Discovery and SAR of spirochromane Akt inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.02.073

A novel series of spirochromane pan-Akt inhibitors is reported. SAR optimization furnished compounds with improved enzyme potencies and excellent selectivity over the related AGC kinase PKA. Attempted replacement of the phenol hinge binder provided compounds with excellent Akt enzyme and cell activities but greatly diminished selectivity over PKA.

Full text of DOI:10.1016/j.bmcl.2011.02.073

Bioorganic & Medicinal Chemistry Letters 21 (2011) 2410–2414
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery and SAR of spirochromane Akt inhibitors
Nicholas C. Kallan a, , Keith L. Spencer a, James F. Blake a, Rui Xu a, Justin Heizer a, Josef R. Bencsik a,
⇑
Ian S. Mitchell a, Susan L. Gloor a, Matthew Martinson a, Tyler Risom a, Stefan D. Gross a, Tony H. Morales a,
Wen-I Wu a, Guy P. A. Vigers a, Barbara J. Brandhuber a, Nicholas J. Skelton b
a Array BioPharma, Inc., 3200 Walnut Street, Boulder, CO 80301, USA
b Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080-4990, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of spirochromane pan-Akt inhibitors is reported. SAR optimization furnished compounds
with improved enzyme potencies and excellent selectivity over the related AGC kinase PKA. Attempted
replacement of the phenol hinge binder provided compounds with excellent Akt enzyme and cell activ-
ities but greatly diminished selectivity over PKA.
Received 17 December 2010
Revised 15 February 2011
Accepted 16 February 2011
Available online 18 February 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keyword:
Akt kinase inhibitors
Akt, also known as protein kinase B (PKB), is an intracellular
anti-apoptotic serine/threonine kinase in the AGC kinase family,
and it is comprised of three isoforms: Akt1, Akt2, and Akt3. These
isoforms have ca. 90% sequence identity in the kinase domain and
97–100% sequence identity in the ATP-binding site.1 The PI3K/Akt
pathway is a major signaling axis for cell growth, proliferation, and
apoptosis, and it is upregulated in many cancers.2 Aberrant Akt sig-
naling can occur via a number of mechanisms, including oncogenic
PI3K mutations, loss of PTEN tumor suppressor function, Akt over-
expression, and therapeutic-induced pathway activation. Owing to
these factors, significant interest exists in finding inhibitors of
components of this pathway for cancer treatment.3
O
O
O
O
HO
HO
N
N
MeO
N
N
SO2 OH
SO2 OH
N
O
N
O
1
2
A
high throughput screening (HTS) campaign identified
M hits (Fig. 1).4 Interestingly, further
µ
µ
M
Akt1 IC50 3.4
PKA IC50 > 85
M
M
Akt1 IC50 3.9
compounds 1 and 2 as low
l
µ
profiling of compound 1 showed that it displayed selectivity
(>25-fold) versus the closely related AGC family kinase protein ki-
nase A (PKA). Achieving general kinase selectivity, especially for
Akt over PKA, is known to be both important and challenging.5 This
became a key driver for further exploring structure–activity rela-
tionships (SAR) in this class of compounds.
Based on HTS hits 1 and 2, an initial set of analogs was prepared
as shown in Scheme 1. Substituted sulfonyl chlorides were reacted
with a variety of primary amines to give the corresponding
sulfonamides 3, which were then deprotonated with NaH and
treated with (R)-epichlorohydrin to generate epoxides 4. The
racemic spirochromane core was prepared by condensing
2050-dihydroxyacetophenone with 1-Boc-3-piperidone to afford
Boc intermediate 5. Boc removal with HCl/dioxane, followed by
Figure 1. HTS hits 1 and 2.
treatment with 4 furnished spirochromanes 7–21 as 1:1 mixtures
of diastereomers at the spiropiperidine stereocenter. In addition,
a set of des-keto spirochromane analogs was also prepared. The ke-
tone in 5 was reduced with NaBH4 to give the corresponding ben-
zylic alcohol, which was removed by hydrogenolysis with Pd/C
under an atmosphere of hydrogen to provide 6. Boc deprotection
with HCl/dioxane, followed by treatment with 4, furnished analogs
22 and 23 as 1:1 mixtures of diastereomers at the spiropiperidine
stereocenter.
Compounds 7–23 (Tables 1 and 2) were evaluated in vitro for
inhibition of Akt16 and in LNCaP cells for formation of phosphory-
lated PRAS40.7 A modest potency difference was observed between
secondary hydroxyl diastereomers (Table 1, compare 7, prepared
⇑
Corresponding author.
E-mail address: nkallan@arraybiopharma.com (N.C. Kallan).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.02.073

N. C. Kallan et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2410–2414
2411
O
O
HO
O
HO
c
e
HO
N
O
O
OH
N
Boc
N
6
5
Boc
Boc
O2
S
b
a
SO2Cl
O2
S
R2
R2
N
R1
d
d
HN
R1
R2
O
NH2
R1
Cl
O
4
3
O
O
HO
HO
O
N
N
R1
R2
R1
R2
N
N
SO2 OH
SO2 OH
22-23
7-21
Scheme 1. Reagents and conditions: (a) TEA, DCM; (b) NaH, DMF; (c) pyrrolidine, MeOH, 50 °C; (d) (i) HCl/dioxane; (ii) 4, TEA, EtOH, 60 °C; (e) (i) NaBH4, MeOH; (ii) Pd/C,
1 atm H2, EtOH.
from (R)-epichlorohydrin, and 8, prepared from (S)-epichlorohy-
drin); thus all further analogs were prepared using (R)-epichloro-
hydrin. The spiropiperidine diastereomers of compound 7 were
also separated via preparative HPLC (7-d1 and 7-d2), and a 20-fold
difference in potency was observed between them.8 For simplicity,
the rest of the compounds were assayed as 1:1 mixtures of diaste-
reomers at the spiropiperidine stereocenter. We also found that re-
moval of the secondary alcohol resulted in >30-fold loss in potency
(data not shown).
HTS lead 1 was preserved in all of the optimized analogs. All com-
pounds in Tables 1 and 2 had a PKA enzyme IC50 >10 lM, with
compound 23 showing >250-fold selectivity.
The more potent phenol-based analogs in Table 2 achieved
excellent PKA selectivity and reasonable enzyme potency, but none
attained cell potency below 2.6 lM observed for 23. Analysis of
several X-ray crystal structures of phenol-based analogs bound to
Akt1 indicated that the phenol OH forms less than optimal hydro-
gen bonding interactions with the hinge residues (Glu228 and
Ala230) due to relatively elongated acceptor–donor distances (cf.,
Fig. 2, compound 21). The phenol O is located 3.5 Å from the nitro-
gen of Ala230 and 3.9 Å from the backbone carbonyl oxygen of
Glu228. Phenol replacements were thus explored in an attempt
For the amine side chain R1, the best substituents were small al-
kyl ethers (Table 1). Ethoxyethyl analog 9 (1.0
oxyethyl derivative 10 (0.3 M) showed improved potencies
relative to methoxymethyl compound 7 (6.5 M), while larger
lM) and isoprop-
l
l
ether substituents lost potency (data not shown). Alkyl groups
gave similar potencies to the corresponding ethers (e.g., 9 vs 11),
but the ether derivatives were preferred due to their lower ClogP
values. Cycloalkyl substituents were also examined and lost po-
tency relative to the ethers (cf. 12). In addition, benzyl derivatives
were explored and showed modest potencies, but these were not
investigated further in order to keep both MW and ClogP low (data
not shown). Ultimately, the ethoxyethyl side chain provided the
optimum balance between potency, size, and ClogP, and this amine
substitution was preserved as we probed sulfonamide SAR.
Modeling suggested that the sulfonamide group would bind in
the somewhat narrow hydrophobic P-loop region of Akt, and in
general, we found that aromatic and heteroaromatic rings substi-
tuted with small lipophilic groups were best. While the meta-
and para-chlorophenyl derivatives 13 and 14 were inactive
Table 1
Amine side chain SAR
O
HO
O
N
R1
N
SO2 OH
N
O
(P20
was 1.0
l
M), the Akt1 potency of the ortho-chlorophenyl analog 15
R1
Akt1 IC50
(
lM)
pPRAS40 IC50 (lM)
a
b
Compd
lM (Table 2). Although larger groups were not tolerated
(data not shown), further exploration at the ortho position pro-
duced sub-micromolar analogs, including the 2,6-dimethyl deriva-
tive 21 (0.3 lM against Akt1). Benzyl and larger bicyclic aromatic
ring systems all lost potency relative to the substituted monocy-
cles, and removal of the arylsulfonyl group completely abolished
activity (data not shown). Turning to the spirochromanone core,
removal of the ketone resulted in further improvement in enzyme
potency (e.g., 9 vs 22). Thus, combining 2,6-dimethyl phenyl sub-
stitution with ketone removal provided the most potent analog
to this point (23). Gratifyingly, the PKA selectivity of the initial
7
7-d1
7-d2
8c
9
10
CH2CH2OMe
CH2CH2OMe
CH2CH2OMe
CH2CH2OMe
CH2CH2OEt
CH2CH2OiPr
CH2(CH2)3CH3
Cyclohexyl
6.5
2.1
40
22.8
1.0
0.3
1.1
4.0
>25
>25
>25
>25
>25
18.6
>25
>25
11
12
a
b
c
Data is an average of three determinations.
Data is an average of two determinations.
(S)-Epichlorohydrin was used in Scheme 1, Step b.

2412
N. C. Kallan et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2410–2414
Table 2
Sulfonamide and core modifications
HO
X
O
N
EtO
N
SO2 OH
R2
a
b
Compd
R2
X
Akt1 IC50
(lM)
pPRAS40 IC50
>25
(lM)
Cl
13
C@O
>20
14
15
16
17
18
C@O
C@O
C@O
C@O
C@O
ꢀ20
1.0
>25
>25
>25
>25
>25
Cl
Figure 2. X-ray structure overlays of 21 (cyan) and 38 (green) bound to Akt1 (PDB
accession codes 3QKK and 3QKL, respectively). Residue Lys179 was removed for
clarity.
Cl
>20
6.2
40-cyano-20-hydroxy-acetophenone was first converted to spi-
rochromane intermediate 31.10 Nitrile reduction with NaBH4/
CoCl2, followed by amine protection gave trifluoroacetamide 32
Iodination with I2/HIO3, followed by deprotection furnished amine
33. Intramolecular lactam formation was effected using a Pd-cata-
lyzed carbonylation to give 34, which after Boc removal and reac-
tion with epoxide 26 provided lactam 35.
In addition to phenol replacements, indazole was explored as a
surrogate (Scheme 4). Diazotization of 4-methoxy-2-methylani-
line, followed by indazole formation gave 36. Friedel–Crafts acyla-
tion (with concomitant phenol deprotection) and subsequent
condensation with Boc-3-piperidone furnished pyrazolo-spirochr-
omane 37. Finally, ketone removal followed by Boc deprotection
and reaction with epoxide 26 provided 38.
The potencies of the phenol replacements are shown in Table 3.
Fluoro and nitrile analogs 27 and 29 exhibited only a slight reduc-
tion in Akt1 enzyme potency compared to phenol 23, although PKA
selectivity diminished somewhat to 50-fold. Interestingly, primary
amide 30 showed improved Akt1 enzyme potency while retaining
PKA selectivity (>385-fold), though its cell potency was still mod-
est. However, in contrast to amide 30, constraining the amide into
a lactam (35) greatly improved cell potency, but unfortunately the
PKA selectivity eroded to 5-fold. In addition, indazole 38 showed
excellent enzyme and cell potencies, but again the PKA selectivity
diminished to 3-fold.
OMe
F
4.4
CN
CF3
19
20
21
C@O
C@O
C@O
1.3
0.5
0.3
>25
>25
17.8
22
23
CH2
CH2
0.07
0.04
6.4
2.6
N
O
a
Data is an average of three determinations.
Data is an average of two determinations.
b
In addition to X-ray crystal structures of the phenol-based ser-
ies of inhibitors (e.g., 21), we obtained a structure of 38 bound to
Akt1 (Fig. 2). The secondary hydroxyls and protonated piperidine
nitrogens of both compounds interact with the side chain carbox-
ylate of Asp292. The secondary hydroxyls also interact with the
catalytic Lys179. The 2,6-dimethyl phenyl groups pack underneath
the hydrophobic P-loop, and the ethoxyethyl side chains form lipo-
philic contacts with the exposed side chain of Phe161 at the tip of
this loop, in addition to a hydrogen bond that is formed between
the backbone NH and one of the sulfonamide oxygens.
The data presented above highlight a distinct difference be-
tween inhibition of Akt1 and PKA, in that inhibition of Akt1 is
not greatly affected by hinge contacts (e.g., 4-fold potency differ-
ence between phenol 23 and indazole 38), whereas PKA is greatly
affected (>330-fold potency difference between 23 and 38). Exam-
ination of Figure 2 shows that all of the other components in both
phenol 21 and indazole 38 — the spiropiperidines, secondary alco-
hols, sulfonamides, and amine side chains — overlay almost
exactly, and the only differences lie in the hinge region. For
indazole 38 bound to Akt1, the indazole NH to Glu228 carbonyl
to improve potency and to remove the potential metabolic liabili-
ties of this moiety.
Unsubstituted and fluoro-spirochromane intermediates 24 and
25 were prepared using 20-hydroxyacetophenone and 50-fluoro-
20-hydroxyacetophenone, respectively (Scheme 2). The Boc group
in 25 was removed, followed by treatment with epoxide 269 to fur-
nish fluoro analog 27. A nitrile phenol replacement was synthe-
sized in a similar manner. Bromination of intermediate 24 with
NBS, followed by lithiation of the resulting bromide with tBuLi
and subsequent treatment with tosyl cyanide afforded nitrile inter-
mediate 28. Boc removal with HCl/dioxane, followed by reaction
with epoxide 26 gave 29. Nitrile 28 was also converted to the pri-
mary amide by treatment with TFA/H2SO4 and subsequent reaction
with epoxide 26 to afford 30.
Other phenol replacements were investigated, such as con-
straining primary amide 30 into a lactam ring (Scheme 3). Thus,

N. C. Kallan et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2410–2414
2413
X
O
X
X
a
b, c
O
O
EtO
N
EtO
N
OH
24 X = H
25 X = F
Boc
N
X = H, F
N
SO2
SO2 OH
O
24
26
27 X = F
d
O
NC
H2N
NC
O
EtO
e, b, c
O
b, c
N
EtO
N
O
N
N
N
SO2 OH
28
Boc
SO2 OH
29
30
Scheme 2. Reagents and conditions: (a) (i) Boc-3-piperidone, pyrrolidine, MeOH, 50 °C; (ii) NaBH4, MeOH; (iii) Pd/C, 1 atm H2, EtOH; (b) HCl/dioxane; (c) 26, TEA, EtOH,
60 °C; (d) (i) NBS, 10 mol % 1 M HCl, acetone; (ii) tBuLi, THF, ꢁ78 °C; TsCN; (e) TFA, H2SO4; then Boc2O, dioxane.
O
H
a
b
c
O
N
NC
O
O
N
CF3
N
NC
OH
31
Boc
32
Boc
O
O
HN
I
O
d
HN
e
EtO
H2N
N
O
O
N
N
N
33
Boc
34
Boc
SO2 OH
35
Scheme 3. Reagents and conditions: (a) (i) Boc-3-piperidone, pyrrolidine, MeOH, 50 °C; (ii) NaBH4, MeOH; (iii) Et3SiH, BF3–OEt2, DCM; (iv) Boc2O, TEA, DCM; (v) Pd/C, 1 atm
H2, EtOAc; (b) (i) NaBH4/CoCl2, EtOH; (ii) TFAA, TEA, DCM; (c) (i) I2, HIO3, MeOH/H2O, reflux; (ii) LiOH–H2O, MeOH/H2O; (d) Pd(PPh3)4, TEA, 1 atm CO, THF, 60 °C; (e) (i) HCl/
dioxane; (ii) 26, TEA, EtOH, 60 °C.
H
N
Table 3
H2N
a
b
N
Potency and selectivity of phenol replacement analogs
OMe
a
a
b
Compd
Akt1 IC50
(
lM)
PKA IC50
(
lM)
PKA/Akt1
pPRAS40 IC50
M)
OMe
36
(
l
27
29
30
35
38
0.18
0.17
0.026
0.010
0.009
>10
8.99
>10
0.053
0.030
>55
54
>385
5
4.38
2.58
2.36
0.035
0.044
N
HN
N
HN
O
O
3
O
a
EtO
c
Data is an average of three determinations.
Data is an average of two determinations.
N
b
N
N
37
SO2 OH
Boc
38
oxygen distance is 2.8 Å, while the indazole N to Ala230 NH is
2.9 Å. From an X-ray co-crystal of 38 bound to PKA (2.9 Å
resolution), the distances are similar (indazole NH to Glu228
carbonyl oxygen 3.2 Å, indazole N to Ala230 NH 3.1 Å; structure
not shown). Consequently, we hypothesize that the improved
Scheme 4. Reagents and conditions: (a) (i) NH4BF4, NaNO2, HOAc/H2O, concd HCl,
0 °C to rt; (ii) KOAc, EtOAc; (b) (i) AlCl3, AcCl, DCE, 60 °C; (ii) Boc-3-piperidone,
pyrrolidine, MeOH, 50 °C; (c) (i) NaBH4, MeOH; (ii) Et3SiH, BF3–OEt2, DCM; then
HCl/dioxane; (iii) 26, TEA, EtOH, 60 °C.

2414
N. C. Kallan et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2410–2414
hinge interactions observed for indazole 38 greatly benefit PKA
inhibition while having modest effects on Akt potency, and thus
PKA selectivity erodes for 38. Note also that while amide 30 and
lactam 35 have comparable Akt1 potencies, their PKA potencies
are vastly different, suggesting that conformational constraint
and the precise location of the hydrogen bond donor greatly influ-
ences PKA binding while not affecting Akt1. Taken together, these
data reinforce the notion that interactions in the hinge region are
particularly sensitive for PKA inhibition and that this phenomenon
could be exploited to find selective compounds.
In summary, a novel series of spirochromane Akt inhibitors was
discovered via HTS. Early SAR with a phenol hinge binder showed
that optimization of the sulfonamide and amine side chain regions,
as well as removal of the ketone from the spirochromane core, im-
proved Akt1 enzyme potency and retained >250-fold selectivity
over PKA. Further exploration of phenol replacements, such as a
lactam and an indazole, produced compounds with much im-
proved cell potency, but the PKA selectivity diminished to 5-fold.
Select analogs were also examined for in vitro metabolism studies,
and most showed very poor microsomal stability across species,
which precluded further progression of these compounds (data
not shown). Efforts to find a new chemical series that combines
favorable Akt cell potency with high PKA selectivity will be re-
ported in a forthcoming publication.
References and notes
1. (a) Hanks, S.; Hunter, T. FASEB J. 1995, 9, 576; (b) Zinda, M. J.; Johnson, M. A.;
Paul, J. D.; Horn, C.; Konicek, B. W.; Lu, Z. H.; Sandusky, G.; Thomas, J. E.;
Neubauer, B. L.; Lai, M. T.; Graff, J. R. Clin. Cancer Res. 2001, 7, 2475; (c) Masure,
S.; Haefner, B.; Wesselink, J. J.; Hoefnagel, E.; Mortier, E.; Verhasselt, P.;
Tuytelaars, A.; Gordon, R.; Richardson, A. Eur. J. Biochem. 1999, 265, 353.
2. For recent reviews, see: (a) Wickenden, J. A.; Watson, C. J. Breast Cancer Res.
2010, 12, 202; (b) Bowles, T. L.; Parsons, C.; Muilenburg, D. J.; Bold, R. J. Curr.
Cancer Ther. Rev. 2009, 5, 288; (c) Sarker, D.; Reid, A. H. M.; Yap, T. A.; de Bono, J.
S. Clin. Cancer Res. 2009, 15, 4799; (d) Franke, T. F. Oncogene 2008, 27, 6473;
(e)Oncogene 2008, 27, 5511; (f) Steelman, L. S.; Stadelman, K. M.; Chappell, W.
H.; Horn, S.; Bäsecke, J.; Cervello, M.; Nicoletti, F.; Libra, M.; Stivala, F.; Martelli,
A. M.; McCubrey, J. A. Expert Opin. Ther. Targets 2008, 12, 1139; (g) Manning, B.
D.; Cantley, L. C. Cell 2007, 129, 1261.
3. (a) Lindsley, C. W. Curr. Top. Med. Chem. 2010, 10, 458; (b) Lindsley, C. W.;
Barnett, S. F.; Layton, M. E.; Bilodeau, M. T. Curr. Cancer Drug Targets 2008, 8, 7;
(c) Lindsley, C. W.; Barnett, S. F.; Yaroschak, M.; Bilodeau, M. T.; Layton, M. E.
Curr. Top. Med. Chem. 2007, 7, 1349. and references cited therein.
4. Compounds 1 and 2 are mixtures of four stereoisomers.
5. (a) Blake, J. F.; Kallan, N. C.; Xiao, D.; Xu, R.; Bencsik, J. R.; Skelton, N. J.; Spencer,
K. L.; Mitchell, I. S.; Woessner, R. D.; Gloor, S. L.; Risom, T.; Gross, S. D.;
Martinson, M.; Morales, T. H.; Vigers, G. P. A.; Brandhuber, B. J. Bioorg. Med.
Chem. Lett. 2010, 20, 5607; (b) Bencsik, J. R.; Xiao, D.; Blake, J. F.; Kallan, N. C.;
Mitchell, I. S.; Spencer, K. L.; Xu, R.; Gloor, S. L.; Martinson, M.; Risom, T.;
Woessner, R. D.; Dizon, F.; Wu, W.; Vigers, G. P. A.; Brandhuber, B. J.; Skelton, N.
J.; Prior, W. W.; Murray, L. J. Bioorg. Med. Chem. Lett. 2010, 20, 7037.
6. For all of the analogs described herein, the Akt2 and Akt3 potencies followed
similar trends to Akt1 potency, albeit with activities ca. 5- to 10-fold less potent.
7. Potencies in the Akt1 enzyme assay as well as the phosphorylation of PRAS40
in LNCaP cells were measured as described in Ref. 5a.
8. The relative configurations at the spiropiperidine stereocenters of 7-d1 and 7-
d2 were not determined.
Acknowledgements
9. Generated as outlined in Scheme
1
using ethoxyethylamine, 2,6-
dimethylphenylsulfonyl chloride, and (R)-epichlorohydrin.
10. An inseparable mixture of the Boc-deprotected styrene and fully reduced core
were formed during step (a) (iii) in Scheme 3. Thus, Boc reprotection and
hydrogenation were required to provide spirochromane intermediate 31
cleanly.
The authors would like to thank James M. Graham and Dr. Fran-
cis Sullivan for helpful discussions and careful proofreading in the
preparation of this Letter.