Discovery and structure-activity study of a novel benzoxaborole anti-inflammatory agent (AN2728) for the potential topical treatment of psoriasis and atopic dermatitis

Article abstract of DOI:10.1016/j.bmcl.2009.03.007

A series of phenoxy benzoxaboroles were synthesized and screened for their inhibitory activity against PDE4 and cytokine release. 5-(4-Cyanophenoxy)-2,3-dihydro-1-hydroxy-2,1-benzoxaborole (AN2728) showed potent activity both in vitro and in vivo. This compound is now in clinical development for the topical treatment of psoriasis and being pursued for the topical treatment of atopic dermatitis.

Full text of DOI:10.1016/j.bmcl.2009.03.007

Bioorganic & Medicinal Chemistry Letters 19 (2009) 2129–2132
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery and structure–activity study of a novel benzoxaborole
anti-inflammatory agent (AN2728) for the potential topical treatment
of psoriasis and atopic dermatitis
*
Tsutomu Akama , Stephen J. Baker, Yong-Kang Zhang, Vincent Hernandez, Huchen Zhou,
Virginia Sanders, Yvonne Freund, Richard Kimura, Kirk R. Maples, Jacob J. Plattner
Anacor Pharmaceuticals, Inc., 1020 E. Meadow Circle, Palo Alto, CA 94303, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of phenoxy benzoxaboroles were synthesized and screened for their inhibitory activity against
PDE4 and cytokine release. 5-(4-Cyanophenoxy)-2,3-dihydro-1-hydroxy-2,1-benzoxaborole (AN2728)
showed potent activity both in vitro and in vivo. This compound is now in clinical development for
the topical treatment of psoriasis and being pursued for the topical treatment of atopic dermatitis.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 5 January 2009
Revised 2 March 2009
Accepted 4 March 2009
Available online 9 March 2009
Keywords:
Anti-inflammatory
Phosphodiesterase
Oxaborole
Psoriasis
Atopic dermatitis
Psoriasis is a chronic skin disorder caused by inflammatory cell
infiltration into the dermis and epidermis, and is accompanied by
keratinocyte hyperproliferation.1 Once triggered, a strong T-cell re-
sponse is mounted, and a cascade of cytokine and chemokine pro-
duction is induced. Down-regulation of certain cytokines and
chemokines is considered to be a good approach to treatment,
an antibacterial/anti-inflammatory agent8 and AN2690 as an anti-
fungal agent.9,10 As our boron-containing compound library grew
and was screened in various assay systems, a series of 5-pheno-
xybenzoxaborole derivatives was found to exhibit inhibitory activ-
ity against the release of cytokines, such as TNF-a and IFN-c, from
peripheral blood mononuclear cells (PBMCs) stimulated by lipop-
olysaccaride (LPS) or phytohemagglutinin (PHA). We also discov-
ered that this class of compounds inhibited the PDE4 enzyme as
part of its mechanism of action. Herein we describe the discovery
of a novel series of benzoxaborole derivatives including AN2728
(compound 5b, Fig. 1) as an optimized developmental candidate
and we report the structure–activity relationships (SAR) in terms
of PDE4 enzyme inhibition and cytokine release inhibition.
5-Phenoxybenzoxaborole derivatives were synthesized as
shown in Schemes 1 and 2. The diaryl ether scaffold was made
by a nucleophilic aromatic substitution. The formyl group of com-
pound 2 was protected as an acetal to avoid self condensation of 2.
The formyl group of 4 was reduced to the alcohol and protected as
a THP ether. The boron atom was then introduced by halogen–me-
tal exchange with n-BuLi in the presence of borate, which is known
as the in situ quench protocol.11 Upon deprotection of the alcohol
by HCl, the resulting hydroxymethyl group spontaneously cyclized
to afford the desired oxaboroles. The carboxy derivative (5e) was
obtained by the base hydrolysis of the corresponding cyano com-
pound (5b). Amide derivatives (5f and 5g) were synthesized from
5e using regular EDC/HOBt conditions. Some compounds of the
and indeed, the biologics targeting TNF-a demonstrate the effec-
tiveness of this approach.2–4 However, biologics have intrinsic
challenges, such as limited administration route, side effects, qual-
ity control and production cost. Small molecule approaches to treat
psoriasis include systemic or topical steroids, cyclosporine, psora-
len plus UVA (PUVA), retinoids, methotrexete, and vitamin D3 ana-
logs.2,3 Atopic dermatitis is an allergic skin disorder, which is
typically treated with topical steroids, antihistamines, and calci-
neurin inhibitors.5 However, there is still a need for new treatment
with improved safety profile. Recently phosphodiesterase 4 (PDE4)
inhibitors have been in development for such skin diseases. CC-
10004 is in development as an oral treatment for psoriasis and ato-
pic dermatitis.5 AWD-12-281 was, until recently, in development
for the topical treatment of atopic dermatitis.5,6 In addition, roflu-
milast is under Phase 1 development for both diseases.7
We previously reported several classes of boron-containing
compounds as potential therapeutic agents, including AN0128 as
* Corresponding author. Tel.: +1 650 543 7527; fax: +1 650 543 7660.
E-mail address: takama@anacor.com (T. Akama).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.03.007

2130
T. Akama et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2129–2132
F
NO2
Me
NO2
Me
a
R
R
R
+
O
OH
R
F
O
SO3Me
6
7
8
N
N
Cl
OEt
Br
O
Br
H
N
NH
b,c
d-f
HO
N
OH
OMe
O
F
O
9
Me
O
O
10
O
CC-10004
Cl
AWD-12-281
OH
Cl
O
N
OH
B
R
R
g-i
NC
B
11a: H
O
O
O
R
11e: 4-Me
N
O
11b: 3-CN
O
11
11f
: 4-OMe
11g: 4-CF3
H
Cl
11c
: 4-F
O
F
11d: 4-Cl
5b (AN2728)
Roflumilast
Scheme 2. Reagents and conditions: (a) K2CO3, DMF, 70 °C, overnight; (b) SnCl2,
HCl, EtOH; 50 °C, 2 h (2 steps 74–99%; (c) t-BuONO, HBr, CuCN, MeCN, H2O, 0 °C to
rt, overnight (22–86%); (d) NBS, AIBN, CCl4, reflux, 2 h; (e) NaOAc, DMF, 70 °C,
overnight; (f) 1 M NaOH, MeOH, rt, 1 h (3 steps 30–63%); (g) 3,4-dihydro-2H-pyran,
camphorsulfonic acid, CH2Cl2, rt, 2 h (93–99%); (h) (i-PrO)3B, n-BuLi, THF, À78 °C to
rt, 3 h; (i) 6 M HCl, THF, rt, 3 h (22–57%).
Figure 1. PDE4 inhibitors aiming at skin inflammatory diseases.
series (11a–g) were synthesized from phenols (6) and 5-fluoro-2-
nitrotoluene (7) as shown in Scheme 2.
Regioisomers of 4-cyanophenoxy derivatives (17a–c) were syn-
thesized as shown in Scheme 3 in similar ways to make com-
pounds 5.
The compounds were tested against partially purified PDE4
from U937 cells13 and for inhibition of cytokine release from
PBMCs stimulated by LPS or PHA.14 The results are summarized
in Table 1. 5-Phenoxy derivative (11a) showed an IC50 value of
OH
Br
4
F
+
+
NC
CHO
6
b
NC
12
12
13a
(6-F)
13b (4-F)
3
6
Br
6.8
lM against PDE4; however, it did not show significant inhibi-
4
OH
O
tion against cytokine release (IC50 >10
l
M). When various func-
b,c
Br
NC
CHO
tional groups were introduced to the 5-phenoxy group, the
activities changed drastically. 4-Cyano (5b), 4-morpholinocarbonyl
(5g), and 4-trifluoromethyl (11g) derivatives showed potent activ-
ity against PDE4 (IC50 0.49, 0.57 and 0.45 lM, respectively), which
were more potent than the classic PDE4 inhibitor rolipram in this
O
16a
(6-isomer)
F
16b (4-isomer0
16c (3-isomer)
O
15
1b
a
system. Compounds 5b and 5g also showed a broad range of potent
d-f
NC
OH
Br
Br
OH
7
6
O
B
CHO
CHO
HO
HO
O
14
2
3
4
a
17a
(4-isomer)
17b (6-isomer)
17c (7-isomer)
Br
Br
b,c
R
+
R
O
O
Scheme 3. Reagents and conditions: (a) ethylene glycol, p-TsOH, toluene, reflux,
6 h (77%); (b) K2CO3, DMF, 100 °C, overnight (62–99%); (c) 3 M HCl, THF, reflux, 2 h
(62%); (d) NaBH4, MeOH, rt, 1 h; (e) 3,4-dihydro-2H-pyran, camphorsulfonic acid,
CH2Cl2, rt, 2 h; (f) (i-PrO)3B, n-BuLi, THF, À78 °C to rt, 3 h (3 steps 43–69%). (See
above mentioned references for further information.)
F
O
4
CHO
1a-d
OH
B
d-g
O
R
O
5
cytokine release inhibition. Compound 5g showed three to four
a and IL-5, comparable potency
against IL-2 and IFN- , and better activity against IL-10 compared
to rolipram. The IC50 values of compound 5b against the panel
were three to six times less active than those of rolipram. 4-Fluoro-
phenoxy derivative (11c) showed comparable PDE4 inhibition to
11a, and slight inhibition against IL-2 and IL-5 (55% and 53% inhi-
fold higher IC50 values against TNF-
R
c
5a: 2-CN
5b
: 4-CN
5c: 4-SO2NEt2
h
5d
: 4-SO2N(CH2CH2)2NMe
5e: 4-CO2H
5f
: 4-CONEt2
5g: 4-CON(CH2CH2)2O
i
bition at 10 lM, respectively). Introduction of chloro (11d), methyl
(11e), and methoxy (11f) groups resulted in slight improvement of
PDE4 inhibition. However, no significant cytokine release inhibi-
tion was observed. Activity was also diminished when the cyano
group was relocated from para to ortho (5a) or meta (11b)
positions.
Scheme 1. Reagents and conditions: (a) ethylene glycol, p-TsOH, toluene, reflux,
6 h (quant.); (b) K2CO3, DMF, 100 °C, overnight (82–96%); (c) 3 M HCl, THF, reflux,
2 h (80–100%); (d) NaBH4, MeOH, rt, 1 h (quant.); (e) 3,4-dihydro-2H-pyran,
camphorsulfonic acid, CH2Cl2, rt, 2 h (quant.); (f) (i-PrO)3B, n-BuLi, THF, À78 °C to rt,
3 h; (g) 6 M HCl, THF, rt, 3 h (37–44%); (h) 6 M NaOH, MeOH, 1,4-dioxane, reflux,
6 days (79%); (i) diethylamine (for 5f) or morpholine (for 5g), EDCI, HOBt, DMAP,
DMF, rt, overnight (41–70%).
Then the 4-cyanophenoxy group was relocated to other posi-
tions from the 5-position. The 4-position analog (17a) showed

T. Akama et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2129–2132
2131
Table 1
Inhibitory activity of benzoxaboroles against PDE4 and cytokine release
OH
7
4
6
B
O
R
5
Compound
R
IC50
(l
M) or inhibition% at 10
l
Ma
PDE4
TNF-a
IL-2
IFN-c
IL-5
IL-10
5a
5b
5c
5d
5e
5f
5g
11a
11b
11c
11d
11e
11f
11g
17a
17b
17c
18
5-(2-CN–PhO)
5-(4-CN–PhO)
5-(4-SO2 NEt2–PhO)
5-(4-SO2N(CH2CH2)2NMe
5-(4-CO2H–PhO)
5-(4-CONEt2–PhO)
5-(4-CON(CH2CH2)2O–PhO)
5-PhO
5-(3-CN–PhO)
5-(4-F–PhO)
5-(4-Cl–PhO)
5-(4-Me–PhO)
5-(4-OMe–PhO)
5-(4-CF3–PhO)
4-(4-CN–PhO)
6-(4-CN–PhO)
7-(4-CN–PhO)
5-MeO
3.7
0.49
3.4
>30
6.4
1.5
0.57
6.8
4.4
7.7
1.3
2.0
2.7
0.45
6.0
>30
>30
8.5
4.0
4.1
7.8
2.4
>10
>10
>10
4.3
1.9
>10
4.7
53%
>10
>10
>10
12%
8.2
>10
>10
>10
0.50
8.0
5.3
0.54
>10
>10
>10
1.3
0.44
>10
8.6
>10
>10
>10
>10
6.1
>10
>10
>10
>10
0.16
0.61
>10
>10
>10
1.4
0.33
>10
3.3
55%
62%
>10
>10
3.6
0.83
>10
>10
>10
0.43
0.17
>10
3.4
15%
29%
29%
39%
2.2
>10
>10
>10
2.0
0.37
>10
8.9
>10
>10
>10
>10
49%
>10
>10
>10
>10
0.88
3.9
>10
>10
>10
>10
0.23
>10
>10
>10
0.23
Rolipram
0.86
a
IC50 values and inhibition % are calculated from means of at least two experiments.
Table 2
some inhibition against PDE4 (IC50 6.0
l
M) as well as against sev-
Inhibitory activity of 5b against PDE isozymes
eral cytokines (IC50s 3.9 M against IL-2 and 8.2
l
l
M against IL-5).
However, other isomers (17b,c) did not show any activity in these
systems. 5-Methoxy derivative (18)9 that lacks the pendant phenyl
ring did not show any activity, either. Those data indicated the
importance of a substituted phenoxy group at the 5-position for
the inhibition against both PDE4 and cytokine release. As for the
substituent, an electron withdrawing group (EWG), such as cyano,
substituted carbamoyl, and trifluoromethyl groups, showed potent
activity against PDE4. However sulfonamides (5c,d) did not, even
though these are strong EWGs. The sulfonamide groups might be
too bulky to be accommodated. Only cyano (5b) and morpholino-
carbonyl (5g) derivatives showed sub-micro molar IC50 values
Compound
IC50 (l
M)a
PDE1A3 (cAMP)
PDE3Cat
PDE4Cat
PDE7A1
5b
Ref.
6.1
6.4
0.11
2.3d
0.73
2.2e
10b
0.84c
a
IC50 values are calculated from means of three experiments. The substrate
M except for PDE7A1 (0.1 M).
8-Methoxymethyl-IBMX.
Cilostazol.
Rolipram.
concentration was 1
l
l
b
c
d
e
BRL-50481.
Table 3. Compounds 5b and 5g showed significant inhibition
against the ear edema caused by phorbol ester after dosing at
1 mg/ear  2 (78% and 68%, respectively). The efficacy was compa-
rable to that of dexamethasone, suggesting that these compounds
have good anti-inflammatory activity as well as skin penetration in
in vivo). Rolipram showed significant inhibition only at a higher
concentration in this system.
In conclusion, we report the identification and SAR of a new ser-
ies of phenoxybenzoxaborole anti-inflammatory agents. These
compounds demonstrated both inhibitory activity against the
PDE4 enzyme and inflammation-related cytokine release. The
SAR illustrated that the substituted phenoxy group at the 5-posi-
tion was important for both activities. Selected compounds 5b
and 5g demonstrated in vivo efficacy in phorbol ester-induced
mouse ear edema model by topical application. Compound 5b
against TNF-
cytotoxicity against mouse fibroblast-derived L929 cells up to
100 M (data not shown).
a, IL-2, and IFN-c. All these compounds did not show
l
The most potent compounds (5b and 5g) showed similar IC50
values against both PDE4 (biochemical) and cytokine release (cell
based) assays, suggesting these compounds have good cell mem-
brane permeability. Interestingly, trifluoromethyl derivative (11g)
was much less active in the cytokine assay, although it was equally
potent as 5b and 5g against PDE4 enzyme. All the compounds syn-
thesized and tested showed no significant inhibition against IL-1b
and IL-4 release (data not shown). The same observation was seen
with other non-oxaborole PDE4 inhibitors.15
In order to investigate the isozyme selectivity, compound 5b
was then tested against the panel of PDE isozymes (PDE1 through
11) by a modified two-step method using recombinant human PDE
enzymes expressed in a baculoviral system.16 Compound 5b
showed the most potent activity against PDE4 catalytic domain,
but it also showed inhibition against PDE1A3, PDE3Cat, and
PDE7A1. The IC50 values are summarized in Table 2. Compound
5b did not show significant inhibition against other isozymes at
Table 3
Inhibitory activity of 5b and 5g against phorbol ester-induced mouse ear edema
Compound
Dose
Inhibition (%)
5b
5g
1 mg/ear  2
1 mg/ear  2
1 mg/ear  2
1 mg/ear  2
3 mg/ear  2
78
68
72
6
10 lM.
Dexamethasone
Rolipram
Compounds 5b and 5g were tested in phorbol ester-induced
mouse ear edema model17 to determine in vivo efficacy and skin
penetration by topical treatment. The results are summarized in
53

2132
T. Akama et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2129–2132
nucleotidase and further incubation at 37 °C for 10 min. Unhydrolyzed cAMP
was bound to AG1-X2 resin, and remaining [3H]Adenosine in the aqueous
phase was quantitated by scintillation counting. Test articles were tested at 10,
(AN2728) was chosen as the developmental candidate, and is cur-
rently in phase 2 clinical trials for the topical treatment of psoriasis
and being pursued for the topical treatment of atopic dermatitis.
3, 1, 0.3, 0.1, 0.03, 0.01, 0.003, and 0.001 lM for IC50 determination.
14. Cytokine assay: Frozen human peripheral blood mononucleocytes (PBMC)
were thawed and centrifuged. Cryopreservation media was aspirated off of the
cell pellet, and the cells were resuspended in fresh culture media (CM)
comprising RPMI 1640 and 10% FBS in 96 well plates. The test article was
References and notes
1. Lowes, M. A.; Bowcock, A. M.; Krueger, J. G. Nature 2007, 445, 866.
2. Shear, N. H. Drug Saf. 2006, 29, 49.
3. Saraceno, R.; DiStefani, A.; Giunta, A.; Chimenti, S. Recent Patents Anti-Infect.
Drug Disc. 2006, 1, 353.
dissolved in DMSO to form
samples were diluted to 100
1, 0.1, and 0.01 M final concentration (n = 3). Inducer (1
or 20 g/mL PHA for IFN , IL-2, IL-5 and IL-10) plus vehicle (1% DMSO) was
a
10 mM sample (DMSO, 100%). The 10 mM
M in CM (DMSO, 1%), then further diluted to 10,
g/mL LPS for TNF-
l
l
l
a
l
c
4. Tzu, J.; Kerdel, F. Dermatol. Ther. 2008, 21, 131.
used as a control. Vehicle without inducer was used as a negative control. Cells
were incubated at 37 °C, 5% CO2. Supernatants were removed at 24 h (for TNF-
5. (a) Bäumer, V.; Hoppman, J.; Rundfeldt, C.; Kietzmann, M. Inflamm. Allergy
2006, 6, 17; (b) Dastidat, S. G.; Rajagopal, D.; Ray, A. Curr. Opin. Invest. Drugs
2007, 8, 364.
6. Giembycz, M. A. Br. J. Pharmacol. 2008, 155, 288.
7. http://nycomed.com/en/Menu/RD/Pipeline/.
a
, IFN
supernatants were thawed, and assayed for TNF-
expression using the fluorochrome-labeled cytokine-specific beads and the
Becton Dickinson FACSArrayTM
c
, and IL-2) or 48 h (for IL-5, and IL-10), and stored at À20 °C. The
a
, IFN , IL-2, IL-5, and IL-10
c
.
8. Baker, S. J.; Akama, T.; Zhang, Y.-K.; Sauro, V.; Pandit, C.; Singh, R.; Kully, M.;
Khan, J.; Plattner, J. J.; Benkovic, S. J.; Lee, V.; Maples, K. R. Bioorg. Med. Chem.
Lett. 2006, 16, 5963.
9. Baker, S. J.; Zhang, Y.-K.; Akama, T.; Lau, A.; Zhou, H.; Hernandez, V.; Mao, W.;
Alley, M. R. K.; Sanders, V.; Plattner, J. J. J. Med. Chem. 2006, 49, 4447.
10. Fernando, R.; Mao, W.; Yaremchuk, A.; Tukalo, M.; Crépin, T.; Zhou, H.; Zhang,
Y.-K.; Hernandez, V.; Akama, T.; Baker, S. J.; Plattner, J. J.; Shapiro, L.; Martinis,
S. A.; Benkovic, S. J.; Cusack, S.; Alley, M. R. K. Science 2007, 316, 1759.
11. Li, W.; Nelson, D. P.; Jensen, M. S.; Hoerrner, R. S.; Cai, D.; Larsen, R. D.; Reider,
P. J. J. Org. Chem. 2002, 67, 5394.
12. Lukin, K.; Hsu, M. C.; Fernando, D.; Leanna, M. R. J. Org. Chem. 2006, 71, 8166.
13. PDE4 assay: PDE4 was partially purified from human U-937 myeloid leukemia
cells. The test article and/or vehicle was incubated with 0.2 mg of enzyme and
1 mM cAMP containing 0.01 mM [3H]cAMP in Tris buffer (pH 7.5) for 20 min at
25 °C. The reaction was terminated by boiling for 2 min and the resulting AMP
was converted to adenosine by addition of 10 mg/ml snake venom
15. (a) Hartmann, G.; Bidlingmaier, C.; Siegmund, B.; Albrich, S.; Schulze, J.;
Tschoep, K.; Eigler, A.; Lehr, H. A.; Endres, S. J. Pharmacol. Exp. Ther. 2000, 292,
22; (b) Claveau, D.; Chen, S. L.; O’Keefe, S.; Zaller, D. M.; Styhler, A.; Liu, S.;
Huang, Z.; Nicholson, D. W.; Mancini, J. A. J. Pharmacol. Exp. Ther. 2004, 310,
752.
16. Thompson, W. J.; Apleman, M. M. Biochemistry 1971, 311.
17. Mouse ear edema model: phorbol 12-myristate 13-acetate (PMA, 5 lg in 20 lL
of acetone) was applied topically to the anterior and posterior surface of the
right ear to groups of 5 CD-1 (Crl.) derived mice (weighing 22 2 g). Test
substances (1 mg in 20 lL) and vehicle (acetone/ethanol/1:1, 20 lL/ear) were
each applied 30 min before and 15 min after PMA challenge. Ear swelling was
then measured by a Dyer model micrometer gauge at 6 h after PMA application
as an index of inflammation. Percent inhibition was calculated according to the
formula: [(Ic À It)/Ic] Â 100%, where Ic and It refer to increase of ear thickness
(mm) in control and treated mice, respectively.