L.; Baugh, M.; Westwood, P.; Bruin, J.; Hamilton,