Bacterial reduction as means for colonic drug delivery: Can other chemical groups provide an alternative to the azo bond?

Article abstract of DOI:10.1021/jm301381a

We compared the rate of colonic bacterial reduction of disulfide and nitro bonds to that of an azo counterpart. The disulfide and nitro reduction rates are comparable to that of azo having a similar molecular structure. We further explored QSAR of bacterial reduction of different nitro compounds giving a Hammett correlation with ρ = 0.553, R² = 0.97. We conclude that disulfide and nitro compounds have an unexploited potential for use in prodrugs and drug delivery systems targeted to the colon.

Full text of DOI:10.1021/jm301381a

Brief Article
pubs.acs.org/jmc
Bacterial Reduction as Means for Colonic Drug Delivery: Can Other
Chemical Groups Provide an Alternative to the Azo Bond?
Sigal Saphier,*,† Avital Haft,‡ and Shlomo Margel‡
†Department of Organic Chemistry, Israel Institute for Biological Research, P.O. Box 19, Ness Ziona 74100, Israel
‡Department of Chemistry, Bar-Ilan Institute of Nanotechnology and Advanced Materials, Ramat-Gan 52900, Israel
S
* Supporting Information
ABSTRACT: We compared the rate of colonic bacterial reduction of disulfide and nitro bonds to that of an azo counterpart.
The disulfide and nitro reduction rates are comparable to that of azo having a similar molecular structure. We further explored
QSAR of bacterial reduction of different nitro compounds giving a Hammett correlation with ρ = 0.553, R2 = 0.97. We conclude
that disulfide and nitro compounds have an unexploited potential for use in prodrugs and drug delivery systems targeted to the
colon.
bacteria are lacking. This is in contrast to the extensive use of
disulfide bonds for intracellular drug delivery systems.14,15 In
the area of mucosal absorption, disulfide bonds are being used
to improve mucoadhesion of carrier systems to the mucosal
membrane in the intestine.16 There is some mention of the
preparation of polymeric colon delivery systems based on
disulfide bonds.17−19 However, to the best of our knowledge, a
detailed study describing preparation and in vivo results has not
been published.
Many aromatic nitro compounds have been shown to
undergo reduction in the colon; however, these reactions were
explored in the past, mainly using drugs and other xeno-
biotics.20−22 The reduction of nitro groups by nitro reductase
have been applied for the design of prodrugs in the fields of
antibody-directed enzyme prodrug therapy (ADEPT) and gene
(or virus) directed enzyme prodrug therapy (GDEPT or
VDEPT).23 However, to date, nitro groups were not
incorporated into colonic prodrugs or drug delivery systems.
The reduction of polymeric azo and nitro dyes by suspensions
of rat cecal bacteria led to the conclusion that the bacterial
reduction may be extracellularly mediated by nonenzymic
electron donors.24 Furthermore, for azo compounds, a
relationship was found between structural features and/or
redox potentials and rate of reduction, which is consistent with
this mechanism.25,26 Such investigations of nitro compounds
are warranted and may improve the ability to predict reduction
efficiency of particular nitro xenobiotic compounds in the
colon22 and aid to design novel, nitro based prodrugs or colonic
drug delivery systems.
INTRODUCTION
■
Specific delivery of oral drugs to the colon is important in
addressing several types of local medical conditions such as
colon cancer and inflammatory bowel disease (IBD)1 and has
potential for systemic applications such as vaccine and peptide
delivery.2−4 There are a number of methods for specific
delivery to the colon such as pH, time, or pressure based
systems. However, biodegradable delivery systems are consid-
ered to be potentially the most accurate.2,5−8 Colonic bacterial
azo reduction has long been employed for the specific delivery
of drugs to the colon. The earliest example is that of the
prodrug sulfasalazine, which upon reduction releases the active
drug 5-aminosalicylic acid (5-ASA), extensively used for the
treatment of IBD. Over the past 2 decades, azo reduction has
been exploited in the design and preparation of many colonic
drug delivery systems in the form of prodrug, polymeric
coatings, and matrices.9 Polymeric colon delivery systems
designed for bacterial degradation were based on hydrolysis of
polysaccharides10,11 or azo bond reduction.12 Despite many
publications on this topic, advancement to new products
utilizing azo bond reduction is lacking. This may be because of
toxicity concerns of the azo group and sluggish reduction rates
of azo polymers and matrices.12 Despite the fact that many
xenobiotics with a large variety of chemical groups have been
shown to be reduced in the gut, the only chemical bond utilized
to date for drug delivery by reduction is the azo bond. Colonic
delivery systems using other reducible groups are almost
nonexistent. During our work on azo polymers for colon drug
delivery,13 we were interested in expanding our knowledge on
other reducible chemical bonds and information about
structural requirements necessary for obtaining rapid and
efficient reduction. Although the bacterial reduction of many
xenobiotics has been evaluated in the past, to the best of our
knowledge, comparison of the reduction rate of different
chemical groups having similar structures by colonic bacteria
has not been published yet.
In this work, we compared the reduction of structurally
related molecules differing in one chemical bond, azo, nitro, or
disulfide, all of which have a potential use in colonic delivery
systems. Reduction rates were compared to that of the
commercial prodrug sulfasalazine, known to be efficiently
reduced in the colon. Subsequently, we focused on the
reduction of different nitro compounds and studied the
While it is clear that disulfides have a reduction potential
suitable for cleavage in the colon, studies showing directly the
reduction of disulfide containing compounds by colonic
Received: September 24, 2012
Published: November 19, 2012
© 2012 American Chemical Society
10781
dx.doi.org/10.1021/jm301381a | J. Med. Chem. 2012, 55, 10781−10785

Journal of Medicinal Chemistry
Brief Article
structure−activity relationship (SAR) of this reaction. These
studies were aimed to expand the “toolbox” of the medicinal
chemist with additional chemical groups to be used in prodrugs
or biodegradable matrices and coatings targeted to the colon
and to enable a more knowledgeable design of such future
systems.
RESULTS AND DISCUSSION
■
A common structure (5-ASA), an anti-inflammatory drug used
for the treatment of IBD and that has been incorporated in
several azo prodrugs including sulfasalazine,25 was used. Since
sulfasalazine is used clinically and known to be reduced almost
entirely within the gut with only a few percent of parent
molecule retrieved from the feces,26 its rate of reduction is
useful as a gold standard in comparing reduction rates for
colonic delivery purposes.
○
Figure 1. Reduction of different chemical bonds: ( ) sulfasalazine,
(
■
▼
) olsalazine (1), ( ) nitro (2b), ( ) disulfide (3).
▲
reduction of disulfide 3 could not be followed to completion.
At ∼2 h, the product thiol began to disappear from the
suspension. This is due to its instability probably owing to the
electron donating groups that facilitate oxidation of the thiol.
This phenomenon was also evident when the thiol was directly
added to cecal suspensions for calibration purposes. Never-
theless, for practical purposes, in a delivery system based on
polymeric matrix or film, drug release following initial reduction
and pore formation may still occur irreversibly.
After establishing the reduction process for the three
chemical groups, we looked more closely at the initial rates
of the different reductions. As can be expected from the single
bond of disulfide 3, it had the fastest initial rate of reduction.
The nitro and azo groups were reduced at a similar initial rate
(Figure 2). Since our ex vivo reducing system is based on rat
The three compounds tested were (1) olsalazine, the
symmetrical azo compound formed from two units of 5-ASA,
which is also a clinically used prodrug; (2b) 5-nitrosalicylic acid,
which may also be regarded as a prodrug of 5-ASA (as a
prodrug, this compound may be superior to some other azo
based prodrugs, as it does not release a residual moiety upon
reduction); (3) a symmetrical disulfide derivative, having an
equivalent structure to that of olsalazine, substituting the azo
bond with a disulfide bond (Chart 1).
Chart 1. Azo, Nitro, and Disulfide Compounds Used
Reduction was evaluated ex vivo in a suspension of rat cecal
content, which is a common model for the evaluation of colonic
reduction and has been found to reduce many molecules to a
similar extent as the human colon.27 Indeed, all three molecules
were found to be reduced by the cecal suspension to obtain the
expected amine or thiol products (Scheme 1).
Scheme 1. Reduction of the Different Chemical Groups
Figure 2. Relative initial rates of reduction of azo (1), nitro (2b), and
disulfide (3).
cecal content, different experimental batches may differ in
absolute rate. For this reason, the rate of sulfasalazine reduction
was remeasured in every experiment as a standard and all other
rates were calculated compared to it (Ksul).
This result emphasizes the potential of disulfides to be used
for colonic delivery. Disulfides could be incorporated into
polymers, as has been done previously for azo compounds, and
the degradation rate of the disulfide polymers may be expected
to proceed more rapidly. Thiol formation at the colonic site
may also lead to enhanced mucoadhesion.29
Overall, the reduction rates of the disulfide and nitro
compounds are of the same magnitude as that of the azo having
a similar structure and as that of sulfasalazine with k/ksul = 0.3−
2.4 for the three chemical groups.
To study further the parameters necessary for obtaining a
higher reduction rate of nitro compounds, we conducted a SAR
study using eight nitro compounds (Chart 2 and Table 1). The
carboxylic acid group present in all nitro compounds served to
gain water solubility, additionally maintaining structural
The reduction was investigated by following the accumu-
lation of the final reduction product (amine or thiol) and not
the disappearance of the starting compound. This was to ensure
that we were studying the full cleavage of the chemical bond.
Following only the starting material disappearance (commonly
performed for colored azo compounds) may lead to over-
estimation of reduction rate, resulting in the design of polymers
or prodrugs that are not completely reduced, leading to sluggish
or incomplete drug release.28
The reduction of sulfasalazine, azo 1, and nitro 2b was
completed within a few hours (Figure 1). In contrast, the
10782
dx.doi.org/10.1021/jm301381a | J. Med. Chem. 2012, 55, 10781−10785

Journal of Medicinal Chemistry
Brief Article
contents, sluggish reaction rates may also result from binding
effects of the diacids to different components of the system.
In ref 33 a correlation was obtained using σ and not σ−
values. When examining such a correlation using our data, most
substituents have similar σ and σ−, leading to similar ρ (0.5285)
and a similar, even higher R2 (0.99), provided that the
dimethylamino substituent is excluded. For the dimethylamine
substituent, σ and σ− values differ greatly because of the
resonance formed by the nitrogen with the aromatic ring which
influences the reaction rate. Attempting to correlate all six
substituents using σ therefore leads to a very weak correlation
with R2 = 0.59.
Chart 2. Structure of Nitro Aromatic Compounds Used for
Hammett Correlation
Table 1. Data on the Different Substituents for Hammett
Correlation
compd
R
σ
σ−
K = k/ksul log(K/K0)
2a
2b
2c
2d
2e
2f
p-H
0
0
K0 = 0.759
0.450
0
The high correlation (R2 > 0.97) obtained in the present
work indicates that the reduction of nitro compounds is most
likely performed extracellularly, mediated by electron donors, as
has been suggested before for azo compounds.24,34,35 Although
the R2 obtained was very high relative to other biological
systems, the ρ was relatively low. In the literature similar values
for chemical reduction of nitro compounds can be found,33
although ρ may also reach to ∼2.31 This indicates that under
the conditions of the cecal contents, the role of the substituents
is relatively reduced.
p-OH
−0.38 −0.37
−0.14 −0.17
−0.63 −0.12
−0.228
−0.100
−0.020
0.125
−0.827
p-CH3
0.603
p-N(CH3)2
p-Cl
0.724
0.24
0.19
1.013
p-COOH
0.44
0.31 (charged) 0.113
0.77 (neutral)
2g
2h
m-COOH
m-CF3
0.35
0.46
0.02 (charged) 0.128
0.56 (neutral)
−0.772
0.41
1.222
0.207
Two of the nitro compounds containing the electron
withdrawing groups p-Cl and m-CF3 were found to undergo
reduction at similar or even somewhat faster rates than
sulfasalazine. The results emphasize the potential of nitro
groups for colon specific drug delivery and the importance of
considering SAR for the design of novel systems.
similarity to 5-ASA. It was assumed that the electronic effects of
the different substituents would be additive. Indeed, the high
correlation obtained has confirmed this assumption.
The σ− constants were used for the correlation, since the
reaction center in the reduction process is expected to have
strong resonance interaction with the substituents. k is the
pseudo-zero-order rate constant derived from the initial
reduction rates.30 By use of the Hammet equation, a high
correlation was observed for the reduction of nitro compounds,
with electron-withdrawing groups accelerating the process: ρ−
= 0.5526, R2 > 0.97 (Figure 3)
CONCLUSIONS
■
The human colon serves as an important target for the
development of prodrugs and drug delivery systems for local
and systemic therapy. The local bacterial population produces a
stable reducing environment with the potential to reduce many
chemical bonds. Although quite a few delivery systems were
designed based on the familiar reduction of azo bonds, other
reducible bonds have not been explored. In addition, little
thought was given to the structure of the azo molecule in
different polymers, leading to slow reduction rates in many of
the azo delivery systems. In this work we have shown that other
groups, namely, disulfide and nitro, are also potential bonds for
the design of novel prodrugs and drug delivery systems to the
colon. The SAR of the reduction of nitro aromatic compounds
has demonstrated that taking under consideration the
electronic effects may be useful for the medicinal chemist in
optimizing the design of new prodrugs and drug delivery
systems targeted to the colon.
Figure 3. Hammet correlation for the reduction of nitro compounds in
cecal contents.
EXPERIMENTAL PROCEDURES
■
General Methods. See Supporting Information for details.
Reduction process was followed and purity of all compounds was
determined using HPLC. HPLC was performed on a Dionex Ultimate
3000 system equipped with an autosampler, a photodiode array
detector, and Chromeleon software system. The mobile phase
consisted of potassium phosphate buffer, pH 6.8, 50 mM (+0.1%
tetrabutylammonium hydroxide), and methanol. Analysis was run at a
flow rate of 1 mL/min and column temperature of 30 °C. A C18 RP
column (Altima, 150 mm × 4.6 mm, 5 μm mean particle size, Alltech)
was used. A gradient method was developed to allow the different
compounds to be analyzed by the same method. All compounds were
determined to be ≥95% pure.
Despite their electron withdrawing nature, two phthalic acids
had very slow reduction rates and did not fit this correlation.
Carboxylic acids are known to be out of correlation especially
when the substituents are negatively charged.30,31 Under the
pH conditions of the current system, these carboxylic acids are
the only charged substituents in a series of neutral ones. Slow
rates of aromatic nitro reduction in the presence of carboxylic
acid substituents have been observed by others. For example, in
a chemical nitro reduction performed under similar pH and
anaerobic conditions using hydrazine hydrate, nitrobenzoic
acids had relatively long reaction times and lower yields and
they were not incorporated into the Hammett correlation
presented.32,33 Specifically, under the conditions of the cecal
Synthesis. Compounds 1 and 2b were commercially available.
Disulfide 3 was prepared starting from 5-ASA through the preparation
of the xantogenate derivative. For reference purposes, the reduction
10783
dx.doi.org/10.1021/jm301381a | J. Med. Chem. 2012, 55, 10781−10785

Journal of Medicinal Chemistry
Brief Article
product 5-thiosalicylic acid 5 was also prepared by chemical reduction
using Zn in acidic conditions (Scheme 2).
(9) (a) Sinha, V. R.; Kumria, R. Colonic drug delivery: prodrug
approach. Pharm. Res. 2001, 18 (5), 557. (b) Sinha, V. R.; Kumria, R.
Microbially triggered drug delivery to the colon. Eur. J. Pharm. Sci.
2003, 18, 3−18.
(10) Chourasia, M. K.; Kain, S. K. Polysaccharides for colon targeted
drug delivery. Drug Delivery 2004, 11, 129−148.
Scheme 2. Synthesis of Disulfide 3 and Its Reduction
Product
(11) Larrosa, M.; Tome-Carneiro, J.; Yanez-Gascon, M. J.; Alcantara,
D.; Selma, M. V.; Beltran, D.; Garcia-Conesa, M. T.; Urban, C.; Lucas,
R.; Tomas-Barberan, F.; Morales, J. C.; Espin, J. C. Preventive oral
treatment with resveratrol pro-prodrugs drastically reduce colon
inflammation in rodents. J. Med. Chem. 2010, 53, 7365−7376.
(12) Van-den-Mooter, G.; Maris, B.; Samyn, C.; Augustijns, P.;
Kinget, R. Use of azo polymers for colon-specific drug delivery. J.
Pharm. Sci. 1997, 86 (12), 1321−1327.
(13) Saphier, S.; Karton, Y. Novel salicylazo polymers for colon drug
delivery: dissolving polymers by means of bacterial degradation. J.
Pharm. Sci. 2010, 99 (2), 804−815.
(14) West, K. R.; Otto, S. Reversible covalent chemistry in drug
delivery. Curr. Drug Discovery Technol. 2005, 2, 123−160.
(15) Saito, G.; Swanson, J. A.; Lee, K.-D. Drug delivery strategy
utilizing conjugation via reversible disulfide linkages: role and site of
cellular reducing activities. Adv. Drug Delivery Rev. 2003, 55, 199−215.
(16) Sarti, F.; Bernkop-Schnurch, A. Chitosan and thiolated chitosan.
Adv. Polym. Sci. 2011, 243, 93−110.
(17) Wilding, I. Site-specific drug delivery in the gasrtointestinal tract.
Crit. Rev. Ther. Drug Carrier Syst. 2000, 17 (6), 557−620.
(18) Schacht, E.; Wilding, I. Process for the Preparation of Azo and/
or Disulfide Polymer Matrix Drug Delivery System for the Site Specific
Delivery of an Active Agent in the Colon. US005,407,682, Apr 18,
1995.
All nitro compounds used for the Hammett correlation and their
aniline derivatives were commercially available except for 5-amino-2-
dimethylaminobenzoic acid (4d), which was obtained from 2-
dimethylamino-5-nitrobenzoic acid (2d) by hydrogenation on Pd/C.
ASSOCIATED CONTENT
* Supporting Information
■
S
General experimental information, syntheses, spectral charac-
terization data, bacterial reduction tests, and HPLC analysis
results. This material is available free of charge via the Internet
at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Phone: (972) 8-9381740. Fax: (972) 8-9381548. E-mail:
sigals@iibr.gov.il.
■
(19) Kudo, Y.; Ueshima, H.; Sakai, K. System for Release in Lower
Digestive Tract. WO/2000/074720, 2000.
(20) Pelkonen, K.; Hanninen, O. Interactions of xenobiotics with the
gastrointestinal flora. Gastrointest. Toxicol. 1986, 193−212 (Chapter
7).
(21) Scheline, R. R. Metabolism of foreign compounds by
gastrointestinal microorganisms. Pharm. Rev. 1973, 25, 451−523.
(22) Pieper, I. A.; Bertau, M. Predictive tools for the evaluation of
microbial effects on drugs during gastrointestinal passage. Expert. Opin.
Drug Metab. Toxicol. 2010, 6 (6), 747−760.
(23) Asche, C.; Dumy, P.; Carrez, D.; Croisy, A.; Demeunynck, M.
Nitrobenzylcarbamate prodrugs of cytotoxic acridines for potential use
with nitroreductase gene-directed enzyme prodrug therapy. Bioorg.
Med. Chem. Lett. 2006, 16, 1990−1994.
(24) Brown, J. P. Reduction of polymeric azo and nitro dyes by
intestinal bacteria. Appl. Environ. Microbiol. 1981, 41 (5), 1283−1286.
(25) Dubin, P.; Wright, L. Reduction of azo food dyes in cultures of
Proteus vulgaris. Xenobiotica 1975, 5 (9), 563−571.
(26) Walker, R.; Ryan, A. J. Some molecular parameters influencing
rate of reduction of azo compounds by intestinal microflora.
Xenobiotica 1971, 1 (4/5), 483−486.
(27) Brown, M. A.; DeVito, S. C. Predicting azo dye toxicity. Crit.
Rev. Environ. Sci. Technol. 1993, 23 (3), 249−324.
(28) Ueda, T.; Yamakoa, T.; Miyamoto, M.; Kimura, Y.; Sasatani, H.;
Kim, S.-I. Bacterial reduction of azo compounds as a model reaction
for the degradation of azo-containing polyurethane by the action of
intestinal flora. Bull. Chem. Soc. Jpn. 1996, 69, 1139−1142.
(29) Bernkop-Schnurch, A. Mucoadhesive systems in oral drug
delivery. Drug Discovery Today: Technol. 2005, 2 (1), 83−86.
(30) (a) Hansch, C.; Leo, A.; Taft, A. W. A survey of Hammett
substituent constants and resonance and field parameters. Chem. Rev.
1991, 91, 165−195. (b) Hansch, C.; Leo, A. In Exploring QSAR,
Fundamentals and Applications in Chemistry and Biology; American
Chemical Society: Washington, DC, 1995.
Notes
The authors declare no competing financial interest.
ABBREVIATIONS USED
■
5-ASA, 5-aminosalicylic acid; ADEPT, antibody-directed
enzyme prodrug therapy; GDEPT, gene directed enzyme
prodrug therapy; VDEPT, virus directed enzyme prodrug
therapy
REFERENCES
■
(1) Friend, D. R. New oral delivery systems for treatment of
inflammatory bowl disease. Adv. Drug Delivery Rev. 2005, 57, 247−265.
(2) Patel, M.; Shah, T.; Amin, A. Therapeutic opportunities in colon-
specific drug-delivery systems. Crit. Rev. Ther. Drug Carrier Syst. 2007,
24 (2), 147−202.
(3) McConnell, E. L.; Fadda, H. M.; Basit, A. Gut instincts:
explorations in intestinal physiology and drug delivery. Int. J. Pharm.
2008, 364, 213−226.
(4) Rajguru, V. V.; Gaikwad, P. D.; Bankar, V. H.; Pawar, S. P. An
overview on colonic drug delivery system. Int. J. Pharm. Sci. Rev. Res.
2011, 6 (2), 197−204.
(5) Das, S.; Chaudhury, A.; Ng, K.-Y. Preparation and evaluations of
zinc-pectin-chitosan composite particles for drug delivery to the colon:
role of chitosan in modifying in vitro and in vivo drug release. Int. J.
Pharm. 2011, 406, 11−20.
(6) Watts, P. J.; Illum, L. Colonic drug delivery. Drug Dev. Ind.
Pharm. 1997, 23 (9), 893−913.
(7) McConnell, E. L.; Short, M. D.; Basit, A. W. An in vivo
comparison of intestinal pH and bacterial as physiological trigger
mechanisms for colonic tergeting in man. J. Controlled Release 2008,
130, 154−160.
(8) Patel, M.; Amin, A. Recent trends in microbially and/or
enzymatically driven colon-specific drug delivery systems. Crit. Rev.
Ther. Drug Carrier Syst. 2011, 28 (6), 489−552.
(31) Jaffe, H. H. A reexamination of the Hammett equation. Chem.
Rev. 1953, 53, 191.
(32) Paula, F. S. d.; Sales, E. M.; Vallaro, M.; Fruttero, R.; Goulart, M.
O. F. The relationship between redox potentials and substituent
10784
dx.doi.org/10.1021/jm301381a | J. Med. Chem. 2012, 55, 10781−10785

Journal of Medicinal Chemistry
Brief Article
constants in biologically active arylazoxy compounds. J. Electroanal.
Chem. 2005, 57 (1), 33−41.
(33) Lauwiner, M.; Rys, P.; Wissmann, J. Reduction of aromatic nitro
compounds with hydrazine hydrate in the presence of an iron oxide
hydroxide catalyst. I The reduction of monosubstituted nitrobenzens
with hydrazine hydrate in the presence of ferrihydrite. Appl. Catal., A
1998, 172, 141−148.
(34) Semde, R.; Pierre, D.; Geuskens, G.; Devleeschouwer, M.;
Moes, A. J. Study of some important factors involved in azo derivative
reduction by Clostridium perfringens. Int. J. Pharm. 1998, 161, 45−54.
(35) Rau, J.; Maris, B.; Kinget, R.; Samyn, C.; Mooter, G. V. d.; Stoiz,
A. Enhanced anaerobic degradation of polymeric azo compounds by
Escherichia coli in the presence of low-molecular-weight redox
mediators. J. Pharm. Pharmacol. 2002, 54, 1471−1479.
10785
dx.doi.org/10.1021/jm301381a | J. Med. Chem. 2012, 55, 10781−10785