Novel dual inhibitors of calpain and lipid peroxidation with enhanced cellular activity

Article abstract of DOI:10.1016/j.bmcl.2005.12.036

A series of dipeptides with dual inhibitory activities on calpain and lipid peroxidation were prepared to target the intracellular calpain. This optimization program focused on the variations of the linker and the N-terminal amino acid of the peptidic cor

Full text of DOI:10.1016/j.bmcl.2005.12.036

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1586–1589
Novel dual inhibitors of calpain and lipid peroxidation
with enhanced cellular activity
b
a
a
b
Serge Auvin,a, Bernadette Pignol, Edith Navet, Morgane Troadec, Denis Carre,
*
´
c
a
b
´
Jose Camara, Dennis Bigg and Pierre-E. Chabrier
aDepartment of Medicinal Chemistry, Ipsen Research Laboratories, Institut Henri Beaufour,
5, Avenue du Canada, 91966 Les Ulis Cedex, France
bDepartment of Neurobiology, Ipsen Research Laboratories, Institut Henri Beaufour,
5, Avenue du Canada, 91966 Les Ulis Cedex, France
cDepartment of BioPharma Profiling and Structural Chemistry, Ipsen Research Laboratories, Institut Henri Beaufour,
5, Avenue du Canada, 91966 Les Ulis Cedex, France
Received 2 November 2005; revised 7 December 2005; accepted 7 December 2005
Available online 27 December 2005
Abstract—A series of dipeptides with dual inhibitory activities on calpain and lipid peroxidation were prepared to target the intra-
cellular calpain. This optimization program focused on the variations of the linker and the N-terminal amino acid of the peptidic
core. Two compounds 6d-05 and 6d-08 exhibited potent intracellular calpain inhibition. The polar surface area and the number of
rotors appeared to be critical descriptors to account for the behavior of these hybrid molecules in the cellular calpain assay.
Ó 2005 Elsevier Ltd. All rights reserved.
Calpains are a superfamily of Ca2+ dependent cysteine
proteases implicated in the modulation of cellular func-
tion. Calpains are present in almost all human cells and
have been implicated and extensively studied in numer-
ous central or peripheral degenerative diseases.1 Most of
these pathophysiological situations are associated with
inflammation and an enhanced formation of free radi-
cals. Such high levels of free radicals have deleterious
consequences for cells. In previous experiments, we dem-
onstrated that the combination of an antioxidant and a
calpain inhibitor protected glial cells from death in a
synergistic manner.2
Previously, we reported the discovery and synthesis of a
series of potent dual inhibitors of lipid peroxidation and
cellular calpain.4 In continuation of this program on hy-
brid molecules, we herein describe recent achievements
in this field. The antioxidant properties of our molecules
originated from phenothiazine and the calpain pharma-
cophore from the 2-hydroxy-tetrahydrofuran group. This
masked aldehyde is thought to form a hemithioketal with
the cysteine residue at the active site of the enzyme and
therefore acts via reversible calpain inhibition.4
In parallel, a theoretical calculation approach was set up
aimed at the identification of molecular descriptors
which best describe the cellular activity of these
molecules.
In the literature, extremely potent calpain inhibitors
have been disclosed. Nevertheless, because of the intra-
cellular calpain location, these inhibitors do not perform
well in cellular inhibition assays.3 Optimization criteria
stemming from the SAR studies seldom match the
prerequisites for intracellular activity. Thus, the design
of powerful, cell penetrating, inhibitors remains a
challenge.
This series of molecules was accessible through a
straightforward four-step synthetic pathway (Scheme
1). The starting aminoacid derivative 25 was prepared
in three steps from commercially available Cbz-L-leucine
and (S)-a-amino-c-butyrolactone. The chain elongation
of 2 with Cbz-protected aminoacids was performed with
HBTU as the coupling reagent and was followed by
hydrogenolysis of the N-protecting group to yield the
primary amines 4. Coupling with either the alkyl car-
boxylic acid derivatives of phenothiazine, 1a-CO2H,6
Keywords: Calpain; Antioxidant; Cellular activity.
*
Corresponding author. Tel.: +33 16092 2003; fax: +33 16907 3802;
e-mail: serge.auvin@ipsen.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.12.036

S. Auvin et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1586–1589
1587
O
1
2
1
2
O
O
1
2
R
R
O
R
R
O
H
H
i
ii
iii
R
R
H
O
H2N
O
N
O
N
N
H
O
N
Cbz
N
N
H
H2N
N
H
R
N
N
H
H
H
OMe
O
O
OMe
OMe
O
OR'
2
3
4
5a-d R' = Me
6a-d R' = H
7a R' = Ac
iv
v
H
H
N
H
N
H
N
N
O
R =
S
S
S
S
1a
1b
1c
1d
Scheme 1. Reagents and conditions: (i) CbzHN-C(R1R2)-CO2H, HBTU, DIEA, DMF, 20 °C, 15 h; (ii) H2 Pd/C 10%, CH2Cl2/EtOH, 20 °C, 2 h;
(iii) 1a-CO2H, 1b-CO2H, 1c-CO2H, or 1d-CO2H, HBTU, DIEA, DMF, 20 °C, 15 h; (iv) 2 N HCl, acetone/H2O, 20 °C, 15 h; (v) 6a, Ac2O, DMAP,
20 °C, 4 h, 60%.
1b-CO2H,7 and 1c-CO2H,8 or with the ether analog 1d-
CO2H9 afforded a series of dipeptides 5a–d with a differ-
ent pattern of substitution between the peptidic core and
the antioxidant. Deprotection of the ketal group was
performed under acidic conditions to give access to the
final hybrid compounds 6a–d. It is noteworthy that each
compound of the 6a–d series exists as an equilibrated
mixture of two diastereoisomers due to the reversible
closing and opening of the hemiketal group in solution.4
One of these compounds, 6a, was acetylated with Ac2O
to generate a prodrug, 7a.
A double L-leucine peptidic chain was selected as the
scaffold for the evaluation of the phenothiazine connec-
tion to the rest of the molecule (Table 1). As expected,
phenothiazines gave a series of very potent antioxidants
which were at least as good as BHT in the LPO test.
Whereas the shortest compound, 6a, built on the 2-carb-
oxyphenothiazine (1a) residue moderately inhibited the
isolated human calpain (IC50 = 85.5 nM), it proved to
be slightly more active than Z-LL-H in the cellular cal-
pain assay (IC50 = 1.8 lM). To see whether a comple-
mentary prodrug strategy would be useful, compound
7a was prepared. However, 7a did not show any
improvement compared to the parent drug 6a
(IC50 = 1.7 lM). Consequently, our efforts were mainly
devoted to the optimization of the N-terminal part. Fur-
ther increases of the alkyl chain (1b and c) in compounds
6b and c were not favorable as shown by the drop in
activity (6c, IC50 = 10 lM) in the cellular calpain test.
The biological activities of compounds 6a–d and 7a were
evaluated in a human isolated calpain 1 enzyme assay,10
with Suc-Leu-Tyr-AMC as substrate, and the intracellu-
lar calpain inhibition was performed in glioma C6 cells
with the same substrate, in the presence of maitotoxin
as a calpain activation inducer.11 The antioxidant poten-
cy of the final molecules was assessed by their ability to
inhibit Fe2+ induced lipid peroxidation (LPO) in rat
brain microsomes.12
Conversely, the analogous compound 6d-08, bearing an
oxygen atom, outperformed not only Z-LL-H in the
intracellular assay (IC50 = 0.6 lM) but also the other
molecules of the series in the LPO test (IC50 = 333 nM).
This beneficial double effect, consecutive to the exchange
of residue 1c by d in the molecule, appointed 6d-08 as
the new lead compound.
Z-LL-H13 and BHT14 (2,6-di-tert-butyl-4-methylphenol)
were chosen as reference compounds for the calpain and
LPO tests, respectively. The results are shown in
Table 1.
Table 1. Human calpain 1, lipid peroxidation, and intracellular calpain inhibitions for 6a–d and 7a
O
O
H
O
N
R
N
N
H
H
O
OR'
a
Human calpain 1 IC50 (nM)
a
LPO IC50 (nM)
a
Compound
R
R0
Intracellular calpain IC50 (lM)
Z-LL-H
BHT
6a
—
—
1a
1a
1b
1c
1d
—
—
H
4.9
—
na
3
—
3250
1095
2310
966
85.5
905
55.4
150
105
1.8
1.7
2.5
10
7a
6b
Ac
H
6c
6d-08
H
3296
333
H
0.6
na = not active.
a Values are means of three experiments.

1588
S. Auvin et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1586–1589
Table 2. Human calpain 1 and intracellular calpain inhibitions with molecular descriptors for 6d-01 to 6d-12
1
R
2
O
O
R
H
N
H
O
O
N
N
N
H
H
O
OH
S
R1
R2
Intracellular calpain IC50 (lM)
pRatiob
Rotors14
FISAc
a
Human calpain 1 IC50 (nM)
a
Compound
6d-01
6d-02
6d-03
6d-04
6d-05
6d-06
6d-07
6d-08
6d-09
6d-10
6d-11
6d-12
—
Me
—
—
Me
—
—
—
i-Pr
—
—
—
—
—
8
25
63
15
30
0.8
0.6
1
ꢁ0.9
0.22
0.66
2.05
2.4
11
11
11
12
13
12
12
13
12
12
13
13
178.6
151.82
125.38
138.13
134.54
123.83
147.67
133.82
121.91
137.04
186.72
167.53
Me
Et
138
88.6
150.9
75.9
553.2
105
n-Pr
i-Pr
1.88
1.54
2.24
1.95
1.65
1.21
1.28
—
i-Bu
16
0.6
4
t-Bu
Ph
354.1
222
5
HO-CH2
H3C-CH(OH)
48.7
38
3
2
a Values are means of three experiments.
b pRatio = ꢁlog(intracellular calpain/human calpain 1).
c FISA: polar surface area.15
To thoroughly study the SAR around 6d-08, an addi-
tional synthetic program was set up to replace the L-leu-
cine of the linker by diversely substituted a-aminoacids
(Table 2). The significance of the lateral chain was
apprehended by using glycine as a starting point in 6d-
01, this compound behaving as an excellent calpain
inhibitor (IC50 = 8 nM). Attempts to improve this
inhibition by a variation of the lateral chain in the linear
or branched alkyl series led only to a decrease in inhib-
itory potency. Moreover, neither the gem-dimethyl sub-
stitution in 6d-03 nor the D-isomer of valine in 6d-07
were valuable. A phenyl residue used as a flat surrogate
of the alkyl chain in compound 6d-10 (IC50 = 222 nM)
did not perform better than 6d-01. To glance over less
lipophilicity, two alcohol derivatives 6d-11 and 6d-12
were prepared and evaluated. Although they displayed
potent IC50s against calpain (ꢀ40 nM), they did not per-
form better than their alkyl counterparts. These results
illustrated that calpain was very sensitive to the pattern
of substitution and that glycine was the most suitable
amino acid at this region.
that Suc-Leu-Tyr-AMC was not a substrate for cathep-
sin B to ensure reliability of the results. Surprisingly,
compound 6d-01 which behaved as an excellent calpain
inhibitor displayed poor activity in the cellular assay
(IC50 = 63 lM). This could be attributed either to a lack
of permeation or to a degradative process occurring in-
side the cell. Among the methyl analogs, the L-alanine
derivative 6d-02 was able to recover some activity,
whereas the gem-dimethyl 6d-03 was not. As this gem
di-substitution was likely to prevent this compound
from proteolysis, the only remaining hypothesis is based
on the permeation properties of the molecules. Increas-
ing the size of the lateral chain from ethyl to isobutyl,
6d-04 to 6d-08, led to a series of powerful inhibitors
(IC50 = 0.6 lM for 6d-05 and 6d-08) of cellular calpain,
except for 6d-07. The L-isomer 6d-06 was sixteen times
more potent than the D-isomer 6d-07 in the cellular cal-
pain inhibition assay which underlines the usefulness of
the L-isomer for this chemical series. Even if these com-
pounds were, on average, 20 times less active than Z-LL-
H on the isolated enzyme, they displayed a consistent
four to five times better performance in the cellular as-
say. In the same way, the most bulky and lipophilic
compounds of this series, 6d-09 and 6d-10, although
considerably less active than Z-LL-H on calpain, exhib-
ited similar performance against cellular calpain. Inter-
estingly, the two alcohol derivatives 6d-11 and 6d-12
also performed as well as Z-LL-H.
Conversely, the peptidic backbone modifications of the
6d series had a limited impact on free radical scavenging.
All the compounds had IC50s in the same range of activ-
ity (ꢀ300–400 nM) in the LPO assay. The antioxidant
potency of the molecules was more influenced by the
phenothiazine substitution (Table 1) than by the rest
of the structure.
The IC50s of the intracellular calpain inhibition (Table
2) correspond to a 4 h incubation time of the molecules
including 3 h in the presence of maitotoxin. It is note-
worthy that this inhibition remained constant between
90 min and 3 h showing that our molecules gave a pro-
longed inhibition.
To complete the inhibitory activity profile of com-
pounds 6d-01 to 6d-12, they were tested against cathep-
sin B, caspase 3, and a-chymotrypsin. Indeed, these
molecules were potent inhibitors of cathepsin B, IC50s
were equivalent to those calculated for the calpain inhi-
bition. In contrast, caspase 3 and a-chymotrypsin were
not inhibited by the compounds.
In the light of these results, the cellular activity could
not be related to either stability or lipophilicity
(ClogP).3d A theoretical calculation approach was
therefore set up to get an insight into the parameters
On the other hand, this series of compounds was evalu-
ated in a cellular calpain assay. Previously, we checked

S. Auvin et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1586–1589
1589
3
2,5
2
1098; (c) Zatz, M.; Starling, A. N. Eng. J. Med. 2005, 352,
2413.
Predicted pRatio
2. Chabrier, P.-E.; Pignol, B.; Auvin, S. PCT Int. Appl. WO
0240016, 2002; . Chem. Abstr. 2002, 136, 380140.
3. (Nonexhaustive list): (a) Li, Z.; Patil, G. S.; Golubski, Z.
E.; Hori, H.; Tehrani, K.; Foreman, J. E.; Eveleth, D. D.;
Bartus, R. T.; Powers, J. C. J. Med. Chem. 1993, 36, 3472;
(b) Li, Z.; Ortega-Vilain, A.-C.; Patil, G. S.; Chu, D.-L.;
Foreman, J. E.; Eveleth, D. D.; Powers, J. C. J. Med.
Chem. 1996, 39, 4089; (c) Lubisch, W.; Hofmann, H. P.;
Treiber, H. J.; Moller, A. Bioorg. Med. Chem. Lett. 2000,
10, 2187; (d) Josef, K. A.; Kauer, F. W.; Bihovsky, R.
Bioorg. Med. Chem. Lett. 2001, 11, 2615; (e) Lubisch, W.;
Moller, A. Bioorg. Med. Chem. Lett. 2002, 12, 1335; (f)
Lubisch, W.; Beckenbach, E.; Bopp, S.; Hofmann, H.-P.;
Kartal, A.; Kaestel, C.; Lindner, T.; Metz-Garrecht, M.;
Reeb, J.; Regner, F.; Vierling, M.; Moeller, A. J. Med.
Chem. 2003, 46, 2404; (g) Gil-Parrado, S.; Assfalg-Mach-
leidt, I.; Fiorino, F.; Deluca, D.; Pfeiler, D.; Schaschke,
N.; Moroder, L.; Machleidt, W. Biol. Chem. 2003, 384,
395; (h) Bihovsky, R.; Tao, M.; Mallamo, J. P.; Wells, G.
J. Bioorg. Med. Chem. Lett. 2004, 14, 1035.
4. Auvin, S.; Pignol, B.; Navet, E.; Pons, D.; Marin, J.-G.;
Bigg, D.; Chabrier, P.-E. Bioorg. Med. Chem. Lett. 2004,
14, 3825.
5. Auvin, S.; Chabrier de Lassauniere P.-E. PCT Int. Appl.
WO 2005/056551 A2, 2005; . Chem. Abstr. 2005, 143,
44075.
6. Chan, C.; Yin, H.; Garforth, J.; McKie, J. H.; Jaouhari,
R.; Speers, P.; Douglas, K. T.; Rock, P. J.; Yardley, V.;
Croft, S. L.; Fairlamb, A. H. J. Med. Chem. 1998, 41, 148.
7. Zupancic, B. Ger. Offen. 2702714, 1978; . Chem. Abstr.
1978, 89, 180029.
1,5
1
0,5
0
-1,50
-1,00
-0,50
0,00
0,50
1,00
1,50
2,00
2,50
3,00
pRatio
-0,5
-1
-1,5
Figure 1. Predicted pRatio versus experimental pRatio of compounds
6d-01 to 6d-12.
which describe the behavior of these molecules in the
cellular assay. For this purpose 3D coordinates of
molecular structures were generated by Corina 3.2
(Molecular Networks GmbH) using the ‘write hydrogen
atoms’ and the ‘neutralize formal charges’ options. The
resulting 3D structures were then used to generate
molecular descriptors such as rotors15 and FISA16 using
QikProp 2.2 (Schrodinger) (Table 2). Multiple linear
regression analysis highlighted the significance of the
number of rotors and the polar surface area (FISA) of
the molecules. The predicted pRatio could thus be de-
scribed as a combination of these two descriptors:
8. Messer, M.; Farge, D.; Guyonnet, J. C.; Jeanmart, C.;
Julou, L. Arzneim. Forsch. 1969, 19, 1193.
predicted pRatio ¼ ꢁ6:479 þ ð0:987Þ ꢂ rotor ꢁ ð0:028Þ
ꢂ FISA.
9. The carboxylic acid was prepared by treatment with LiOH
of the corresponding ethyl ester described in: Yu, M. J.;
McCowan, J. R.; Thrasher, K. J.; Keith, P. T.; Luttman,
C. A.; Ho, P. P. K.; Towner, R. D.; Bertsch, B.; Horng, J.
S.; Um, L. S.; Phebus, L. A.; Saunders, R. D. J. Med.
Chem. 1992, 35, 716.
10. Meyer, S. L.; Bozyczko-Coyne, D.; Mallya, S. K.; Spais,
C. M.; Bihovsky, R.; Kawooya, J. K.; Lang, D. M.; Scott,
R. W.; Siman, R. Biochem. J. 1996, 314, 511.
11. Wang, K. K. W.; Nath, R.; Raser, K. J.; Hajimohamma-
dreza, I. Arch. Biochem. Biophys. 1996, 331, 208.
12. Esterbauer, H.; Schaur, R. J.; Zollner, H. Free Radical
Biol. Med. 1991, 11, 81.
13. Iqbal, M.; Messina, P. A.; Freed, B.; Das, M.; Chatterjee,
S.; Tripathy, R.; Tao, M.; Josef, K. A.; Dembofsky, B.;
Dunn, D.; Griffith, E.; Siman, R.; Sedanhi, S. E.; Biazzo,
W.; Bozyczko-Coyne, D.; Meyer, S. L.; Ator, M. A.;
Bihovsky, R. Bioorg. Med. Chem. Lett. 1997, 7, 539.
14. Sgaragli, G. P.; Valoti, M.; Gorelli, B.; Fusi, F.; Palmi,
M.; Mantovani, P. Br. J. Pharmacol. 1993, 110, 369.
15. Number of nontrivial (not CX3), nonhindered (not
alkene, amides, and small ring) rotatable bonds in the
molecule.
The excellent correlation and predictability of this mod-
el, q2 = 0.846 (cross-validated r2), between the predicted
pRatio and the experimental pRatio (Table 2) are illus-
trated in Figure 1.
Our study has led to the design of a series of dipeptides
which are potent calpain and LPO inhibitors. Some of
them, 6d-05 and 6d-08, show strong inhibitory potencies
in cellular assays with IC50s below 1lM and with a lim-
ited (4- to 6-fold) decrease in cellular activity compared
to enzymatic inhibition. With the help of a calculation
approach, we were able to demonstrate that, in this ser-
ies of compounds, the polar surface area and the num-
ber of rotors of a molecule adequately described their
performance in this cellular calpain inhibition assay.
References and notes
16. Hydrophilic component of the solvent accessible surface
area.
1. (a) Donkor, I. O. Curr. Med. Chem. 2000, 7, 1171; (b)
Branca, D. Biochem. Biophys. Res. Commun. 2004, 322,