Design and identification of selective HER-2 sheddase inhibitors via P1′ manipulation and unconventional P2′ perturbations to induce a molecular metamorphosis

Article abstract of DOI:10.1016/j.bmcl.2007.10.108

In an effort to obtain a MMP selective and potent inhibitor of HER-2 sheddase (ADAM-10), the P1′ group of a novel class of (6S,7S)-7-[(hydroxyamino)carbonyl]-6-carboxamide-5-azaspiro[2.5]octane-5-carboxylates was attenuated and the structure-activity rela

Full text of DOI:10.1016/j.bmcl.2007.10.108

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 159–163
Design and identification of selective HER-2 sheddase inhibitors
via P10 manipulation and unconventional P20 perturbations
to induce a molecular metamorphosis
Wenqing Yao,a,* Jincong Zhuo,a David M. Burns,a Yun-Long Li,a Ding-Quan Qian,a
Colin Zhang,a Chunhong He,a Meizhong Xu,a Eric Shi,a Yanlong Li,b Cindy A. Marando,b
Maryanne B. Covington,b Gengjie Yang,b Xiangdong Liu,b Max Pan,b
Jordan S. Fridman,b Peggy Scherle,b Zelda R. Wasserman,a Gregory Hollis,b Kris Vaddi,b
Swamy Yeleswaram,b Robert Newton,b Steve Friedmanb and Brian Metcalfa,b
aIncyte Corporation, Department of Medicinal Chemistry, Wilmington, DE 19880, USA
bIncyte Corporation, Department of Discovery Biology, Wilmington, DE 19880, USA
Received 28 September 2007; revised 29 October 2007; accepted 30 October 2007
Available online 4 November 2007
Abstract—In an effort to obtain a MMP selective and potent inhibitor of HER-2 sheddase (ADAM-10), the P10 group of a novel
class of (6S,7S)-7-[(hydroxyamino)carbonyl]-6-carboxamide-5-azaspiro[2.5]octane-5-carboxylates was attenuated and the structure–
activity relationships (SAR) will be discussed. In addition, it was discovered that unconventional perturbation of the P20 moiety
could confer MMP selectivity, which was hypothesized to be a manifestation of the P20 group effecting global conformational
changes.
Ó 2007 Elsevier Ltd. All rights reserved.
The human epidermal growth factor receptor-2 (HER-2
or ErbB-2) is a tyrosine kinase receptor that is activated
by ligand induced homo- and heterodimerization of the
extracellular domain (ECD) between two members of
the HER family or by proteolytic cleavage (shedding) of
the ECD.1 Once activated, intracellular signal transduc-
tion pathways are initiated that mediate a diverse range
of essential cellular activities such as cell proliferation, dif-
ferentiation, motility, adhesion, and survival.2 Overex-
pression of the oncogene HER-2/neu has been
associated with aggressive pathogenesis, poor prognosis,
and decreased responsiveness toconventional chemother-
apeutic and hormonal treatment regimes in non-small cell
lung cancer (NSCLC), ovarian cancer, and breast cancer
patients.1a In addition, elevated plasma levels of HER-
2ECD have been associated with metastasis and a de-
crease in disease-free and overall survival in patients with
breast cancer.1b,3 Therefore, inhibition of the protease
responsible for HER-2 shedding is therapeutically desir-
able for treating cancer patients that overexpress HER-2.
We recently identified ADAM-10 as a major source of
HER-2 ectodomain sheddase activity in HER-2 over-
expressing breast cancer cells4a and reported the design,
synthesis, evaluation, and identification of a novel class
of (6S,7S)-7-[(hydroxyamino)carbonyl]-6-carboxamide-
5-azaspiro[2.5]octane-5-carboxylates as the first potent
and selective inhibitors of HER-2 sheddase.4b Now, we
wish to report the results of a more comprehensive
investigation of the structure–activity relationships with
regard to the P10 and P20 substituents (see Scheme 1).
A series of hydroxamic acids of general formula 3 were
prepared using sequential BOP reagent mediated amide
coupling reactions starting from the designed polyfunc-
tionalized chiral building block 1, which was prepared
by the methods that were previously disclosed starting
from L-aspartic acid b-tert-butyl ester and 3-chloro-2-
(chloromethyl)-1-propene.4b,5
Keywords: HER-2 sheddase; ADAM-10; Metalloprotease; MMP; P10
group; P20 group; Global conformation; SAR; Piperidine; Hydroxamic
acid.
Previously we reported that 4-phenyl–piperazine 4, tet-
rahydropyridine 5, and piperidine 6 hydroxamic acids
*
Corresponding author. Tel./fax: +1 302 498 6707; e-mail:
wyao@incyte.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.10.108

160
W. Yao et al. / Bioorg. Med. Chem. Lett. 18 (2008) 159–163
Zn+2
NH
HO
O
P1
O
O
b
a
O
N
O
N
O
n
X
n
Ar
X Ar
OH
N
R1
N
R1
N
H
O
O
P2'
P1'
n: 0, 1
X: C, CH, N
R1: H, C(O)OR, S(O)2R, C(O)R
1
2
3
Scheme 1. Reagents: (a) i—BOP, amine, DIEA, DMF; ii—i-Pr2EtN, R1-Cl; (b) i—TFA, DCM; ii—NH2OH, BOP, DIEA, DMF.
exhibited excellent cellular inhibition of HER-2 shed-
ding (<140 nM) with excellent selectivity against
MMP-1 and -3, modest to excellent selectivity against
MMP-9, and negligible selectivity against MMP-2 (Ta-
ble 1).4b Achieving selectivity toward MMP-2 and -9
proved to be an arduous task, particularly toward
MMP-2, due to the homology of these metalloproteases
within the active site. We found that selectivity against
MMP-2 and -9 could be accomplished by slightly per-
turbing the torsion angle between the P10 aryl ring and
the heterocyclic ring to which it is attached.4b We spec-
ulated that the observed selectivity was due to a subtle
difference at position 223 within the narrow and deep
loop 3 region of the S10 pocket, which for MMP-2, -3,
and -9 is a tyrosine residue and for ADAM-10 is an ala-
nine. For the current study, we hypothesized that the
additional space that is created due to the smaller Ala
residue at 223 of ADAM-10 might be exploited by
substituting the P10 phenyl ring at the 3-position to
obtain MMP selectivity. Thus, a series of 3-substi-
tuted-aryl hydroxamic acids were prepared and the re-
sults are outlined in Table 1. Selectivity was assessed
by comparing the enzymatic binding data of ADAM-
10 to MMP-1, -2, -3, and -9.
The 3-methyl-, 3-ethyl-, and 3-methoxy-phenyl-tetrahy-
dropyridine compounds 7, 8, and 9 did not show any
indication of enhanced selectivity and were found to
be 2- to 13-fold less potent against HER-2 sheddase in
comparison to the unsubstituted phenyl-tetrahydropyri-
dine 5 (Table 1). The 3-isopropyl-phenyl-tetrahydro-
pyridine analog 10 displayed a similar loss in potency
that was compensated for by a profound improvement
in the selectivity against MMP-2 (>15-fold), MMP-9
(10-fold), and MMP-3 (2-fold). Exchanging the 3-iso-
propyl group of 10 for a cyclopropyl group completely
Table 1. Investigation of various substituents in the 3-position of the P10 aryl ring
HO
NH
R3
R5
O
R4
N
N
W
H
O
Compound
Bond
W
R3
R4
R5
HER-2a (nM)
Enzyme Binding (nM)
ADAM-10
MMP-1
MMP-2
MMP-3
MMP-9
4
5
Single
Double
Single
N
C
H
H
H
H
H
H
H
H
H
H
Me
H
Me
F
52
18
118
33
>5000
3878
>5000
1892
>5000
>5000
>5000
—
195
5
>5000
1237
>5000
359
650
113
>5000
111
190
110
2500
—
H
H
6
CH
C
H
H
137
42
324
25
446
13
7
Double
Double
Double
Double
Double
Double
Double
Double
Double
Double
Single
Me
Et
H
8
C
H
240
194
229
162
86
33
40
21
50
—
—
9
C
OMe
i-Pr
Cyclopropyl
Me
i-Pr
Me
F
H
10
11
12
13
14
15
16
17
18
19
20
21
22
C
H
73
48
1110
24
4999
—
C
H
C
H
85
>5000
—
433
2500
159
14
>5000
—
>5000
—
C
CN
CN
H
902
129
24
NAb
184
594
42
154
28
C
>5000
>5000
—
3500
>5000
—
>5000
916
—
C
15
99
C
Cl
H
Cl
H
H
H
H
H
H
4999
63
N
N
N
N
CH
CH
Me
Me
Cl
H
187
38
>5000
—
—
>1000
—
—
1268
—
Single
Single
CN
H
<5
32
14
—
—
Single
Single
Single
i-Pr
Me
i-Pr
H
NA
689
139
1102
194
16
—
>5000
—
>5000
391
1387
—
>5000
>5000
H
604
—
H
a Values are obtained from a BT-474 cellular HER-2 ECD shedding assay.
b (NA, not active).

W. Yao et al. / Bioorg. Med. Chem. Lett. 18 (2008) 159–163
161
abated the MMP-2 selectivity. The 3,5-dimethyl deriva-
tive 12 exhibited a substantial gain in selectivity toward
MMP-2 (>5-fold) and MMP-9 (17-fold) in comparison
to the unsubstituted parent phenyl compound 5; albeit
with a concomitant loss in HER-2 sheddase potency.
responding piperazine analog 20 was prepared and
was found to be inactive against both HER-2 shed-
dase and MMP-2, which may imply that the tetrahy-
dropyridine and piperazine ring systems have subtle
conformational differences that must orient the at-
tached phenyl ring and thus the 3-isopropyl group
slightly differently within the narrow S10 pocket.
Our previously disclosed exploratory studies indicated
that placement of an electron withdrawing group
(EWG) in the 4-position of the P10 aromatic ring may
lead to a cogent increase in potency.4b Therefore, a cya-
no group was placed in the 4-position of the MMP-
selective 3-isopropyl analog 10 and the 3,5-dimethyl-
phenyl tetrahydropyridine analog 12 in an effort to
boost the potency while maintaining the selectivity.
The 4-cyano-3-isopropyl-phenyl analog 13 was found
to be less potent than 10 in both the enzymatic
ADAM-10 and cellular HER-2 assays. In contrast, the
4-cyano-3,5-dimethyl-phenyl analog 13 was 3-fold more
potent than 12 in the enzymatic ADAM-10 binding as-
say and retained the excellent selectivity against other
MMPs; despite a slight loss in cellular HER-2 potency.
To further explore the role of the P10 heterocycle in ori-
enting the aryl ring within the S10 pocket, the 3-methyl
and 3-isopropyl-phenyl piperidine compounds 21 and
22 were prepared. The 3-methyl analog 21 was found
to be only slightly more selective than the unsubstituted
parent phenyl piperidine compound 6 against MMP-2
and was 5-fold less selective against MMP-9. In con-
trast, the 3-isopropyl analog 22 displayed an 80-fold
improvement in selectivity toward MMP-2 in compari-
son to 6 and a 19-fold accession in ADAM-10 potency.
Comparison of the piperidine compound 22 to the tetra-
hydropyridine analog 10 indicates a 4-fold increase in
ADAM-10 potency and a 6-fold increase in selectivity
toward MMP-2. The improved binding profile of 22
may be a manifestation of a distortion in the dihedral
angle between the P10 heterocyclic ring and the aromatic
ring, as previously mentioned.
Alternatively, the aryl ring of the 3,5-dimethyl-phenyl
tetrahydropyridine analog 12 was made electron defi-
cient by directly substituting the methyl groups for
EWGs. Thus, the 3,5-difluoro-derivative 15 was pre-
pared and found to have the desired increase in potency
in comparison to 12; however, the selectivity against
MMP-2 was abrogated. The slightly larger 3,5-di-
chloro-phenyl derivative 16 showed diminutive MMP-
2 binding and HER-2 activity, possibly indicating that
the outer limits of the spatial diameter within the S10
pocket were reached. Overall, these results imply that
the electronic disposition of the aromatic ring does not
play as crucial a role in determining the binding proper-
ties for the tetrahydropyridine series as for the previ-
ously published piperazine series.4b This is in
agreement with the initial proposed hypothesis that the
role of the EWG in the piperazine series is to promote
coplanarity between the piperazine and phenyl rings
by facilitating electron donation from the piperazine
nitrogen lone pair to the electron deficient aromatic p-
system.4b For the tetrahydropyridine series, conjugative
factors predispose the P10 heterocycle and aryl ring to be
coplanar; therefore, steric interactions appear to be the
main determinant for the observed metalloprotease
binding affinities for this series.
In summary, excellent selectivity toward MMP-2 and -9
can be achieved by the introduction of a small substitu-
ent at the 3-postion of the P10 phenyl ring. The effect
that the 3-aryl substituent has on the metalloprotease
binding profile is contingent on the P10 heterocycle,
since this predetermines the spatial orientation of the
P10 aryl ring and the appended 3-substituent. Based on
the data, it is plausible that the P10 3-aryl substituent
is occupying the space created by the Ala 223 of
ADAM-10 and is interacting unfavorably with Tyr
223 of MMP-2 and -9, as originally proposed. The 3-iso-
propyl moiety appears to be the optimal group for this
small lipophilic pocket rendering analogs 10 and 22 with
excellent MMP-2 and -9 selectivity. The piperidine ring
emerges as the optimal P10 heterocycle for orienting the
3-isopropyl phenyl moiety within the fastidious S10
pocket. Furthermore, the data suggests that the diame-
ter of the S10 pocket constricts with the following hierar-
chy: HER-2 > MMP-2 > MMP-9.
Previously, it was found that the metalloprotease bind-
ing properties could be attenuated by substitution of
the core scaffold piperidine N–H.4b It was determined
that simple N-methylation of the core nitrogen of com-
pounds 4 and 5 resulted in a 2-fold loss in potency in the
enzymatic ADAM-10 assay and a 4- and 9-fold loss in
HER-2 cellular potency, respectively. Installation of a
methyl carbamate to the core nitrogen of 5 had a similar
effect on the binding profile; however, the pharmacoki-
netic properties of the carbamate were superior to the
parent N–H compound 5. In an effort to further explore
the SAR of the P20 group, a series of scaffold N-substi-
tuted analogs were prepared and the results are depicted
in Table 2.
This hypothesis was further explored by preparing a
series of 3-aryl substituted piperazine compounds (Ta-
ble 1, compounds 17–20). The 3-methyl aryl pipera-
zine analog 17 had a higher binding affinity toward
MMP-2 than HER-2 sheddase. Efforts to modify the
torsion angle between the phenyl and piperazine rings
in a favorable manner, as previously discussed, by the
addition of a cyano group in the 4-position to afford
analog 18 afforded another potent MMP-2 inhibitor.
Replacement of the methyl group of 17 with an
EWG to afford the 3-Cl analog 19 resulted in a nota-
ble gain in HER-2 potency, which was inconsequential
due to a lack of MMP-2 selectivity. Based on the
encouraging MMP-2 selectivity of the 4-(3-iso-pro-
pyl-phenyl)tetrahydro-pyridine compound 10, the cor-
Comparison of the phenyl-tetrahydropyridine N-carba-
mates 23, 24, and 25 to their parent N–H compound 5 re-

162
W. Yao et al. / Bioorg. Med. Chem. Lett. 18 (2008) 159–163
Table 2. Investigation of the P20 and P10 substituents
HO
NH
O
O
N
R1
R2
Compound
R1
R2
HER-2a (nM)
Enzyme bindinga (nM)
ADAM-10
MMP-1
MMP-2
MMP-3
MMP-9
23
24
25
26
27
28
29
30
31
32
33
34
CO2Me
CO2Bu-i
Cbz
4-Ph-Tetrahydropyridine
4-Ph-Tetrahydropyridine
4-Ph-Tetrahydropyridine
4-Ph-Tetrahydropyridine
4-Ph-Tetrahydropyridine
4-Ph-Piperidine
68
150
431
72
59
19
34
40
51
19
37
27
97
81
23
26
>5000
>5000
>5000
>5000
>5000
—
39
35
4000
2601
1669
>5000
>5000
—
428
451
309
160
677
—
37
50
C(O)Me
SO2Me
C(O)OMe
CO2THPb
SO2Me
H
93
53
169
26
472
106
86
4-Ph-Piperazine
4-Ph-Piperazine
>5000
>5000
>5000
—
4000
>5000
>3000
—
2623
1499
>3000
—
45
92
3-Ph-Pyrrolidine
3-Ph-Pyrrolidine
960
>5000
13
C(O)OMe
H
C(O)OMe
289
13
3-Ph-Pyrrolid-3-ene
3-Ph-Pyrrolid-3-ene
2500
>5000
450
>5000
77
3048
34
488
a Values are obtained from a BT-474 cellular HER-2 ECD shedding assay.
b 4-Tetrahydropyran (THP).
veals that the selectivity toward MMP-3 and -9 is mark-
edly enhanced, while the ADAM-10 and MMP-2 binding
affinities remain impervious. The observed loss in cellular
HER-2 potency as the alkoxy group distended may be
attributed to a change in physiochemical properties. The
phenyl-tetrahydropyridine N-acetyl and methylsulfona-
mide compounds 26 and 27 possessed a binding profile
that closely resembled that of the corresponding methyl
carbamate 23. The phenyl–piperidine N-methyl-carba-
mate 28 had a 19-fold increase in MMP-2 selectivity and
a 17-fold boost in ADAM-10 binding in comparison to
the parent N–H compound 6. The 4-phenylpiperazine
carbamate 29 exhibited a considerable improvement in
potency and an ꢀ2-fold gain in selectivity against
MMP-2 and a 13-fold improvement in selectivity toward
MMP-9 in comparison to the parent N–H compound 4.
The methylsulfonamide analog 30 had a metalloprotease
binding profile similar to the carbamate 29. The 3-phenyl-
pyrolidine methyl carbamate analog 32 had a 5-fold in-
crease in MMP-2 selectivity in comparison to the parent
N–H compound 31. The 3-phenylpyrolid-3-ene methyl
carbamate derivative 34 had an even more pronounced
amelioration in selectivity toward MMP-2 (37-fold),
MMP-9 (40-fold), MMP-3 (11-fold), and MMP-1 (2-fold)
in comparison to the parent N–H compound 33. For both
of the methyl carbamate analogs 32 and 34, the ADAM-
10 affinity was similar to their corresponding parent N–H
compounds 31 and 33; however, both analogs suffered a
ꢀ3-fold loss in cellular HER-2 sheddase potency. This
discrepancy can be explained by the alteration in physio-
chemical properties.
Overall, it appears that substitution of the scaffold piper-
idine N–H with an EWG can result in increased selectiv-
ity, especially with regard to MMP-3 and -9, often with
either negligible or positive repercussions to the desired
HER-2 potency. It is speculated that the observed influ-
ence that the P20 group has on the metalloprotease bind-
ing affinity is not due to direct interactions between the P20
substituent and the S20 subsite, since the S20 subsite is be-
lieved to be primarily solvent exposed. Instead, the varia-
tion in binding affinitymay be an indirect manifestation of
the P20 group inducing global conformational changes
due to A(1,3) strain, thus shifting the orientation of the
P10 group within the highly sensitive S10 pocket. The data
presented in Table 2 support this hypothesis indicating
that the influence conferred by the P20 substituent on
the enzyme binding profile is interconnected to the P10
group and that the nature of the P20 group, so long as it
is an EWG, is insignificant. It is also reasonable that the
P20 substituent may have an allosteric affect on the
HER-2 sheddase active site, thus altering the disposition
of the S10 pocket and the concomitant interactions with
the P10 group.
In conclusion, the metalloprotease binding profile can be
attenuated to selectively inhibit HER-2 sheddase by
installing a substituent at the 3-position of the P10 aro-
Table 3. PK properties of 23 and 34
Compound
h-PB (% free)
Mouse PKa (po)
AUC(0ꢁ24 h) (nM h)
T1/2 (h)
Cmax (nM)
23
34
30
34
3900
2800
7700
3000
1.9
1.8
a Average of three animals administered at 10 mg/kg as a suspension in 10% DMAC in 0.5% methylcellulose.

W. Yao et al. / Bioorg. Med. Chem. Lett. 18 (2008) 159–163
163
matic ring or an EWG, such as a carbamate, on the scaf-
References and notes
fold piperidine nitrogen. In both cases the heterocycle to
which the P10 aromatic ring is attached plays a critical role
in determining the effect that the substituent will have
on the metalloprotease binding profile. Future SAR
studies will examine the corollaries between modifying
both of these variables simultaneously on the same
compound.
1. (a) Mass, R. D. Int. J. Radiat. Oncol. Biol. Phys. 2004, 58,
932; (b) Carney, W. P.; Neumann, R.; Lipton, A.; Leitzel,
K.; Ali, S.; Price, C. P. Clin. Chem. 2003, 49, 1579; (c)
Arribas, J.; Borroto, A. Chem. Rev. 2002, 102, 4627.
2. Marmor, M. D.; Skaria, K. B.; Yarden, Y. Int. J. Radiat.
Oncol. Biol. Phys. 2004, 58, 903.
3. Molina, M. A.; Saez, R.; Ramsey, E. E.; Garcia-Barchino,
M.-J.; Rojo, F.; Evans, A. J.; Albanell, J.; Keenan, E. J.;
Lluch, A.; Garcia-Conde, J.; Baselga, J.; Clinton, G. M.
Clin. Cancer Res. 2002, 8, 347.
4. (a) Liu, P. C. C.; Liu, X.; Li, Y.; Covington, M.; Wynn, R.;
Huber, R.; Hillman, M.; Yang, G.; Ellis, D.; Marando, C.;
Katiyar, K.; Bradley, J.; Abremski, K.; Stow, M.; Rupar,
M.; Zhuo, J.; Li, Y.-L.; Lin, Q.; Burns, D.; Xu, M.; Zhang,
C.; Qian, D.-Q.; He, C.; Sharief, V.; Weng, L.; Agrios, C.;
Shi, E.; Metcalf, B.; Newton, R.; Friedman, S.; Yao, W.;
Scherle, P.; Hollis, G.; Burn, T. C. Cancer Biol. Ther. 2006,
5, 657–664; (b) Yao, W.; Zhou, J.; Burns, D. M.; Xu, M.;
Zhang, C.; Li, Y.-L.; Qian, D.-Q.; He, C.; Weng, L.; Shi,
E.; Lin, Q.; Agrios, C.; Burn, T. C.; Caulder, E.; Covington,
M. B.; Fridman, J. S.; Friedman, S.; Katiyar, K.; Hollis, G.;
Li, Y.; Liu, C.; Liu, X.; Marando, C. A.; Newton, R.; Pan,
M.; Scherle, P.; Taylor, N.; Vaddi, K.; Wasserman, Z. R.;
Wynn, R.; Yeleswaram, S.; Jalluri, R.; Bower, M.; Zhou,
B.-B.; Metcalf, B. J. Med. Chem. 2007, 50, 603.
Several lead compounds were identified that possessed
superior selectivity against MMP-2 and -9 in compari-
son to the formerly reported lead compound 23
(INCB3619).4b Most notably compound 34 exhibited a
2-fold improvement in potency accompanied by a 19-
fold and a 16-fold enhancement in selectivity toward
MMP-2 and -9, respectively. Discrete oral pharmacoki-
netic murine studies indicated that 23 had a slightly
higher plasma exposure than 34, but based on the excep-
tional potency and selectivity exhibited by 34 further
pharmacokinetic and pharmacodynamic studies are
warranted and impending (Table 3).
Acknowledgments
The authors would like to thank Mei Li, Laurine Galya,
Karl Blom and Jin Liu of the analytical group and Kamna
Katiyar from the applied technology group.
5. Zhou, J.; Burns, D. M.; Zhang, C.; Xu, M.; Weng, L.; Qian,
D.-Q.; He, C.; Lin, Q.; Li, Y.-L.; Shi, E.; Agrios, C.;
Metcalf, B.; Yao, W. Synlett 2007, 460.