Synthesis and evaluation of dithiolethiones as novel cyclooxygenase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.11.045

3H-1,2-Dithiole-3-thiones substituted with a 3,5-di-tert-butyl-4-hydroxyphenyl (DTBHP) or a 3,5-di-tert-butyl-4-methoxyphenyl group at the C5 position were prepared and their ability to inhibit the cyclooxygenase isoenzymes, COX-1 and COX-2 was evaluated. Both compounds were potent inhibitors of COX-2 (relative to rofecoxib), and while the phenol was a weak inhibitor of COX-1, the methyl ether gave no measurable inhibition. Docking studies of the two compounds into the COX-1 and -2 active sites showed that the methyl ether could only fit in the COX-2 active site whereas the phenol could be docked into both COX-1 and -2. This study reports a new mode for inhibitor binding to COX-1 and -2 and a novel structural scaffold for the development of COX-2 selective inhibitors.

Full text of DOI:10.1016/j.bmcl.2008.11.045

Bioorganic & Medicinal Chemistry Letters 19 (2009) 459–461
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and evaluation of dithiolethiones as novel cyclooxygenase inhibitors
Shannon D. Zanatta a,b, David T. Manallack c, Bevyn Jarrott a,d, Spencer J. Williams b,
*
a Howard Florey Institute, Parkville, Vic. 3010, Australia
b School of Chemistry and Bio21 Molecular Science and Biotechnology Institute, University of Melbourne, 30 Flemington Rd, Parkville, Vic. 3010, Australia
c Medicinal Chemistry and Drug Action, Monash Institute of Pharmaceutical Sciences, Monash University, 381 Royal Pde Parkville, Vic. 3052, Australia
d Department of Pharmacology, University of Melbourne, Parkville, Vic. 3010, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
3H-1,2-Dithiole-3-thiones substituted with a 3,5-di-tert-butyl-4-hydroxyphenyl (DTBHP) or a 3,5-di-tert-
butyl-4-methoxyphenyl group at the C5 position were prepared and their ability to inhibit the cycloox-
ygenase isoenzymes, COX-1 and COX-2 was evaluated. Both compounds were potent inhibitors of COX-2
(relative to rofecoxib), and while the phenol was a weak inhibitor of COX-1, the methyl ether gave no
measurable inhibition. Docking studies of the two compounds into the COX-1 and -2 active sites showed
that the methyl ether could only fit in the COX-2 active site whereas the phenol could be docked into both
COX-1 and -2. This study reports a new mode for inhibitor binding to COX-1 and -2 and a novel structural
scaffold for the development of COX-2 selective inhibitors.
Received 16 September 2008
Revised 12 November 2008
Accepted 13 November 2008
Available online 18 November 2008
Keywords:
Cyclooxygenase
Dithiolethiones
Inhibitor
Ó 2008 Elsevier Ltd. All rights reserved.
Modelling
Inflammation is a major cause of pain and contributes to tissue
injury and dysfunction, in particular for arthritis and inflammatory
bowel syndrome.1 Inflammation has also been implicated in the
pathogenesis of neurodegenerative conditions such as Alzheimer’s
and Parkinson’s diseases.2 Non-steroidal anti-inflammatory drugs
(NSAIDs) (e.g., aspirin and diclofenac) are some of the most com-
monly used and prescribed drugs. Most NSAIDs prevent the syn-
thesis of prostaglandins through inhibiting the constitutive
enzyme cyclooxygenase-1 (COX-1) as well as cyclooxygenase-2
(COX-2), an inducible enzyme that is abundant in activated macro-
phages and other cells at sites of inflammation. Associated with
their clinical use are side effects that include gastrointestinal ulcer-
ation, bleeding and impaired renal function. Non-selective inhibi-
tion of COX enzymes has been highlighted as a potential cause of
these adverse side effects, as COX-1 is responsible for the synthesis
of endogenous prostaglandins that are believed to be important for
maintenance of gastric mucosa.3 A series of COX-2 selective inhib-
itors, the coxibs (eg celecoxib and rofecoxib) were developed and
introduced over the last 10 years. Coxibs have a reduced incidence
of gastrointestinal ulceration and bleeding; however, there have
been recent concerns over cardiovascular adverse effects of rofec-
oxib.4 These concerns have resulted in all but one coxib, celecoxib,
being voluntarily withdrawn from the market. Some studies have
suggested that rofecoxib’s adverse cardiac events may not be a
class effect but rather an intrinsic chemical property related to
its metabolism.5 For this reason novel scaffolds with COX-2
selective inhibitory activity need to be found and evaluated for
their anti-inflammatory effects.
It has been demonstrated that the 3,5-di-tert-butyl-4-hydroxy-
phenyl containing drug 1 is a potent anti-inflammatory, which acts
by dual inhibition of COX-2 and lipooxygenase-5 (Fig. 1).6 As well,
conjugates of non-selective NSAIDs with hydrogen sulfide releas-
ing dithiolethiones such as the diclofenac ester 2, displayed
enhanced anti-inflammatory activities and reduced gastrointesti-
nal toxicity.7 It was suggested that release of hydrogen sulfide
from the dithiolethione fragment might interfere with early stages
of inflammation. Hydrogen sulfide has been identified as an
important signaling molecule in the stomach and nervous system
and acts to suppress adherence of leukocytes to the vascular
MeO2S
MeO2S
tBu
OH
tBu
O
Et
O
rofecoxib
1
Cl
O
O
NH
Cl
S
S
S
2
* Corresponding author. Tel.: +61 3 8344 2422; fax: +61 3 9347 8124.
E-mail address: sjwill@unimelb.edu.au (S.J. Williams).
Figure 1. Assorted non-steroidal anti-inflammatory drugs.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.11.045

460
S. D. Zanatta et al. / Bioorg. Med. Chem. Lett. 19 (2009) 459–461
Table 1
endothelium and their subsequent migration into underlying tis-
sue.7 Accordingly, these two studies led us to ask whether a hybrid
3H-1,2-dithiole-3-thione bearing a 3,5-di-tert-butylhydroxyphenyl
(DTBHP) group at the C5 position might act as novel COX
inhibitors.
COX-1 and COX-2 inhibition of 3, 4 rofecoxib and indomethacin
Compound
IC50 (nM)
COX-1
COX-2
3
4
3640
>30,000
—
7.2
We targeted the synthesis of 3, which bears a dithiolethione
group directly joined to the 4-position of the di-tert-butylphenol,
and its methyl ether 4. Compounds 3 and 4 were prepared from
2,6-di-tert-butylphenol as shown in Scheme 1. 2,6-Di-tert-butyl-
phenol was converted to the acetophenone 5 by treatment with
acetic acid and trifluoroacetic anhydride.8 Towards 3, the phenol
of 5 was protected by a MOM (methoxymethyl) group by treat-
ment with MOMCl, prepared in situ from dimethoxymethane, acet-
yl chloride and ZnCl2 according to Berliner,9 affording the acetal 6.
Methoxycarbonylation10 of 6 using dimethyl carbonate in the pres-
ence of NaH afforded the b-ketoester 7. Treatment of 7 with phos-
phorus pentasulfide (P4S10) and elemental sulfur in the presence of
hexamethyldisiloxane according to Curphey11 smoothly afforded
the dithiolethione 312 with concomitant removal of the MOM
group. Towards 4, the phenol was converted to the methyl ether
8 by treatment with methyl iodide and potassium carbonate.
Methoxycarbonylation of 8 as above afforded the b-ketoester 9.
Treatment of 9 with P4S10 and elemental sulfur in the presence
of hexamethyldisiloxane afforded the dithiolethione 4.13
47.2
74.5
—
Rofecoxib
Indomethacin
28.2
Table 2
Glide scores for binding of 3 and 4 to COX-1 and COX-2
Compound
GlideScore
1ht5a (COX-1)
4coxa (COX-2)
3
4
À14.50
À14.48
À14.08
Unable to dock
a
Protein data bank ID.
compound 4 is unable to fit into the COX-1 binding cavity unless
the VDW constraints are relaxed.
Broadly speaking, the inability of compound 4 to dock into
COX-1 without relaxation of VDW constraints, and the ability of
3 to dock without such constraint relaxation, are in agreement
with the enzyme inhibition study.
The potencies of 3 and 4 as inhibitors of COX-1 and COX-2
enzymes were determined using microsomal COX-1 isolated from
human platelets and recombinant human COX-2 expressed in Sf21
insect cells (Table 1).14 These data reveal that both compounds are
potent inhibitors of COX-2 (relative to rofecoxib), and exhibit good
selectivity relative to COX-1. Most interestingly, modification of 3
with a methyl group gives 4, which no longer inhibits COX-1 activ-
ity up to a concentration of 30 lM.
To understand the binding of compounds 3 and 4 to COX-1 and
COX-2 we performed an initial docking study using Glide15,16 with-
in the Maestro package.17
This preliminary experiment revealed that compounds 3 and 4
could only bind into the COX-2 structure. However, if the van der
Waals (VDW) radii were reduced to below 0.8 Å during the docking
run then it was possible to fit each compound into COX-1, high-
lighting the smaller size of this binding site.
In order to more fully investigate the docking of 3 and 4 we
allowed protein flexibility within the docking protocol. Table 2
shows the results of the induced fit docking experiment giving
the best energy pose for each case. Each docking experiment
provided multiple ligand poses; however, these were essentially
identical to one another. The docking studies demonstrate that
Figure 2. Diagram showing selected residues of COX-2 adjacent to compound 3.
OH
OMOM
OMOM
tBu
tBu
tBu
tBu
tBu
tBu
d
e
OH
R
b
c
tBu
tBu
O
S
S
O
O
O
OMe
S
S
6
3
7
OMe
tBu
R= H
OMe
tBu
OMe
tBu
a
tBu
tBu
tBu
R = COMe 5
d
e
O
O
OMe
S
S
8
4
9
Scheme 1. Reagents and conditions: (a) TFAA, HOAc, 85%; (b) (i) AcCl, (MeO)2CH2, ZnCl2, toluene (ii) 5, (iPr)2EtN, 71%; (c) MeI, K2CO3, acetone, 83%; (d) Me2CO3, NaH, THF,
97%; (e) P4S10, S, (Me3Si)2O, 43%.

S. D. Zanatta et al. / Bioorg. Med. Chem. Lett. 19 (2009) 459–461
461
Figure 3. Left: Overlay of indomethacin (carbon atoms colored grey) and 3 (carbon atoms colored green) in the binding site of COX-2; Right: Overlay of flurbiprofen methyl
ester (carbon atoms colored grey) and 3 (carbon atoms colored green) in the binding site of COX-1.
3. Radi, Z. A.; Khan, N. K. Exp. Toxicol. Pathol. 2006, 58, 163.
4. Tacconelli, M. L.; Capone, M. L.; Patrignani, P. Curr. Pharm. Des. 2004, 10,
Examination of the binding modes of compounds 3 and 4 into
COX-1 and COX-2 reveals them to be almost identical. The only sig-
589.
nificant difference was seen for compound 4 where R513 was
moved allowing binding to occur. In all cases the oxygen atom
made a hydrogen bond with S530, the length of which varied from
2.82 to 3.39 Å. Figure 2 shows a diagram of the residues in close
proximity to compound 3 highlighting the hydrogen bond with
S530
An overlay of compound 3 with the bound structure of indo-
methacin reveals that the dithiolethione ring coincides with the
location of the methoxy group of indomethacin while the tert-butyl
groups are placed in a similar location to the ethanoic acid and
chlorophenyl side chains of indomethacin (Fig. 3). This figure also
shows the overlay of compound 3 with flurbiprofen methyl ester
(FME). In this case the dithiolethione ring does not overlap with
the other structure while the tert-butyl groups are oriented
towards the binding locations of the propanoate and 4-phenyl side
chains of FME.
5. Mason, R. P.; Walter, M. F.; McNulty, H. P.; Lockwood, S. F.; Byun, J.; Day, C. A.;
Jacob, R. F. J. Cardiovasc. Pharmacol. 2006, 47, S7.
6. Moreau, A.; Rao, P. N. P.; Knaus, E. E. Bioorg. Med. Chem. 2006, 14, 5340.
7. Li, L.; Rossoni, G.; Sparatore, A.; Lee, L.-C.; Del Soldato, P.; Moore, P. K. Free
Radic. Biol. Med. 2007, 42, 706.
8. Nishinaga, A.; Shimizu, T.; Yasushi, T.; Matsuura, T.; Hirotsu, K. J. Org. Chem.
1982, 47, 2278.
9. Berliner, M. A.; Belecki, K. J. Org. Chem. 2005, 70, 9618.
10. Mordant, E.; de Andrade, C.; Touati, R.; Ratovelomanana-Vidal, V.; Ben Hassine,
E. G. J.-P. Synthesis 2003, 2405.
11. Curphey, T. J. Tetrahedron Lett. 2000, 41, 9963.
12. Compound 3: The crude reaction mixture was applied directly to silica gel and
purified by flash chromatography (5% EtOAc/petrol). The residue was
recrystallised
from
EtOAc/petrol
to
afford
5-(3,5-di-tert-butyl-4-
hydroxyphenyl)-3H-1,2-dithiole-3-thione 3 as a yellow-brown solid (540 mg,
43%); mp 180–183 °C; 1H NMR (500 MHz, CDCl3) d 1.47 (s, 18H, t-Bu Â2), 5.72
(s, 1H, OH), 7.41 (s, 1H, H4), 7.48 (s, 2H, Ar); 13C NMR (125 MHz, CDCl3) d 30.3
(C(CH3)3), 34.8 (CH3), 123.5 (C4), 124.5 (C5), 134.7, 137.5, 158.1, 175.1 (Ar),
215.2 (CS); IR
m
3429, 2960, 1593, 1514, 1419, 889, 715 cmÀ1; HRMS ESI+
[M+H]+ = 339.0908, requires 339.0911 for C17H22OS3; Microanalysis: Found C,
60.37; H, 6.65. C17H22OS3, requires C, 60.31; H, 6.55%.
In conclusion we have reported the design and discovery of two
novel dithiolethiones that display potent anti-COX-2 activity.
Molecular modelling of these compounds in the active sites of
COX-1 and COX-2 provide a good explanation for their selectivity.
These compounds are interesting lead structures for the develop-
ment of hydrogen sulfide releasing anti-inflammatory drugs.
13. Compound 4: The crude reaction mixture was applied directly to silica gel and
purified by flash chromatography (5% EtOAc/petrol). The residue was
recrystallised
from
EtOAc/petrol
to
afford
5-(3,5-di-tert-butyl-4-
methoxyphenyl)-3H-1,2-dithiole-3-thione
4 as an orange solid, (117 mg,
51%); mp 106–107 °C; 1H NMR (400 MHz, CDCl3) d 1.45 (s, 18H, t-Bu Â2),
3.74 (s, 3H, OCH3), 7.41 (s, 1H, H4), 7.53 (s, 2H, Ar); 13C NMR (100 MHz, CDCl3)
d 31.8 (C(CH3)3), 36.2 (CH3), 64.6 (OCH3), 125.5 (C4), 126.2 (C5), 135.2, 145.6,
163.5, 174.2 (Ar), 215.2 (CS); IR
m 2948, 2865, 1744, 1587, 1498, 1304, 1110,
782 cmÀ1; HRMS ESI+ [M+H]+ = 353.1063, requires 353.1062 for C18H24OS3;
Microanalysis: Found C, 61.33; H, 6.85; S, 27.34. C18H23OS3, requires C, 61.32;
H, 6.86; S, 27.28%.
Acknowledgments
14. COX-1 and COX-2 assays: Aliquots either of a microsomal preparation of human
platelets or insect Sf21 cells were preincubated with dilutions of the
compounds in 1% DMSO for 15 min at 37 °C. The enzyme reaction was
We are grateful to Kamani R. Subasinghe for early synthetic
studies towards 3 and 4. We acknowledge the support of the Na-
tional Health and Medical Research Council of Australia and the
Australian Research Council.
started by the addition of arachidonic acid (100 lM for COX-1 or 0.3 lM for
COX-2) and incubated for 15 min at 37 °C. The reaction was then stopped and
the amount of prostaglandin E2 formed was quantitated using an ELISA assay.
15. Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T. A.; Klicic, J. J.; Mainz, D. T.;
Repasky, M. P.; Knoll, E. H.; Shelley, M.; Perry, J. K.; Shaw, D. E.; Francis, P.;
Shenkin, P. S. J. Med. Chem. 2004, 47, 1739.
References and notes
16. Halgren, T. A.; Murphy, R. B.; Friesner, R. A.; Beard, H. S.; Frye, L. L.; Pollard, W.
T.; Banks, J. L. J. Med. Chem. 2004, 47, 1750.
17. Schrödinger, LLC, Portland, Oregon, USA.
1. Wallace, J. Trends Pharmacol. Sci. 2007, 28, 501.
2. Wilms, H.; Zecca, L.; Rosenstiel, P.; Sievers, J.; Deuschl, G.; Lucius, R. Curr.
Pharm. Des. 2007, 13, 1925.