A double prodrug system for colon targeting of benzenesulfonamide COX-2 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.09.071

The design, synthesis and delivery potential of a new type of benzenesulfonamide cyclo-oxygenase-2 (COX-2) inhibitor prodrug is investigated using celecoxib. The approach involves a double prodrug that is activated first by azoreductases and then by cyclization triggering drug release. We studied the intramolecular aminolysis of the acylsulfonamide. The cyclization was surprisingly rapid at physiological pH and very fast at pH 5. The prodrug is activated specifically under conditions found in the colon but highly stable in the presence of human and rodent intestinal extracts. Finally, the prototype with celecoxib was transported much more slowly in the Caco-2 transepithelial model than the parent. The design therefore shows significant promise for the site specific delivery of benzenesulfonamide COX-2 inhibitors to the colon.

Full text of DOI:10.1016/j.bmcl.2011.09.071

Bioorganic & Medicinal Chemistry Letters 21 (2011) 6636–6640
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
A double prodrug system for colon targeting of benzenesulfonamide
COX-2 inhibitors
Juan F. Marquez Ruiz a, Kinga Kedziora a, Brian Keogh c, Jacqueline Maguire c, Mary Reilly c,
Henry Windle b, Dermot P. Kelleher b, John F. Gilmer a,
⇑
a School of Pharmacy and Pharmaceutical Sciences, Trinity College, Dublin 2, Ireland
b School of Medicine, Trinity College, Dublin 2, Ireland
c Opsona Therapeutics Ltd, The Trinity Centre for Health Sciences, Institute of Molecular Medicine, St. James’s Hospital, Dublin 8, Ireland
a r t i c l e i n f o
a b s t r a c t
Article history:
The design, synthesis and delivery potential of a new type of benzenesulfonamide cyclo-oxygenase-2
(COX-2) inhibitor prodrug is investigated using celecoxib. The approach involves a double prodrug that
is activated first by azoreductases and then by cyclization triggering drug release.
We studied the intramolecular aminolysis of the acylsulfonamide. The cyclization was surprisingly
rapid at physiological pH and very fast at pH 5. The prodrug is activated specifically under conditions
found in the colon but highly stable in the presence of human and rodent intestinal extracts. Finally,
the prototype with celecoxib was transported much more slowly in the Caco-2 transepithelial model than
the parent. The design therefore shows significant promise for the site specific delivery of benzenesulfon-
amide COX-2 inhibitors to the colon.
Received 1 September 2011
Revised 17 September 2011
Accepted 19 September 2011
Available online 22 September 2011
Keywords:
COX-2 inhibitor
Colon cancer
Celecoxib
Azoreductase
Cyclization
Ó 2011 Elsevier Ltd. All rights reserved.
Colorectal cancer is the third leading cause of cancer-related
death in the western world with incidence in the range 25–35
per 100,000.1 Most colonic cancers are preceded by the develop-
ment of small benign colonic polyps which may take several years
to accumulate the mutations required for frank malignant change.
The disease is incurable in its later stages and there is an urgent
need for new therapies and chemopreventative agents.
Colorectal cancer rates are reduced in patient groups chroni-
cally using aspirin or other cyclo-oxygenase (COX) inhibitory
drugs.2 The role of COX-2 in the aetiology and progression of the
disease is well characterized and the enzyme is an important
pharmacological target.3–7 The COX-2 inhibitor celecoxib (Celeb-
rex) was approved by the FDA (1999) as add-on therapy for pa-
tients with familial adenomatous polypsosis (FAP, 400 mg BID).
Subsequent randomized clinical trials affirmed the validity of
celecoxib in controlling polyp progression. In the adenoma preven-
tion with celecoxib trial (APC), which recruited very high risk pa-
tients, high dose celecoxib (800 mg) was associated with a 45%
reduction in adenoma recurrence. The APC trial is however usually
remembered for the discovery of an increase in risk of adverse
thrombotic events in the celecoxib treatment group, which had
significance far beyond the trial. Clinical evidence supporting a
class effect soon followed and the most selective COX-2 inhibitors
were withdrawn from the market. Unopposed systemic COX-2
inhibition is now widely accepted to increase the risk of a cardio-
vascular event.8 Accordingly, COX-2 inhibitors are indicated in only
the very highest risk FAP patients.
Colorectal targeting of COX-2 inhibitors is considered a promis-
ing strategem for reducing systemic exposure while focusing the
beneficial COX-2 inhibitory effects on the relevant tissue.9–13 Ap-
proaches so far have relied on formulations for local drug release.
In this context it is notable that there are a number of clinically
used azo-prodrugs of 5-amino salicylic acid (5-ASA) for treating
inflammatory bowel disease (IBD) that act by locally releasing
the drug, reducing systemic exposure and intestinal metabolism.
Azoprodrugs are activated by azoreductases secreted by the colo-
nic micoflora.14 The approach is most suitable and directly applica-
ble to those drugs, such as 5-ASA, that bear a primary aromatic
amine.
We recently reported the design of a prodrug type that can site
specifically deliver prednisolone to the colon using azoreductase
as a vector (Fig. 1a).15 The ester prodrug first undergoes reduction
by azoreductase liberating an amino ester, followed by spontaneous
cyclization. The concept could in principle be applied to benzenesul-
fonamide COX-2 inhibitors provided that: (i) the required acylation
of the sulfonamido group could be accomplished synthetically and
(ii) intramolecular aminolysis of acylsulfonamide intermediate
could happen at realistic rates following azoreductase activation
(Fig. 1b).
⇑
Corresponding author. Tel.: +353 1 896 2795; fax: +353 1 896 2793.
E-mail address: gilmerjf@tcd.ie (J.F. Gilmer).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.09.071

J. F. Marquez Ruiz et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6636–6640
6637
This paper concerns proof of principle studies along these lines
along with stability and transport characteristics. In the design, the
5-ASA pendant group is intended to suppress systemic absorption
by increasing mass and polarity, but it could in theory impart other
interesting characteristics. There is mechanistic evidence for an
antiproliferative effect of 5-ASA16–19 that is consistent with clinical
analysis suggesting it has a colorectal cancer chemopreventative
effect.20,21
The synthesis of 2 was carried out by attachment of 5 which
was obtained as reported previously. The condensation between
the carboxylic acid of 5 and the sulfonamide of 1 was accomplished
using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC),
with an excess of 4-dimethylaminopyridine (DMAP), required be-
cause of the low reactivity of the sulfonamide (Scheme 1). We used
polymer supported EDC22 in order to facilitate workup of the reac-
tion by filtration of the polymer. To eliminate the DMAP we used
polymer supported sulfonic acid. Finally 6 was treated with TFA
at room temperature for 4 h to remove the tertbutyl ester group
to yield 2.23
likely to undergo passive diffusion to the same extent as 1 follow-
ing oral administration.
BOC-3 (Scheme 2) was required in order to study the
intramolecular lactamization of 3. The condensation between 1
and BOC-aminophenylpropionic acid was performed using similar
chemistry to before.26 Deprotection of BOC-3 was carried out in
dichlormethane at 0 °C with trifluoroacetic acid for several min,
the volatiles removed under a stream of nitrogen and the residue
dissolved immediately in the appropriate buffer solution before
starting kinetic studies. Lactamization of 3 was studied in BBS at
the lower intestinal pH range 5–8 (37 °C) and at I = 0.12, using
HPLC to measure the disappearance of the parent and the appear-
ance of 1 and dihydroquinolone (DHQ) 4. The method was vali-
dated for precision and accuracy (recovery) at 1.1–16.6 lM for
each analyte. Lactamization was unexpectedly rapid given the
hydrolytic stability of 2 and it was strikingly pH dependent (Figs.
3 and 4). Apparent first-order rate constants were obtained at each
pH by nonlinear regression which allowed an estimation of cycliza-
tion half-lives. These decreased from around 10 h at pH 7.4 to
around 10 min at pH 5.5. The kinetics of aminolysis of acylsulfona-
mides by amines has not been studied previously. Acylsulfona-
mides are weak organic acids with pKa values in the range of
carboxylic acids, for which they are sometimes employed as mim-
ics (parecoxib (Fig. 1.) has a reported pKa of 4.9). The pH-rate
behavior of cyclization is likely to be influenced by the ionization
state of the acylsulfonamide and anilide amino groups (nucleo-
phile). In its anionic form, the sulfonamide amide group is resistant
to nucleophilic attack. Cyclization rate at pH 5 is rapid because of
the presence of the neutral anilide (pKa ꢁ 4.3) and sulfonamide.
In order to examine both of the necessary mechanistic
components of drug release, that is, azo reduction and cyclization,
2 was incubated under anaerobic conditions in a brain–heart-
infusion (BHI) suspension that had been inoculated withClostridium
perfringens, an azoreductase secreting anaerobe from the human
intestine. Incubation and analysis was carried out as recently de-
scribed.27 Processing of 2 was measured using HPLC along with
the appearance of celecoxib (1) and DHQ (4). Consumption of 2
was complete in 6 h (85% in 4 h) (Fig. 5). The amino intermediate
(3) was also evident and it too disappeared over time leading to for-
mation of celecoxib (1) and DHQ 4. The experiment was conducted
at pH 7.4 at which the HPLC and extraction methods were validated
but evidently this is not the optimum pH for the cyclization of 3. The
pH in the gastrointestinal tract (radiotelemetric capsule) varies from
6.4 in proximal small intestine to 7.3 in the distal ileum, falling
sharply to 5.7 in the caecum and rising through the colon to 6.6 in
the rectum.28 This pH range is optimal for reduction and cyclization
from 2 suggesting that 2 has suitable characteristics for colon target-
ing of celecoxib.
An important parameter determining the potential capacity of 2
to deliver 1 into the colon is its stability under the conditions in the
small intestine and stomach. The stability of 2 was monitored by
HPLC following incubation at 20 lM in freshly aspirated duodenal
or gastric juices from IBD patients (37 °C) (n = 4 in duplicate). Com-
pound 2 could be recovered unchanged from the intestinal prepa-
rations after 24 h, with no evidence of formation of 1. Compound 2
was also stable in human and rat plasma (<5% hydrolysis at 24 h in
50% plasma buffered at pH 7.4, 37 °C). Unsurprisingly, 2 was also
stable (<1% hydrolysis) in borate buffer solution at pH 5 and 7.4
over 24 h (HPLC). The stability characteristics of acylsulfonamides
have been studied as potential prodrugs in the generic sense and
specifically as prodrugs of celecoxib and valdecoxib. Parecoxib
sodium, developed originally for IV administration, is a more water
soluble propionamide of valdecoxib.23 Parecoxib is converted to
valdecoxib in man but its acetyl analog is reported to be incom-
pletely converted in vivo to the parent.24 The acyl, propionyl and
butyryl amides of celecoxib were reported to be stable in rat plas-
ma over 24 h.25 The resistance of 2 towards hydrolysis is not sur-
prising in this context and indeed it is a useful feature provided
that activation at the target tissue occurs with appropriate kinetics.
The transport characteristics of 1 and 2 were assessed using the
transepithelial Caco-2 model in order to predict the potential for
passive diffusion and extraction from the gut (Fig. 2). In the apical
to basolateral direction (AP>BL, the permeability coefficient (Papp
)
for 1 was 1.88eꢀ04 cm/s whereas for 2 it was 2.94eꢀ06 cm/s. Mass
balance suggests that 98% of celecoxib (1) was transported in the
AP>BL direction whilst only 18% of 2 was transported in the AP>BL
study.
Secretory transport was more similar for 1 and 2 although cele-
coxib permeability was still higher (Papp 1.24eꢀ05 vs 4.33eꢀ06 cm/
s). Thus mass balance showed 46% of celecoxib transported versus
the 19% of 2 transported in BL>AP. The data suggests that 2 is not
A double prodrug design involving azoreductase and cyclization
can be applied to benzenesulfonamides because of the surprising
rate of intramolecular aminolysis of acylsulfonamides. The
phenylpropionic acid acyl adduct of celecoxib is highly stable in
O
O
O
O
S
O
N
H
(CH2)2
SO2NH2
(CH2)2
Prednisolone
HO
O
OH
N
N
N
CF3
N
CF3
N
N
O
N
O
N
OH
OH
2
O
OH
OH
1
5-ASA
5-ASA
Prednisolone-ASA prodrug
Figure 1a. Prednisolone prodrug, celecoxib (1), celecoxib prodrug design (2).

6638
J. F. Marquez Ruiz et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6636–6640
O
S
O
O
O
S
O
O
N
H
N
H
N
N
N
N
N
H2N
N
CF3
CF3
3
NH2
HO
colonic microflora
2
O
OH
HO
5-ASA
O
OH
O
S
O
O
O
S
O
N
H
NH2
N
O
N
N
CF3
N
H
O
celecoxib
parecoxib
4 DHQ
1
Figure 1b. Mechanism of drug release by spontaneous intramolecular lactamization of 3 following azoreductase activation of 2.
O
S
O
N
O H
1
+
N
N
N
N
a
b
CF3
OH
2
N
O
N
O
5
HO
O
6
CO2H
O
Scheme 1. Synthetic route to celecoxib prodrug 2. Reagents and conditions: (a) celecoxib (1), P-EDC, DMAP in ClCH2CH2Cl: t-butanol (1:1), A-15; (b) TFA, DCM.
0.0100
A
B
0.004
0.003
0.002
0.001
0.000
0.0075
0.0050
0.0025
0.0000
0
1000 2000 3000 4000 5000 6000 7000 8000
Time (sec)
0
1000 2000 3000 4000 5000 6000 7000 8000
Time (sec)
Figure 2. (A) Absorptive transport of 1 (h), 2 (N): (B) Secretory transport of 1 (h), 2 (N).
NHBoc
O
S
O
N
O
NHBoc
O H
N
a
N
CF3
OH
+
1
BOC-3
7
Scheme 2. Synthetic route to BOC-3. Reagents and conditions: (a) celecoxib (1), P-EDC, DMAP in ClCH2CH2Cl: t-butanol (1:1), A-15.

J. F. Marquez Ruiz et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6636–6640
6639
100
100
75
50
25
0
A
B
75
50
25
0
0
100
200
300
400
500
0
10 20 30 40 50 60 70 80
Time (min)
Time (min)
Figure 3. Lactamization of 3 (s) with formation of celecoxib, 1 (h), DHQ, 4 (ꢀ) at A: pH 6.4 and B, 5.5 both at 37 °C and I = 0.05.
propionic acid linker, transepithelial diffusion is effectively sup-
pressed in the Caco-2 model. The design may find most use applied
to COX-2 inhibitory sulfonamides that possess highest COX-2
selectivity and potency.
0.075
0.050
0.025
0.000
k
Acknowledgments
This work was supported by the Enterprise Ireland and OPSONA
Therapeutics LTD under the Innovations Partnership Program
(IP/2007/503).
5.0
5.5
6.0
6.5
pH
7.0
7.5
8.0
References and notes
Figure 4. Profile of kobs for the lactamization of 3 (d) versus pH at 37 °C, I = 0.05
(each point represents the mean of three determinations: Rate data are in Table 1).
1. Cooper, K.; Squires, H.; Carroll, C.; Papaioannou, D.; Booth, A.; Logan, R. F.;
Maguire, C.; Hind, D.; Tappenden, P. Health Technol. Assess 14, 1.
2. Iwama, T. J. Gastroenterol. 2009, 44(Suppl 19), 72.
3. Wu, A. W.; Gu, J.; Ji, J. F.; Li, Z. F.; Xu, G. W. World J. Gastroenterol. 2003, 9, 1990.
4. Wu, W. K.; Sung, J. J.; Lee, C. W.; Yu, J.; Cho, C. H. Cancer Lett. 2010, 295, 7.
5. Gustafsson, A.; Andersson, M.; Lagerstedt, K.; Lonnroth, C.; Nordgren, S.;
Lundholm, K. Int. J. Oncol. 2010, 36, 469.
6. Gardiner, P. S.; Gilmer, J. F. Mini-Rev. Med. Chem. 2003, 3, 461.
7. Loukanov, T.; Kirilov, M.; Furstenberger, G.; Muller-Decker, K. Clin. Invest. Med.
2010, 33, E22.
8. Furberg, C. D.; Psaty, B. M.; FitzGerald, G. A. Circulation 2005, 111, 249.
9. Krishnaiah, Y. S.; Satyanarayana, V.; Kumar, B. D.; Karthikeyan, R. S. J. Drug
Target 2002, 10, 247.
10. Shrivastava, P. K.; Shrivastava, S. K. Curr. Drug Delivery 2010, 7, 144.
11. Sinha, V. R.; Bhinge, J. R.; Kumria, R.; Kumar, M. Drug Delivery 2006, 13, 221.
12. Haupt, S.; Zioni, T.; Gati, I.; Kleinstern, J.; Rubinstein, A. Eur. J. Pharm. Sci. 2006,
28, 204.
Table 1
Rate data for the cyclization of the 3 in aqueous conditions (I = 0.05) in the pH range
5.5–7.4 at 37 °C and I = 0.05 (n = 3).
pH
kobs SEM (minꢀ1
)
t1/2 (min)
7.4
7
0.00109 0.00007
0.0053 0.0004
0.0115 0.0002
0.0168 0.0002
0.0234 0.0003
0.045 0.004
631.8
130.4
60.3
41.2
29.5
15.2
10.9
10.2
6.4
6.28
6.2
6.11
6.0
5.52
0.063 0.016
0.068 0.022
13. Thakral, N. K.; Ray, A. R.; Majumdar, D. K. J. Mater. Sci. Mater. Med. 2010, 21,
2691.
14. Roldo, M.; Barbu, E.; Brown, J. F.; Laight, D. W.; Smart, J. D.; Tsibouklis, J. Expert.
Opin. Drug Delivery 2007, 4, 547.
15. Ruiz, J. F.; Radics, G.; Windle, H.; Serra, H. O.; Simplicio, A. L.; Kedziora, K.;
Fallon, P. G.; Kelleher, D. P.; Gilmer, J. F. J. Med. Chem. 2009, 52, 3205.
16. Stolfi, C.; Fina, D.; Caruso, R.; Caprioli, F.; Sarra, M.; Fantini, M. C.; Rizzo, A.;
Pallone, F.; Monteleone, G. Biochem. Pharmacol. 2008, 75, 668.
17. Schwab, M.; Reynders, V.; Steinhilber, D.; Stein, J. Cancer Lett. 2009, 273, 98.
18. Fang, H. M.; Mei, Q.; Xu, J. M.; Ma, W. J. World J. Gastroenterol. 2007, 13, 2872.
19. Clapper, M. L.; Gary, M. A.; Coudry, R. A.; Litwin, S.; Chang, W. C.; Devarajan, K.;
Lubet, R. A.; Cooper, H. S. Inflamm. Bowel Dis. 2008, 14, 1341.
20. Velayos, F. S.; Terdiman, J. P.; Walsh, J. M. Am. J. Gastroenterol. 2005, 100, 1345.
21. Rubin, D. T.; Cruz-Correa, M. R.; Gasche, C.; Jass, J. R.; Lichtenstein, G. R.;
Montgomery, E. A.; Riddell, R. H.; Rutter, M. D.; Ullman, T. A.; Velayos, F. S.;
Itzkowitz, S. Inflamm. Bowel Dis. 2008, 14, 265.
15
10
5
22. Sturino, C. F.; Labelle, M. Tetrahedron Lett. 1998, 39, 5891.
23. Spectroscopic data for 2: IRvmax(KBr): 3434.84, 2923.82, 1642.30,
0
0
4
8
12
16
20
24
1095.99 cmꢀ1 1H-NMR (DMSO-d6) d: 9.55 (s, 1H, NH), 8.32 (s, 1H), 7.97 (s,
.
Time (h)
1H), 7.84 (d, 1H, J 5.8 Hz), 7.56 (d, 1H J = 5.24 Hz), 7.44 (d, 2H, J = 5.04 Hz), 7.36
(m, 2H), 6.80 (d, 1H, J = 6.04 Hz), 6.62 (d, 1H, J = 5.8 Hz), 6.11 (s, 1H), 2.64 (t, 2H
J = 5.24 Hz). 13C-NMR (DMSO) d: 169.90 (C, C@O, C-16), 150.18 (C, C-5), 140.01
(C, C-4), 130.61 (C, C-28), 130.11 (CH, C-9), 127.19, 125.97, 124.84, 122.48,
119.56, 118.10, 115.78, 115.78, 115.77 (CH, C-14), 115.37 (CH, C-6), 107.50
(CH, C-26), 38.37 (CH2, C-2), 27.17 (CH2, C-3), 21.06 (CH3, C-34). HRMS:
Expected (MꢀH+) = 678.16287, Found (MꢀH+) = 678.16284.
Figure 5. Progress curve following incubation of
2
(h) in a suspension of C.
perfringens under anaerobic conditions at pH 7.4 and 37 °C and the evolution of
celecoxib (d), DHQ (4), 3 (s). Also shown is 2 in BHI (ꢀ) under similar conditions
(n = 3).
24. Talley, J. J.; Bertenshaw, S. R.; Brown, D. L.; Carter, J. S.; Graneto, M. J.; Kellogg,
M. S.; Koboldt, C. M.; Yuan, J.; Zhang, Y. Y.; Seibert, K. J. Med. Chem. 2000, 43,
1661.
intestinal fluid except when azoreductase secreting organisms are
present under anaerobic conditions. Then, rapid drug disappear-
ance occurs, generating anilide 3, drug release from which is rapid
especially in the pH range encountered in the diseased colon. The
design can be optimized with respect to physicochemical and
transport characteristics but in its simplest form, with the
25. Mamidi, R. N.; Mullangi, R.; Kota, J.; Bhamidipati, R.; Khan, A. A.; Katneni, K.;
Datla, S.; Singh, S. K.; Rao, K. Y.; Rao, C. S.; Srinivas, N. R.; Rajagopalan, R.
Biopharm. Drug Dispos. 2002, 23, 273.
26. Spectroscopic data for BOC-3. IRvmax(KBr): 3437.21, 1564.30, 1415.85,
.
1261.62 cmꢀ1 1H-NMR (MeOH-d4) d: 9.48 (s, 1H, NH), 7.89 (d, 2H,

6640
J. F. Marquez Ruiz et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6636–6640
J = 8.52 Hz), 7.45 (d, 2H J = 9.04 Hz), 7.37 (d, 1H, J = 7.56 Hz), 7.21 (d, 3H
105.29 (CH, C-20), 79.46 (C, C-11), 39.44 (CH2, C-3), 27.34 (CH3, C-12), 26.60
(CH2, C-3), 22.34 (CH3, C-28). HRMS: Expected (MꢀNa+) = 651.1875, Found
(MꢀNa+) = 651.11865.
J = 8.04 Hz), 7.14 (d, 2H, J = 8.04 Hz), 6.96 (t, 1H, J = 7.52 Hz), 6.92 (d, 1H,
J = 6.52 Hz), 6.5 (s, 1H, NH), 2.84 (t, 2H, J = 7 Hz), 2.48 (t, 2H J = 7.04 Hz), 2.40 (s,
3H), 1.53 (s, 9H). 13C-NMR (MeOD) d: 178.64 (C, C@O, C-1), 155.13 (C@O, C-10),
145.48 (C, C-21), 143.87 (C, C-19), 141.05 (C, C-16), 139.44 (C, C-25), 135.59 (C,
C-13), 134.96 (C, C-5), 132.19 (C, C-4), 131.01 (C, C-22), 129.22 (CH, C-26, C-
24), 129.19 (CH, C-23, C-27), 128.62 (CH, C-14, C-18), 127.80 (CH, C-7, C-9),
125.98 (CH, C-8), 124.89 (C, C-29), 124.82 (CH, C-6), 124.65 (CH,C-17, C-15),
27. Ruiz, J. F.; Kedziora, K.; Windle, H.; Kelleher, D. P.; Gilmer, J. F. J. Pharm.
Pharmacol. 2011, 63, 806.
28. Fallingborg, J.; Christensen, L. A.; Ingeman-Nielsen, M.; Jacobsen, B. A.;
Abildgaard, K.; Rasmussen, H. H. Aliment. Pharmacol. Ther. 1989, 3, 605.