Discovery of a potent, selective and orally bioavailable 3,9-diazaspiro[5.5]undeca-2-one CCR5 antagonist

Article abstract of DOI:10.1016/j.bmcl.2008.10.115

Replacement of the cyclic carbamate in our previously disclosed 1-oxa-3,9-diazaspiro[5.5]undecan-2-one template led to the discovery of two novel series of 3,9-diazaspiro[5.5]undecane and undeca-2-one CCR5 antagonists. The synthesis, SAR, and antiviral ac

Full text of DOI:10.1016/j.bmcl.2008.10.115

Bioorganic & Medicinal Chemistry Letters 19 (2009) 209–213
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of a potent, selective and orally bioavailable
3,9-diazaspiro[5.5]undeca-2-one CCR5 antagonist
*
Hanbiao Yang , Xiao-Fa Lin, Fernando Padilla, Stephen D. Gabriel, Gabrielle Heilek, Changhua Ji,
*
Surya Sankuratri, André deRosier, Pamela Berry, David M. Rotstein
Roche Palo Alto LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Replacement of the cyclic carbamate in our previously disclosed 1-oxa-3,9-diazaspiro[5.5]undecan-2-one
template led to the discovery of two novel series of 3,9-diazaspiro[5.5]undecane and undeca-2-one CCR5
antagonists. The synthesis, SAR, and antiviral activities of these two series are described. One compound
(32) was found to have attractive combination of antiviral potency, selectivity, and pharmacokinetic pro-
file. The asymmetric synthesis of 32 was also accomplished and both enantiomers were equally potent.
Ó 2008 Elsevier Ltd. All rights reserved.
Received 20 September 2008
Revised 21 October 2008
Accepted 27 October 2008
Available online 31 October 2008
Keywords:
CCR5 antagonist
3,9-Diazaspiro[5.5]undecane
3,9-Diazaspiro[5.5]undecan-2-one
Anti-HIV-1
The development of new agents for the treatment of HIV-1/AIDS
remains a necessity, particularly due to the ongoing problem of
resistance development to current therapies.1 Viral entry inhibitors
are a validated target for controlling HIV-1 infection, particularly
blockade of the CCR5 receptor, which serves as one of the major
viral binding sites.2,3 Efforts at Roche have led to the discovery of
a series of potent CCR5 antagonists represented by structure 1, fea-
intermediate 9, which was converted to analogs 10–23 (Tables 1
and 2) via Boc deprotection and coupling. Alternatively, starting
from chiral amine 24, Boc protection, debenzylation and Strecker
reaction of 25 with N-Cbz-4-piperidone gave versatile intermedi-
ate 26, which was further converted to analogs 27 and 28 (Tables
1 and 2).
Synthesis of 5-butyl-3,9-diazaspiro[5.5]undeca-2-one 5-based
analogs is outlined in Scheme 3. Alkylation of 5 with alkyl bromide
or tosylate was achieved with NaH and DMF at elevated tempera-
ture (100 °C). However, large excesses of bromide and tosylate
(>5 equiv) were needed to achieve reasonable conversion. To our
turing
a
1-oxa-3,9-diazaspiro[5.5]undecan-2-one
template
(Fig. 1).4 In a continuation of our work, two new templates were
proposed. Inspired by this proposal, we developed a convenient
synthesis of 9-benzyl-1-butyl-3,9-diazaspiro[5.5]undecane 4 and
undecan-2-one 5 via Michael addition of the lithium enolate 2 to
the tetrasubstituted olefin acceptor 3 (Scheme 1).5 Herein, we re-
port novel and potent CCR5 antagonists featuring these new
templates.
O
R2
O
O
Two routes were developed for the synthesis of 3,9-diazaspi-
ro[5.5]undecane 4-based analogs (Scheme 2). Coupling of 4 with
tetrahydropyran-4-carboxylic acid gave amide 6, which was then
debenzylated and subjected to a Strecker reaction promoted by
Ti(O-iPr)4 to give nitrile 7. Treatment of 7 with excess methyl mag-
nesium bromide (5–8 equiv) at room temperature led to the ex-
pected displacement of the cyanide by methyl, as well as an
unexpected cleavage of the amide bond to yield amine 8. Reaction
of 8 with a variety of sulfonyl, sulfamyl, and acid chlorides afforded
R3
N
N
N
R1
n-Bu
1
US 2005176703 A1
R2
N
O
N
N
R3
R1
X
n-Bu
* Corresponding authors. Tel.: +1 650 855 5697; fax: +1 650 855 6585 (H.Y.); tel.:
+1 650 852 1947 (D.M.R.).
X = H, H or O
E-mail addresses: hanbiaoyang@gmail.com (H. Yang), david.rotstein@roche.com
(D.M. Rotstein).
Figure 1.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.10.115

210
H. Yang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 209–213
OLi
OEt
n
-Bu
CN
2
+
CO2Et
THF, -78 oC
N
NH
X
N
Ph
90%
CO2Et
CN
CO2Et
Ph
n-Bu
N
n
-Bu
4 X = H, H
5 X = O
Ph
3
Scheme 1.
Route 1:
O
b, c
a
N
N
N
NH
Ph
Ph
4
n-Bu
6
n-Bu
O
X
CN
N
O
Cl
R
d
BocN
N
BocN
N
NH
X= CO or SO2
e
8
n-Bu
7
n-Bu
O
O
f, g
N
N
N
X
BocN
N
N
X
R1
R
R
n
-Bu
10-23
9
n-Bu
see Tables 1 and 2 for structures
Route 2:
h, a
i
NBoc
N
N
N
Ph
n-Bu
n-Bu
25
24
O
CN
N
X
N
N
NBoc
CbzN
N
R
N
n-Bu
26
n-Bu
N
27-28
see Tables 1 and 2 for structures
Scheme 2. Reagents and conditions: (a) EDCI, HOBT, DIPEA, tetrahydropyran-4-carboxylic acid, DCM, rt, 71%; (b) Pd/C, EtOH, H2 (60 psi); (c) N-Boc-4-piperidone, Ti(O-iPr)4,
DCE; then Et2AlCN; (d) CH3MgBr, THF, rt, 58% over three steps; (e) triethyl amine, DCM; (f) TFA, DCM, rt; (g) EDCI, HOBT, DIPEA, acid, DCM, rt; (h) Boc2O, iPr2EtN, DCM;
(i) N-Cbz-4-piperidone, Ti(O-iPr)4, DCE; then Et2AlCN.
delight, under phase transfer conditions (solid NaOH, K2CO3, tetra-
butylammonium bromide (TBAB), 2 equiv of alkyl bromide or tos-
ylate, toluene, 90 °C), the alkylation occurred smoothly to give
intermediate 29, which was transformed to analogs 30–37
(Table 3) following procedures outlined in Scheme 2.
For the 3,9-diazaspiro[5.5]undecane 4-based analogs, the choice
of types of R2 group was important for antiviral activity (Table 1).
For example, compound 10 with an amide on the right-hand
nitrogen was inactive. Replacing the left-hand 4,6-dimethyl-5-pyr-
imidinecarboxyamide in 10 with 6-cyano-2,4-dimethyl-3-pyridin-
ecarboxyamide in 11 reinstalled some of the activity
(CCFIC50 = 170 nM). In addition, carbamate 12 and amine 13 were
also inactive. On the other hand, benzenesulfonamide 14 and urea
27 were quite potent (CCFIC50 < 1 nM and =4 nM, respectively).
While both 14 and 27 had promising potency in CCF, screening
in our definitive antiviral assay7 showed 14 to have superior po-
tency (CCFIC50s: 1.6 nM for 14 and 53 nM for 27).
The requisite alkyl tosylates or bromides that were utilized in
Scheme 2 were available commercially or prepared according to
Scheme 3. Reduction of 4-methoxycyclohexyl carboxylic acid with
NaBH4 and BF3 etherate in THF gave alcohol 38, which was tosylat-
ed to give isomers 39 and 40. Separation of 39 and 40 was achieved
by flash column chromatography on silica gel (Scheme 4).
The synthesized compounds were screened for CCR5 antago-
nism activity by using radiolabelled RANTES binding assay and
CCR5-mediated antiviral activities employing a cell–cell fusion
(CCF) assay.6,7 Our initial structure–activity relationship (SAR)
studies on the 1-oxa-3,9-diazaspiro[5.5]undecan-2-one series sug-
gested that a hindered amide (R1, Table 1) with restricted rotation
was optimal for antiviral activity, therefore, only a few selected R1
groups were employed throughout our study and emphasis was
placed on finding the optimal R2 group.
Further profiling of 14 identified its weakness—high clearance
in in vitro human liver microsomal assay (Table 2). Replacement
of the phenyl group with 2-pyridyl (compound 15) did improve
the microsomal stability without much loss in potency
(CCFIC50 = 4 nM). However, the clearance (310 ll/min/mg) was still
considered to be high. The methyl sulfonamide 16 was stable, but
unfortunately, a 100-fold loss in potency was observed compared
to 14. A bulkier cyclopropyl group (compound 17) increased po-

H. Yang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 209–213
211
Table 1
Table 2
RANTES binding and cell–cell fusion activity of diazaspiroundecane analogs.
RANTES binding, cell–cell fusion activity and HLM of diazaspiroundecane sulfonamide
series.
O
R2
O
N
N
N
O
O
N
N
N
S
R1
n-Bu
R
N
n-Bu
N
Compound
R1
R2
IC50 (nM)
RANTES
CCFa
Compound
R
IC50 (nM)
HLMb
773
310
7
(ll/min/mg)
RANTES
19
CCFa
<1
O
O
10
11
>500
217
>5000
O
O
N
14
15
16
17
18
19
20
21
22
23
N
170
18
25
22
60
22
27
11
78
64
19
4
100
7
NC
N
N
N
O
O
12
349
>5000
O
N
100
207
280
96
13
14
>500
19
>5000
<1
O
N
N
N
N
O
O
7
N
S
O
6
O
27b
35
4
N
N
H
N
a
N
Values are means of two experiments.
7
b
(R)-Enantiomer.
N
1
276
70
N
O
tency (CCFIC50 = 7 nM), however, at the expense of microsomal sta-
bility. Saturated heterocycles, such as 4-substituted tetrahydropy-
ran (compound 19), morpholine (compound 21), and 4-substituted
piperidines (compounds 22 and 23), are tolerated. To our disap-
pointment, none of those analogs provided desirable combination
of potency and microsomal stability.
O
10
14
3
N
N
O
H
Parallel investigation of 3,9-diazaspiro[5.5]undeca-2-one series
proved to be quite successful (Table 3). Compound 30 retained
CCR5 antiviral potency comparable to its 1-oxa-3,9-diazaspi-
ro[5.5]undecan-2-one counterpart (structure 1, Fig. 1). This con-
firms our earlier hypothesis that the cyclic carbamate moiety in
1 served as a linker to place all the substituents at the proper ori-
entation. Replacement of the 4-tetrahydropyran with a para-fluo-
rophenyl (compound 31) led to a substantial loss of potency
(CCFIC50 = 100 nM). On the other hand, the tetrahydropyran moi-
ety can be replaced by a cyclohexyl group with 1,4-trans substitu-
tion (compounds 32–34). Both ethers (methyl and ethyl) and the
polar hydroxyl group are tolerated. In addition, the tetrahydropy-
ran can also be replaced with 4-substituted piperidine (compound
35) with comparable activity. However, the difluoroethyl piperi-
dine analog (compound 36) led to reduced activity.
60
28c
53
CF3
a
Values are means of two experiments.
Human liver microsomal clearance.
(R)-Enantiomer.
b
c
We also investigated modification of the butyl side chain (Table 4).
Replacement of the butyl with either propoxy (compound 41) or
phenyl (compound 42) led to a substantial loss of activity. This is
in agreement with our CCR5 homology model where the butyl group
binds in a narrow hydrophobic pocket within CCR5 receptor.8
O
a or b
N
N
N
O
N
N
O
N
NH
O
R
Ph
R
Ph
R1
n-Bu
n-Bu
n-Bu
30-37
29
5
see Table 3 for structures
Scheme 3. Reagents and conditions: (a) NaH, alkyl bromide or tosylate, DMF 100 °C; (b) solid NaOH, TBAB, K2CO3, toluene, 90 °C.

212
H. Yang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 209–213
Table 3
Table 4
RANTES binding and cell–cell fusion activity of diazaspiroundecane series.
Effect of side-chain modification.
O
O
N S
O
N
O
N
N
O
R2
N
N
R1
R2
N
N
n
-Bu
R1
Compound
R1
R2
IC50 (nM)
Compound
R1
R2
IC50 (nM)
RANTES
CCFa
RANTES
CCFa
400
N
30
30
19
30
37
76
60
4
O
O
41
80
N
N
N
N
N
N
N
42
104
1900
N
N
N
N
N
N
31
32
33
34
35
36
100
F
a
Values are means of two experiments.
Table 5
Pharmacokinetic properties of 32 in rat.
5
OEt
Parameter
IV (0.5 mg/kg)a
PO (2 mg/kg)a
Tmax (h)
Cmax (ng/mL)
T1/2 (h)
—
—
1.00
203
4.19
4.33
910
—
3.80
2.65
334
24.1
4.71
—
6
OMe
MRT (h)
AUC (ng h/mL)
CL (mL/min/kg)
Vdss (L/kg)
%F
—
66
a
10
8
OH
Values are means of two experiments.
tion for CCR1, CCR2, CCR3, CCR4, CCR6, and CXCR4 (all with
IC50 > 50 M). In addition, 32 showed no significant CYP inhibition
against major subtypes such as 3A4, 1A2, 2C9, 2C19, and 3D6. Fur-
l
O
N
O
thermore, 32 had no significant hERG inhibition activity with an
in vitro IC20 value of 3 lM at 37 °C.
Encouraged by the rat PK and selectivity data for 32, we pur-
sued an asymmetric synthesis (Scheme 5). One of the key steps
was alkylation of the chiral template 4310 (99% ee) with tosylate
39. Surprisingly, under the previously employed phase transfer
conditions (solid NaOH, tetrabutylammonium bromide, K2CO3, tol-
uene, 90 °C), the desired product 44 was obtained in merely 50% ee.
After extensive optimization, it was found that running the reac-
tion at room temperature for 2 d gave 44 in 94% yield and 96%
ee, which was further enriched to >99% ee by chiral HPLC purifica-
tion.11 Compound 44 was then converted to (R)-enantiomer 45 fol-
lowing procedures outlined in Scheme 1. The (S)-enantiomer 46
was prepared in a same fashion.
70
31
26
N
CH2CHF2
OH
37
<2.5
NC
N
a
Values are means of two experiments.
OEt
OEt
Compounds 32, 45 and 46 were evaluated in peripheral blood
mononuclear cell (PBMC) viral replication assay with 10% FBS
using an R5-tropic HIV-1Ba-L strain. The results were shown in Ta-
ble 6. Interestingly, both enantiomers of 32 displayed similar activ-
ities with IC90 values of 4 and 3 nM for 45 and 46, respectively. This
result is in sharp contrast to our earlier series represented by gen-
eral structure 1, in which only the (S)-enantiomer is active. Efforts
to explain this discrepancy by invoking our homology model8 are
inconclusive due to its low resolution.
a
b
EtO
+
EtO
CO2H
OH
38
TsO
TsO
40
39
Scheme 4. Reagents and conditions: (a) NaBH4, BF3 etherate, THF, reflux, 90%; (b)
p-toluenesulfonyl chloride, pyridine, rt, 7% of 39 and 7% of 40.
In conclusion, replacement of the spirocyclic carbamate in our
1-oxa-3,9-diazaspiro[5.5]undecan-2-one series led to the discov-
ery of two novel series of CCR5 antagonists. While the 3,9-diazaspi-
ro[5.5]undecane series provided potent compounds in CCF assay,
its further development was hampered by high in vitro human
microsomal clearance. On the other hand, 3,9-diazaspi-
ro[5.5]undeca-2-one series also yielded potent analogs. Analog
32 showed an attractive combination of antiviral potency, selectiv-
Several compounds from the 3,9-diazaspiro[5.5]udecan-2-one
series were selected for pharmacokinetic evaluation in rat. Among
them, 32 displayed attractive pharmacokinetic properties as
shown in Table 5. It has 66% bioavailability and an half life of
4.19 h (PO), which compared favorably with other clinically ad-
vanced CCR5 antagonists.9 Compound 32 was also quite selective
against other chemokine receptors: it inhibits RANTES binding to
CCR5 with an IC50 value of 24 nM but shows no significant inhibi-

H. Yang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 209–213
213
N
O
N
solid NaOH, 39, TBAB
NH
O
N
Ph
K2CO3, toluene, r.t.
Ph
n
-Bu
n
-Bu
44
94% yield, 96% ee
43 (>99% ee)
OEt
O
N
O
N
N
O
N
N
N
O
N
N
N
n-Bu
n-Bu
N
45
46
OEt
OEt
Scheme 5.
Table 6
References and notes
Peripheral blood mononuclear cell (PBMC) viral replication assay results for
compounds 32, 45 and 46.
1. Perno, C.-F.; Moyle, G.; Tsoukas, C.; Ratanasuwan, W.; Gatell, J.; Schechter, M. J.
Med. Virol. 2008, 80, 565.
Compound
Chirality
PBMC assaya
2. For reviews, see: (a) Repik, A.; Richard, K. H.; Clapham, P. R. Curr. Opin. Investig.
Drugs 2007, 8, 130; (b) Palani, A.; Tagat, J. J. Med. Chem. 2006, 49, 2851; (c)
Kazmierski, W.; Bifulco, N.; Yang, H.; Boon, L.; DeAnda, F.; Watson, C.; Kenakin,
T. Bioorg. Med. Chem. Lett. 2003, 11, 2663. and references therein.
3. For recent publications, see: (a) Ernst, J.; Dahl, R.; Lum, C.; Sebo, L.; Urban, J.;
Miller, S. G.; Lundström, J. Bioorg. Med. Chem. Lett. 2008, 18, 1498; (b) Thoma,
G.; Beerli, C.; Bigaud, M.; Bruns, C.; Cooke, N. G.; Streiff, M. B.; Zerwes, H.-G.
Bioorg. Med. Chem. Lett. 2008, 18, 2000. and references therein.
4. For a patent application, see: Gabriel, S. D.; Rotstein, D. M. US 2005176703 A1,
2005; Chem. Abstr. 2005, 143, 211895.
IC50 (nM)
IC90 (nM)
32
45
46
Racemic
R
S
1.4
0.8
0.5
7.5
4
3
a
Values are means of two experiments.
ity, and a favorable pharmacokinetic profile. In addition, asymmet-
ric synthesis of 32 was accomplished and both enantiomers were
equally potent in PBMC assay with IC90 values of 4 and 3 nM,
respectively.
[5]. Yang, H.; Lin, X-F.; Padilla, F.; Rotstein, D. M. Tetrahedron Lett. 2008, 49, 6371.
6. Ji, C.; Zhang, J.; Cammack, N.; Sankuratri, S. J. Biomol. Screen. 2006, 11, 65.
7. Compounds active in cell–cell fusion assay are in general also potent in RANTES
binding assay. However, the ranking of compounds in these two assays does
not agree well. In the SAR discussion, only IC50 from cell–cell fusion assay will
be used. For a similar observation, see Ref. 3a.
Acknowledgments
8. Kondru, R.; Zhang, J.; Ji, C.; Mirzadegan, T.; Rotstein, D.; Sankuratri, S.; Dioszegi,
M. Mol. Pharmacol. 2008, 73, 289.
9. Walker, D.; Abel, S.; Comby, P.; Muirhead, G.; Nedderman, A.; Smith, D. Drug
Metab. Dispos. 2005, 33, 587.
10. For an asymmetric synthesis of 43, see Ref. 5.
11. Chiral HPLC condition: chiralpak AD column, 95/5 hexane/EtOH as solvent at
1.4 mL/min. Compound 44 has a retention time of 9.29 min.
The authors thank Eric Sjogren for a critical review of this man-
uscript, Marquis L. Cummings, Jr. for chiral HPLC purification,
Yanzhou Liu for NMR analysis, Lijing Chen for synthesizing com-
pound 46, and Naina Patel for formulation study.