Lead optimization of isocytosine-derived xanthine oxidase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2012.11.057

We report our attempts at improving the oral efficacy of low-nanomolar inhibitors of xanthine oxidase from isocytosine series through chemical modifications. Our lead compound had earlier shown good in vivo efficacy when administered intraperitoneally but

Full text of DOI:10.1016/j.bmcl.2012.11.057

Bioorganic & Medicinal Chemistry Letters 23 (2013) 834–838
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Lead optimization of isocytosine-derived xanthine oxidase inhibitors
Komal Bajaj a, Sandeep Burudkar a, Pranay Shah a, Ashish Keche a, Usha Ghosh a, Prashant Tannu b,
Smriti Khanna c, Ankita Srivastava d, Nitin J. Deshmukh e, Amol Dixit e, Yogesh Ahire e, Anagha Damre b,
a
Kumar V. S. Nemmani e, Asha Kulkarni-Almeida d, Chandrika B-Rao c, , Rajiv Sharma ,
⇑
H. Sivaramakrishnan a
a Department of Medicinal Chemistry, Piramal Healthcare Limited, Goregaon (E), Mumbai 400 063, Maharashtra, India
b Department of Drug Metabolism and Pharmacokinetics, Piramal Healthcare Limited, Goregaon (E), Mumbai 400 063, Maharashtra, India
c Discovery Informatics, Piramal Healthcare Limited, Goregaon (E), Mumbai 400 063, Maharashtra, India
d Department of High Throughput Screening and Biotechnology, Piramal Healthcare Limited, Goregaon (E), Mumbai 400 063, Maharashtra, India
e Department of Pharmacology, Piramal Healthcare Limited, Goregaon (E), Mumbai 400 063, Maharashtra, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 July 2012
Revised 25 October 2012
Accepted 16 November 2012
Available online 29 November 2012
We report our attempts at improving the oral efficacy of low-nanomolar inhibitors of xanthine oxidase
from isocytosine series through chemical modifications. Our lead compound had earlier shown good
in vivo efficacy when administered intraperitoneally but not orally. Several modifications are reported
here which achieved more than twofold improvement in exposure. A compound with significant
improvement in oral efficacy was also obtained.
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Xanthine oxidase inhibitors
Isocytosine
Structure–activity relationship
Docking
Pharmacokinetics
Purine oxidation in eukaryotes is catalyzed by enzymes called
xanthine oxidoreductases (XOR) that catabolize physiological sub-
strate hypoxanthine to xanthine and then to uric acid (UA) by
hydroxylation.1 Hyperuricemia results from increased production
or reduced elimination of UA from the body, or both.2 It is associ-
ated with gout and other systemic disorders like cardiovascular
and metabolic diseases.3 Gout, a painful and potentially disabling
form of arthritis that has been recognized since ancient times,
has seen an increase in prevalence in recent years.4
Treatment regimes for gout include uric acid lowering agents,
which are recommended along with lifestyle changes like weight
loss, reduction in intake of alcohol and purine-rich foods, etc.5
Anti-gout drugs include XOR inhibitors which reduce production,
uricosurics which enhance excretion (e.g., probenecid, sulfinpyra-
zone, benzbromarone and isobromindione), and urate oxidase
which mops up excess UA (pegloticase, approved by US FDA for se-
vere treatment-refractory chronic gout). Allopurinol, a hypoxan-
thine analog that is a mechanism-based inhibitor of XOR,3 and
febuxostat, a non-purine inhibitor of XOR,6 are the approved first
and second lines of UA production inhibitors in the market. Colchi-
cine, a neutrophil motility and activity inhibitor, is often adminis-
tered along with other anti-gout therapies to control treatment-
related gout flares. New generation small molecules like RDEA594,
a uricosuric7 and BCX4208, a purine nucleoside phosphorylase
inhibitor that works upstream of XOR,8 are currently in phase II/
III clinical trials in combination with allopurinol or febuxostat.
These compounds showed inadequate efficacy or response as sin-
gle agents in clinical trials indicating the continuing importance
of XOR inhibitors in the treatment of gout.
We had earlier reported the discovery of a novel scaffold with
isocytosine moiety, compound 1, whose in vitro XOR inhibitory
activity was better than that of allopurinol (Fig. 1).9 Docking stud-
ies showed important interactions of this molecule in the active
site of XOR and provided directions for structure-based modifica-
tions leading to 2, with significant improvement in in vitro activity.
Further optimization efforts resulted in 3 with significantly higher
in vitro activity and in vivo efficacy by intraperitoneal (ip) admin-
istration (Fig. 1).10 However, this compound had oral bioavailabil-
ity of only 8–12% in rats. To obtain good oral efficacy, equivalent or
better in vitro potency and much higher oral exposure was
required.
Abbreviations: XOR, xanthine oxidoreductase; XO, xanthine oxidase; PK, phar-
macokinetics; AUC, area under curve; XDH, xanthine dehydrogenase; ip, intraperi-
toneal; po, per oral; SAR, structure–activity relationship; UA, uric acid; sUA, serum
UA.
⇑
Corresponding author. Tel.: +91 22 30818714; fax: +91 22 30818036.
E-mail address: chandrika.rao@piramal.com (C. B-Rao).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.11.057

K. Bajaj et al. / Bioorg. Med. Chem. Lett. 23 (2013) 834–838
835
scribed earlier.9,14 Many of these substitutions resulted in
compounds with low nanomolar in vitro activity, equivalent to
febuxostat.
R1
N
N
R2
When the isobutoxy group on 3 was replaced by the smaller and
less hydrophobic methoxy group (4), activity was reduced.
Branched alkoxy groups gave very potent compounds such as 5,
6 and 7, the last being the most potent with an IC50 of 4 nM. Benzyl
substitution at R (8) maintained potency at nanomolar levels. Fluo-
rination, substitutions like –CF3 and t-butyl (9–12), and cyclic sat-
urated rings (13, 14) also gave nanomolar IC50 values (Table 1).
Having obtained a plethora of potent compounds, subsequent
designs were aimed at getting better oral exposure. Our aim was
to obtain Cmax of at least 10 times the IC50 of the compound and
sufficiently high area under the curve (AUC) of plasma concentra-
tion vs. time profile to be able to show efficacy in the 5 h time per-
iod of the in vivo assay.9,10 Precipitation at concentrations over
R
R3
1
3
Compound
Compound
2
Compound
R1=OH, R2=NH2
R1=OH, R2=NH2
R1=OH, R2=NH2
R3=H, R=OCH3
In vitro IC50 = 9.4 μΜ
In vivo reduction =
84% (100 mg/kg, po)
R3=CN, R=OCH2CH(CH3)2
In vitro IC50 = 0.02 μΜ
In vivo reduction =
12% (200 mg/kg, po)
124% (100 mg/kg, ip)
R3=H, R=OCH2CH(CH3)2
In vitro IC50 = 0.6 μΜ
In vivo reduction=
24% (100 mg/kg, po)
Figure 1. Milestone compounds obtained with isocytosine scaffold. In vivo efficacy
was assessed by percentage reduction in UA levels in hyperuricemic rats treated
with compound vs vehicle.
30 lM of several compounds in this series and the presence of
Towards this end, the R position of 3 (Fig. 1), which approaches
the solvent front, was explored (Scheme 1)9,10 to obtain a range of
molecules with equivalent or enhanced potencies (Table 1). These
molecules were docked to xanthine dehydrogenase (XDH) co-crys-
tallized with piraxostat (PDB code 1VDV11) using XP docking pro-
tocol within GLIDE module12 of Schrodinger Suite.13 All of these
structures docked well, with better scores than allopurinol, and
some equivalent to febuxostat. In vitro xanthine oxidase (XO) inhi-
bition assay was carried out on a high throughput system as de-
the highly polar isocytosine group led us to suspect that both sol-
ubility and permeability could be issues. Depending on the as-
sessed cause of the problem with oral absorption, conventional
medicinal chemistry approaches include increasing polarity to im-
prove solubility and hydrophobicity to enhance permeability.
Hence, acid and ester functionalities were introduced
(Scheme 1)9,10 on some of the potent compounds like 7, 8 and 14
(Table 2). All the derivatives performed well in docking, retained
R1
R1
N
N
R2
N
Cl
N
O
Br
B
35
O
a
R2
b
R
Miyaura-Ishiyama
Boration
R
R
CN
Suzuki-Miyaura
CN
CN
Coupling
4-32
33
34
Scheme 1. Synthetic scheme for compounds in Tables 1–3: Reagents and conditions: (a) Pd(dppf)Cl2ÁCH2Cl2, dioxan, bis-pinacolato diborane, potassium acetate, 120 °C, 8 h.
(b) Pd(PPh3)4, DMF, Na2CO3, microwave, 140 °C, 1–2 h.
Table 1
Potent analogs of compound 3 with modifications at R position
Compound
ID
R
Docking scores (kcal/
mol)
In vitro IC50
M)
Compound
ID
R
Docking scores (kcal/
mol)
In vitro IC50
M)
(
l
(l
F3C
F3C
3
4
À10.98
À10.26
0.018 0.029
0.160 0.040
10
11
À10.89
À10.79
0.009 0.001
0.011 0.003
O
O
O
F
O
O
5
6
7
À10.96
À11.28
À11.06
0.010 0.001
0.01 3 0.001
0.004 0.001
12
13
14
À10.72
À10.63
À11.15
0.019 0.030
0.018 0.024
0.010 0.007
O
O
O
O
O
O
O
8
9
À11.19
À11.13
0.017 0.035
0.007 0.077
Aflopurinol
Febuxostat
À8.47
4.500 0.060
0.020 0.003
F
À12.05

836
K. Bajaj et al. / Bioorg. Med. Chem. Lett. 23 (2013) 834–838
Table 2
Analogs of compound 3 with polar groups to improve PK
Compound ID
R
Docking scores (kcal/mol)
In vitro IC50 (lM)
COOH
15
À12.65
À12.12
0.005 0.001
0.015 0.003
O
COOCH3
16
O
O
O
COOH
17
18
19
20
À11.01
À11.18
À12.37
À10.74
0.005 0.000
0.009 0.001
0.009 0.000
0.019 0.002
COOCH3
COOH
O
O
COOCH3
Allopurinol
Febuxostat
À8.47
4.200 0.200
0.020 0.006
À12.05
Table 3
PK studies at 10 mg/kg oral dose in rats of selected compounds with potent in vitro activities
PK parameters
3
3 (2 mg/kg,
7
15
16
Released
14
19
20
Released
23
25
25 (2 mg/kg,
iv)
15
19
iv)
Cmax (ng/mL)
65.72
0.23
0.19
1191.85
4.19
0.03
49.15
0.16
0.50
2168.70
6.60
14.03 2779.68
13.76
0.04
1.00
19.06
0.05
4.00
NC
NDa 126.24
308.40 1124.20 2666.17
Cmax
(
lM)
0.04
0.75
2.04
8.46
4.00
11.60
—
—
—
0.34
2.00
3.90
1.15
0.25
9.88
4.19
0.75
1.35
9.94
0.03
5.96
Tmax (h)
NCb
Elimination half-life
(h)
11.45
0.63
10.69
4.00
17.96
AUClast (ng* h/mL)
%F
202.69 471.19
8.60
—
—
—
296.53 19469.00 44.81 22699.83
—
—
—
—
99.68
—
109.35 119.41
—
—
—
—
94.00
97.30
—
—
—
—
—
795.24
415.81 1999.82 2991.35
—
—
—
—
—
33.87
90.22
—
—
—
—
30.24
95.88
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
90.52
49.58
13.37
—
—
—
88.42
21.20
—
Co (ng/mL)
1360.78
3825.58
4279.22
—
3716.38
5498.68
639.64
—
Vd (mL/kg)
CL (mL/h/kg)
Plasma stabilityc
Liver microsomal
stabilityc
96.00
91.50
106.13 0.00
90.57 0.00
—
—
Liver S9 fraction
65.60
—
—
98.86
93.58
—
86.80
101.17 0.00
—
13.49
71.30
—
stabilityc
a
b
c
ND: not detected.
NC: not calculated.
% Parent remaining after 60 min in in vitro assay.
*
indicating multiplication sign needs to be added between ng and h.
IC50 values in nanomolar range and hence were suitable for further
investigation. Plasma pharmacokinetics (PK) studies10,15 were
performed at 10 mg/kg per oral (po) dose in Sprague Dawley rats
(Table 3). Compound 7 was more potent than 3, but did not show
much gain in AUC. The acid derivative of 7 (15) showed significant
improvement in PK over both 7 and 3. The ester derivative, 16, re-
leased 15 in vivo with higher Cmax and AUC than 15 administered
directly. Compound 14 was as active as 3 but failed to achieve bet-
ter Cmax. Compound 19, the acid derivative of 14 did not achieve
better Cmax. This may be due to low permeability of 19. Its ester
derivative, 20, released 19 in sufficient quantities so that its Cmax
and AUC were much better than those of 3. Thus, the approach
of increasing polarity using acid and ester groups provided com-
pounds with better PK profile in terms of both Cmax and AUC. At
R1 position the acid derivative (21) docked well but resulted in a
drop in activity to 400 nM while the amide derivative (22) per-
formed poorly in both docking and in vitro assay (Table 4).
placed by –CH3 at R2 position (24, Table 4), the compound did
not dock in the correct pose and was inactive. Removal of –OH
and –NH2 groups, tried earlier when SAR was done for 1 and 2,
had resulted in complete loss of activity as loss of H-bonds pre-
vented the molecules from anchoring well in the active site.9,10
However, in the presence of –CN which compensated for this loss,
correct docked poses were obtained (Fig. 2) albeit with a decrease
in dock scores (23 and 25 in Table 4). As we hoped, the loss of po-
tency could be compensated by improved PK. Cmax of 23 and 25
was about five and 18 times better, respectively, than that of 3
(Fig. 3, Table 3). AUClast for 23 and 25 was, respectively, about
two and 10 times better than for 3. Thus, 25 led to greater improve-
ment in oral exposure than 23, with Cmax being about 52 times the
IC50 value.
Since the –OH group had a greater effect on PK than the –NH2
group, deshydroxyl forms of some of the other potent compounds
in Tables 1 and 2 were synthesized (26–32, Scheme 1) and tested
(Table 4). In general, reduction in in vitro potency was observed
for all the deshydroxylated compounds while Cmax and AUC
increased.
The other approach was to add hydrophobic groups or remove
polar groups to enhance permeability. Removal of –CN from 3
would result in 2 and not serve our purpose. When –NH2 was re-

K. Bajaj et al. / Bioorg. Med. Chem. Lett. 23 (2013) 834–838
837
Table 4
Analogs of compound 3 with modifications at R1, R2 and R positions
Compound
ID
R1
R2
R
Docking
scores
In vitro IC50
M)
Compound
ID
R1 R2
R
Docking
scores
In vitro IC50
M)
(
l
(l
(kcal/mol)
(kcal/mol)
F3C
21
COOH
NH2
À12.74
0.400 0.300 28
H
NH2
À8.44
0.080 0.007
F
O
O
22
23
24
CONH2 NH2
À6.23
À10.45
À7.53
3.900 0.700 29
0.060 0.010 30
H
H
H
H
NH2
NH2
NH2
NH2
À8.56
À8.38
À8.83
0.400 0.007
0.050 0.007
0.010 0.001
O
O
O
O
OH
OH
H
O
O
O
COOCH3
COOH
CH3
>30
31
25
26
H
H
NH2
NH2
À8.22
À8.63
0.080 0.003 32
À8.26
À8.47
0.010 0.000
4.200 0.200
O
O
F
0.150 0.020 Allopurinol
F3C
27
H
NH2
À8.67
0.060 0.004 Febuxostat
À12.05
0.020 0.010
O
efficacy of allopurinol and greater decrease in sUA than higher
doses of 16, 20 and 23 (Fig. 4B and C). Thus 25 shows trend for dose
dependent efficacy profile, with oral ED50 between 10 and 30 mg/
kg, which is a vast improvement on 3. 31 and 32 were tested
in vivo but no reduction was seen in sUA levels (data not shown).
This is consistent with the low efficacy seen for the hydroxyl ana-
log, 20.
Major metabolic liabilities anticipated were conjugation reac-
tions on hydroxyl group at R1, O-dealkylation at R and hydroxyl-
ation reactions. However, most compounds in this series were
found to be stable in in vitro metabolic stability studies. Some
changes introduced to improve PK, nevertheless, enhanced the lia-
bilities. Compound 20 was not detected in plasma, liver micro-
somes and liver S9 fraction while the de-esterified form, 19 was
found. Likewise 31 showed poor stability (plasma 28%, liver micro-
somes 0% and liver S9 fraction 11% remaining at 60 min) but its de-
esterified form, 32 was seen. Both 20 and 32 may additionally un-
dergo other phase I and II biotransformations. Reduced stability of
23 in both liver microsomes and S9 fraction (Table 3) suggests in-
crease in phase I and II metabolism, which may be responsible for
its poor efficacy. Removal of –OH in 25 seems to have reduced lia-
bility for phase II metabolism (better S9 fraction stability) as com-
pared to 3 and 23, but increased phase I metabolism, possibly to
generate active metabolites.
Figure 2. Docked poses of compounds 23 (green) and 25 (orange). H-bond of –CN
group with Asn768 anchors the compounds in the active site even though 23 loses
H-bonds with Glu802 and 25 with Thr880. This figure has been generated using
Pymol.16
In vivo efficacy studies9,17 were carried out for a few selected
compounds (Fig. 4). Compound 25 showed 49% as much reduction
in serum uric acid (sUA) level17 as allopurinol (Fig. 4A) at 10 mg/kg
oral dose. At 30 mg/kg po dose 25 showed almost 70% of the
Figure 3. PK of compounds (A) 23 and (B) 25.

838
K. Bajaj et al. / Bioorg. Med. Chem. Lett. 23 (2013) 834–838
Figure 4. Antihyperuricemic effect of 16, 20, 23 and 25 along with allopurinol in hyperuricemic rats. sUA levels after treatment by (A) 10 mg/kg of 25, (B) higher doses of 16,
20, 23 and 25, (C) AUC of sUA levels over 5 h time period of study. Percent values indicate decrease in sUA levels with respect to vehicle.
2. Adams, J. U. Nat. Biotechnol. 2009, 27, 309.
3. Pacher, P.; Nivorozhkin, A.; Szabó, C. Pharmacol. Rev. 2006, 58, 87.
4. Crittenden, D. B.; Pillinger, M. H. Bull. NYU Hosp. Jt. Dis. 2011, 69, 257.
It is observed (Fig. 4) that 25 is equally efficacious as allopurinol
at 3 h time point, suggesting that the compound may act more
slowly than allopurinol. This could be due to slower absorption
but more importantly, it could be due to metabolism by liver en-
zymes (Table 3). One of the possible metabolites of 25 is 3, which
is also an active compound and may be generated in sufficient
quantities after a time lag. This slow onset of activity of 25 suggests
that a study in chronic hyperuricemic rat model could lead to a
better understanding of its efficacy potential in a chronic disease
like gout.
In summary, using virtual screening and enzymatic assay, we
identified a novel isocytosine-containing scaffold for xanthine oxi-
dase inhibition with 3 compounds showing micromolar IC50s.9 By
structure-based design methods, we obtained compounds with im-
proved in vitro potency in low nanomolar range and potent activity
in hyperuricemic rat model by ip route but not by oral route.10 Fur-
ther medicinal chemistry approaches to lead optimization along
with structural considerations resulted in a potent compound
(25) with good oral activity.
5. Kim, K. Y.; Schumacher, R. H.; Hunsche, E.; Wertheimer, A. I.; Kong, S. X. Clin.
Ther. 2003, 25, 1593.
6. Okamoto, K.; Eger, B. T.; Nishino, T.; Kondo, S.; Pai, E. F.; Nishino, T. J. Biol. Chem.
1848, 2003, 278.
7. Mapa, J. B.; Pillinger, M. H. Curr. Opin. Investig. Drugs 2010, 11, 499.
8. Bantia, S.; Parker, C.; Upshaw, R.; Cunningham, A.; Kotian, P.; Kilpatrick, J. M.;
Morris, P.; Chand, P.; Babu, Y. S. Int. Immunopharmacol. 2010, 10, 784.
9. B-Rao, C.; Kulkarni-Almeida, A.; Katkar, K. V.; Khanna, S.; Ghosh, U.; Keche, A.;
Shah, P.; Srivastava, A.; Korde, V.; Nemmani, K. V. S.; Deshmukh, N. J.; Dixit, A.;
Brahma, M.; Bahirat, U.; Doshi, L.; Sharma, R.; Sivaramakrishnan, H. Bioorg.
Med. Chem. 2012, 20, 2930.
10. Khanna, S.; Burudkar, S.; Bajaj, K.; Shah, P.; Keche, A.; Ghosh, U.; Desai, A.;
Srivastava, A.; Kulkarni-Almeida, A.; Deshmukh, N. J.; Dixit, A.; Brahma, M. K.;
Bahirat, U.; Doshi, L.; Nemmani, K. V. S.; Tannu, P.; Damre, A.; B-Rao, C.;
Sharma, R.; Sivaramakrishnan, H. Bioorg. Med. Chem. Lett http://dx.doi.org/
10.1016/j.bmcl.2012.10.029.
11. Fukunari, A.; Okamoto, K.; Nishino, T.; Eger, B. T.; Pai, E. F.; Kamezawa, M.;
Yamada, I.; Kato, N. J. Pharmacol. Exp. Ther. 2004, 311, 519.
12. Friesner, R. A.; Murphy, R. B.; Repasky, M. P.; Frye, L. L.; Greenwood, J. R.;
Halgren, T. A.; Sanschagrin, P. C.; Mainz, D. T. J. Med. Chem. 2006, 49, 6177.
13. Glide, version 5.7, Schrödinger, LLC., New York, NY, 2011.
14. Detailed in vitro assay protocol was described earlier.9 Briefly, XO was added to
the test compound or vehicle control and reaction was initiated by addition of
xanthine. The UV absorbance indicating the formation of UA was measured at
30 min at ambient temperature. The IC50 values were calculated for test
compounds. Allopurinol or febuxostat was used as positive control.
15. Oral PK parameters of the compounds were assessed using a suspension [1 mg/
mL] in 0.5% CMC and Tween 80. Sprague Dawley rats were weighed and the
compounds were administered orally (10 mg/kg, n = 8). Blood samples were
drawn at 0.08, 0.17, 0.25, 0.5, 0.75, 1.0, 2.0, 4.0, 6.0, 8.0 and 24 h after dosing
following alternate bleeding protocol (four rats per time point). Concentrations
of the compounds in plasma were determined using an ESI-LC/MS/MS method
developed at Piramal Healthcare Ltd PK parameters were determined by non-
compartmental analysis using WinNonlin Professional (Version 4.1).
16. The PyMOL Molecular Graphics System, Version 1.2r2, Schrödinger, LLC.
17. Hyperuricemic rat model was generated by administering potassium oxonate
at a dose of 250 mg/kg ip to overnight fasted rats, as described earlier. Test
compounds were administered 1 h later and sUA level was measured at 0, 2, 3
and 5 h time points. Percent reduction in sUA level was computed as AUC over
the time period of observation for treated animals with respect to vehicle
controls.9 Reference compound (allopurinol) was administered at 10 mg/kg po
while other compounds were tested at different doses by po route.
Acknowledgments
We are grateful to Dr. Somesh Sharma for his continuous sup-
port to this structure-guided drug discovery project and to Profes-
sor Krishna Iyer for his useful inputs on PK studies. We would like
to thank Vaidehi Korde for in silico metabolite predictions and
property predictions.
Supplementary data
Supplementary data (supplementary table) associated with this
article can be found, in the online version, at http://dx.doi.org/
10.1016/j.bmcl.2012.11.057.
References and notes
1. Okamoto, K.; Kawaguchi, Y.; Eger, B. T.; Pai, E. F.; Nishino, T. J. Am. Chem. Soc.
2010, 132, 17080.