Mutual prodrugs containing bio-cleavable and drug releasable disulfide linkers

Article abstract of DOI:10.1016/j.bioorg.2013.06.007

We report herein the design and synthesis of several representative examples of novel mutual prodrugs containing nine distinct types of self-immolative drug-releasable disulfide linkers with urethane, ester, carbonate, or imide linkages between the linker and any two amine/amide/urea (primary or secondary) or carboxyl or hydroxyl (including phenolic)-containing drugs. We also report drug release profiles of a few representative mutual prodrugs in biological fluids such as simulated gastric fluid and human plasma. We also propose plausible mechanisms of drug release from these mutual prodrugs. We have also conducted a few mechanistic studies based on suggested sulfhydryl-assisted cleavage of mutual prodrugs and characterized a few important metabolites to give support to the proposed mechanism of drug release from the reported mutual prodrugs.

Full text of DOI:10.1016/j.bioorg.2013.06.007

Bioorganic Chemistry 49 (2013) 40–48
Contents lists available at SciVerse ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
Mutual prodrugs containing bio-cleavable and drug releasable disulfide
linkers
Arun K. Jain a, Machhindra G. Gund a, Dattatraya C. Desai b, Namdev Borhade a,
Subrayan P. Senthilkumar a, Mini Dhiman b, Naveen K. Mangu a, Sunil V. Mali a,
Nauzer P. Dubash a, Somnath Halder a, Apparao Satyam a,
⇑
a Medicinal Chemistry Division, Piramal Life Sciences, Piramal Enterprises Ltd., Nirlon Complex, Goregaon East, Mumbai 400 063, India
b Analytical Chemistry Division, Piramal Life Sciences, Piramal Enterprises Ltd., Nirlon Complex, Goregaon East, Mumbai 400 063, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 January 2013
Available online 3 July 2013
We report herein the design and synthesis of several representative examples of novel mutual prodrugs
containing nine distinct types of self-immolative drug-releasable disulfide linkers with urethane, ester,
carbonate, or imide linkages between the linker and any two amine/amide/urea (primary or secondary)
or carboxyl or hydroxyl (including phenolic)-containing drugs. We also report drug release profiles of a
few representative mutual prodrugs in biological fluids such as simulated gastric fluid and human
plasma. We also propose plausible mechanisms of drug release from these mutual prodrugs. We have
also conducted a few mechanistic studies based on suggested sulfhydryl-assisted cleavage of mutual pro-
drugs and characterized a few important metabolites to give support to the proposed mechanism of drug
release from the reported mutual prodrugs.
This article is dedicated to Professors
H. Junjappa and H. Ila for their outstanding
contribution to the chemistry of Sulfur.
Keywords:
Prodrugs
Controlled release
Synthesis
Ó 2013 Elsevier Inc. All rights reserved.
Preclinical pharmacokinetics
Stability
Mutual prodrugs
Drug releasable disulfide linkers
Self-immolative bio-cleavable linkers
Sulfhydryl-assisted cleavage
Thiol-assisted cleavage
1. Introduction
drugs in a combination therapy to yield a molecular entity called
‘mutual prodrug’ or ‘codrug’ [1–4]. Moreover, there could be
Combination therapy or polytherapy involving two or more
therapeutic agents has been in use initially for the treatment of dis-
eases such as AIDS/HIV, cancer, leprosy, malaria and tuberculosis
and its use is now extended further to the treatment of many other
disorders such as cardiovascular, cardiometabolic, gastrointestinal
and musculoskeletal. The rationale for using such combination pill
or polypill is to harness the potential synergistic or additive effects
of constituent drugs leading to better treatment outcomes. How-
ever, it could be a difficult task sometimes to formulate a combina-
tion pill or polypill containing two or more active pharmaceutical
ingredients with different physicochemical properties and such a
hurdle could be overcome by covalently coupling the constituent
potential advantages such as improved delivery and pharmacoki-
netic properties in delivering the co-administered agents as a sin-
gle chemical entity.
By definition, a mutual prodrug consists of two pharmacologi-
cally active agents coupled together so that each acts as a promo-
iety for the other agent and vice versa [1]. Constituent drugs in a
mutual prodrug can be covalently coupled either directly or via a
suitable linker. Two classic examples of approved mutual prodrugs
are estramustine phosphate (1a) [5] and sultamicillin (1b) [6]
(Fig. S1, Supporting Information). Estramustine is used in the treat-
ment of prostate cancer and it is composed of an antiandrogenic
17a-estradiol and an alkylating agent normustard [HN(NCH2CH2-
Cl)2] linked via a carbamate linkage. The phosphate group on the
estradiol moiety was introduced to improve its water solubility.
Both the carbamate and phosphate groups in 1a are hydrolyzed
in vivo to release the constituent drugs. Sultamicillin is a mutual
prodrug with synergistic action and is composed of an irreversible
⇑
Corresponding author. Address: Special Projects Department, Medicinal Chem-
istry Division, Piramal Life Sciences, Piramal Enterprises, Ltd., 1 Nirlon Complex,
Goregaon East, Mumbai 400 063, India. Fax: +91 22 3081 8036.
E-mail address: apparaosvgk@hotmail.com (A. Satyam).
0045-2068/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.bioorg.2013.06.007

A.K. Jain et al. / Bioorganic Chemistry 49 (2013) 40–48
41
-lactamase inhibitor sulbactam and the antibiotic ampicillin con-
nected via a double ester linkage which can be hydrolyzed
in vivo to release both the drugs simultaneously.
Ideally, a mutual prodrug design must meet the following crite-
ria: (a) the mutual prodrug itself should be inactive or less active
than the parent drug(s); (b) both the constituent drugs must pos-
sess at least one derivatizable functional group such as an amine,
carboxyl, hydroxyl [7]; (c) the linker should be bio-cleavable
(either enzymatically or non-enzymatically); (d) the linker or the
fragments released from the linker should be nontoxic; (e) release
of parent drugs from the mutual prodrug must occur rapidly to en-
sure therapeutically effective levels of both the constituent drug(s)
at the site of action [2].
As an extension of our mission to exploit the bio-cleavable and
drug-releasable disulfide linker technology [8–11], we report here-
in the design and synthesis of several representative examples of
novel mutual prodrugs containing nine distinct types of bio-cleav-
able, self-immolative and drug-releasable disulfide linkers (Fig. 1).
We also report a few mechanistic studies to determine the drug re-
lease profiles of a few representative mutual prodrugs in biologi-
cally relevant fluids such as simulated gastric fluid and plasma
and propose plausible mechanisms of drug release from these mu-
tual prodrugs.
2. Results and discussion
In this study, in addition to exploiting our disulfide linker tech-
nology [8–11], we also aimed at identifying pairs of representative
drugs with functional group(s) that are essential to generate
appropriate bio-cleavable covalent linkages between the constitu-
ent drug(s) and the disulfide linker. Consequently, some of the mu-
tual prodrugs synthesized in this study may not be of any
therapeutic value, but they do serve as good models to subse-
quently design potentially useful mutual prodrugs of appropriate
pairs of drugs that would offer the desired additive and/or syner-
gistic effects.
2.1. Design of mutual prodrugs
The most important feature in our design of mutual prodrugs 2–
10 (Fig. 1) is the presence of an appropriately positioned disulfide
group that is at b-position to bio-labile linkages connecting constit-
uent drugs at both sides of the linker. While drug release from es-
ter-bearing mutual prodrugs could occur in vivo via enzymatic
hydrolysis, we logically propose here a plausible sulfhydryl-as-
sisted cleavage of the mutual prodrugs containing carbonate, ester,
urethane and imide linkages as shown in Scheme 1. For brevity, we
present here a generic form of mutual prodrug that depicts a drug-
releasable self-immolative disulfide linker and all four types of
linkages [i.e., carbonate, ester, imide and urethane] between the
constituent drugs and the linker. We thus logically suggest that
an intracellular sulfhydryl-containing species such as cysteine or
glutathione (RSH or its thiolate anion RSꢀ at biological pH) would
attack the disulfide bond in the mutual prodrug to give the frag-
ments [M1] and [M2]. The thiolate anion in [M1] would trigger
cyclization by attacking the proximal carbonyl group of carbon-
ate/urethane/imide linkage (route a) to release the first drug
(D1-XH) along with the cyclic metabolite ethylene monothiolcar-
bonate [M3] [12]. Alternatively, the metabolite M1 could undergo
cyclization followed by decarboxylation (route b) to give thiirane
(M4) and free drug (D1-XH). The ester-bearing intermediate
metabolite [M2] could undergo sulfhydryl-assisted cleavage (route
a) to give thiirane (M4) and free drug (D2-CO2H). Alternatively, the
metabolite M2 could undergo hydrolysis by plasma esterases
(route b) to release the second drug (D2-CO2H) and the
Fig. 1. Mutual prodrugs containing a bio-labile disulfide linker with urethane, ester,
carbonate or imide linkages between the linker and two drugs with appropriate
derivatizable amide/urea or amine (primary or secondary) or carboxyl or hydroxyl
(including phenol) functional groups.
RS-S-containing conjugate [M5] that could be excreted. If RSH in-
volved is L-cysteine, then the metabolite M5 would be a cys-
teine-S-conjugate and such cysteine S-conjugates could possibly
undergo b-lyase catalyzed breakdown into pyruvate, ammonium
ion and a reactive persulfide species [13]. The persulfide species
thus formed could rapidly react with
and a reactive cysteine-S-conjugate persulfide species which could
release elemental sulfur to regenerate the -cysteine S-conjugate.
L-cystine to give L-cysteine
L
This cycle could probably continue until all relevant species are ex-
creted or metabolized into some biologically inert species.
It is also plausible to think here that the cysteine residues pres-
ent in intracellular proteins could as well participate in such disul-
fide exchange reactions with the exogenous disulfide-containing
compounds resulting in modified proteins which could lead to
some unwarranted biological response. However, by having intra-
cellular concentrations of cysteine and glutathione reduced (GSH)

42
A.K. Jain et al. / Bioorganic Chemistry 49 (2013) 40–48
1
2
RS
Mutual Prodrug
(A general formula)
RSSR
RSH
RS
pH~7.4
Ester hydrolysis
M4
[
]
M3
[
]
bye sterases
b
Route 'a'
a
a
1
2
1
Route 'a'
Route 'b'
a
b
2
b
Released Drug1
a
Released Drug2
Route 'b
'
M2
[
]
M1
[
]
CO2
M4
[
]
M5
[
]
When RSH = L-cysteine
2
β
-Lyase reaction
+
-
4
3
2
2
Ammonium ion persulfide species
Pyruvate
L-cysteine S-conjugate
L-cystine
L-cysteine
o
2
2
L-Cysteine S-conjugate persulfide
Wherein, D1-XH = Drug1, where, X=O, NR (R = H or a bond linked to the D1 residue), CONH, NCONH; D2-CO2H =
Carboxyl-containing Drug2; RSH = Cellular Cysteine or Glutathione reduced; RSSR = Cystine or Glutathione oxidized.
Scheme 1. Plausible mechanisms of drug release by sulfhydryl-assisted and/or esterase-assisted cleavage of the reported mutual prodrugs in vivo.
in
l
M and mM range, respectively, cells maintain appropriate RE-
drugs (i.e., valproic acid or nicotinic acid or naproxen) either as free
acids in presence of DCC, DMAP and triethylamine or as acid chlo-
rides in presence of triethylamine, yielded the corresponding mu-
tual prodrugs 3a–e. Treatment of urea or carbamate-containing
drugs (i.e., carbamazepine or levetiracetam) with oxalyl chloride
under refluxing condition yielded the corresponding acyl isocya-
nates 17a–b which were then treated with appropriate intermedi-
ate alcohol 15a or 15d to yield the corresponding mutual prodrugs
4a–c.
DOX balance. It is also known that GSH is a substrate for glutathi-
one transferases (GSTs), which are considered as a major family of
detoxification enzymes. Thus, cells can maintain sufficient quanti-
ties of thiolate anions (i.e., Cys-Sꢀ or GSꢀ) when it is necessary to
neutralize any exogenous disulfide-containing species entering
the cells. This could minimize the possibility of protein cysteine
residues from participating in such undesirable disulfide exchange
reactions.
The mutual prodrugs 5a–b, 6a–b and 7a–e were synthesized as
shown in Scheme 3. Thus, treatment of the diol 11 with appropri-
ate carboxyl-containing drugs (i.e., naproxen or ketoprofen or val-
proic acid or nicotinic acid or ibuprofen or aspirin) either as free
acids in presence of DCC, DMAP and triethylamine or as acid chlo-
rides in presence of triethylamine yielded the corresponding inter-
mediate alcohols 18a–f. The intermediate alcohols 18a and 18c
were treated with appropriate carboxyl-containing drugs (i.e.,
indomethacin and nicotinic acid, respectively) as acid chlorides
in presence of triethylamine to yield the corresponding mutual
prodrugs 5a–b. The intermediates 18c–e were treated with the
appropriate acyl isocyanates 17a or 17b (Scheme 2) to yield the
corresponding mutual prodrugs 7a–c. The intermediates 18a–b
were treated with phosgene in presence of triethylamine to yield
the corresponding formyl chlorides 19a–b, which were then trea-
ted with paracetamol to yield the corresponding mutual prodrugs
6a–b. Aspirin acid chloride, which was freshly made by treating
aspirin with oxalyl chloride in presence of catalytic amount of
DMF, was treated with silver cyanate to yield the corresponding
acyl isocyanate 17c, which was then reacted with the intermediate
alcohol 18a to yield the mutual prodrug 7d. Ibuprofen acid chlo-
ride, which was freshly made by treating ibuprofen with oxalyl
chloride in presence of catalytic amount of DMF, was treated with
ammonia to yield ibuprofen amide. This amide was treated with
oxalyl chloride under reflux conditions to yield the acyl isocyanate
2.2. Synthesis of mutual prodrugs
Synthesis of mutual prodrugs 2a–d, 3a–e and 4a–c was per-
formed as depicted in Scheme 2. Firstly, the commercially available
2,20-dithiodiethanol (11) was treated with acetic anhydride in
presence of pyridine to yield the mono-acetyl derivative 12. Treat-
ment of 12 with 1,10-carbonyldiimidazole (CDI) or diphosgene
yielded the corresponding imidazolide (13a) or formyl chloride
(13b) intermediates which were then treated with appropriate
amino-containing drugs (i.e., lamotrigine or amlodipine or olanza-
pine or gabapentin ethyl ester) in presence of a suitable base such
as triethylamine or diisopropylethylamine to yield the correspond-
ing intermediates 14a–d. The intermediates 14a–d were deacety-
lated to give the corresponding intermediate alcohols 15a–d
which were then treated with a phosgene equivalent (i.e., CDI or
diphosgene or 4-nitrophenyl chloroformate) to yield the corre-
sponding imidazolide (16a) or formyl chloride (16b) or 4-nitro-
phenoxy (16c) intermediates, respectively. Treatment of
intermediates 16a–c with appropriate amino-containing drugs
(i.e., gabapentin or gabapentin ethyl ester or lisinopril diethyl dies-
ter or fluoxetine) in presence of a suitable base such as diisopropyl-
ethylamine or triethylamine or aqueous sodium bicarbonate
yielded the corresponding mutual prodrugs 2a–d. Treatment of
alcohol intermediates 15a–d with appropriate carboxyl-containing

A.K. Jain et al. / Bioorganic Chemistry 49 (2013) 40–48
43
Scheme 2. Synthesis of mutual prodrugs 2a–d, 3a–e and 4a–c. Reagents and conditions: (a) Ac2O, pyridine, 42%; (b) diphosgene, CH2Cl2 (or) CDI, THF; (c) D1NHR, DIPEA or
TEA, 24–79%; (d) 1N K2CO3, MeOH:THF (3:1), 60–83%; (e) diphosgene, CH2Cl2 (or) CDI, THF/DMF/CH2Cl2 or 4-nitrophenyl chloroformate, CH2Cl2; (f) D2NHR, DIPEA or TEA or
NaHCO3 aq., 12–93%; (g) D2-CO2H, DCC, DMAP, TEA, CH2Cl2, 20–79%; (h) D2-COCl, TEA, CH2Cl2, 35–90%; (i) (COCl)2, DCE or CH2Cl2, reflux; (j) 15a or 15d, CH2Cl2, RT, 38–50%.
Scheme 3. Synthesis of mutual prodrugs 5a–b, 6a–b and 7a–e. Reagents and conditions: (a) D1-CO2H, DCC, DMAP, TEA, THF, 35%; (b) [D1-COCl] (freshly prepared by oxalyl
chloride and DMF method), TEA, THF, 69–77%; (c) [D2-COCl] (in case of indomethacin, it was freshly prepared by oxalyl chloride and DMF method), TEA, THF, 89–95%; (d) 17
(Scheme 2), DCE or CH2Cl2, 12 h, 42–73%; (e) phosgene, TEA, CH2Cl2; (f) D2-OH (paracetamol), TEA, CH2Cl2, 67%; (g) AgNCO, benzene, reflux (85–90 °C), 2 h; (h) 18a, CH2Cl2,
RT, 16 h, 25%; (i) (i) oxalyl chloride, DMF (2 drops), CH2Cl2, 4 h; (ii) 0 °C, NH3 gas, 30 min, quantitative; (j) (COCl)2, DCE, reflux; (k) 18b, CH2Cl2, RT, 16 h, 47%.
intermediate 17d, which was then treated with the intermediate
alcohol 18b to afford the mutual prodrug 7e.
Synthesis of mutual prodrugs 8a–b and 9a–b was carried out as
shown in Scheme 4. Thus, the diol 11 was treated with acetic anhy-
dride in presence of pyridine or chloroacetyl chloride in presence
of triethylamine to yield the corresponding mono-acetyl or
mono-chloroacetyl derivatives, which were then treated with
diphosgene in presence of triethylamine to yield the corresponding
formyl chlorides 20a and 20b. Treatment of 20a and 20b with hy-
droxyl-containing drugs tramadol and paracetamol, respectively,

44
A.K. Jain et al. / Bioorganic Chemistry 49 (2013) 40–48
Scheme 4. Synthesis of mutual prodrugs 8a–b and 9a–b. Reagents and conditions: (a) (i) acetic anhydride, pyridine, 42%; (ii) diphosgene, TEA, CH2Cl2; (b) (i) chloroacetyl
chloride, TEA, CH2Cl2, 61%; (ii) diphosgene, TEA, CH2Cl2; (c) (i) [Trt. DMAP]+ Clꢀ, CH2Cl2, 30%; (ii) DSC, ACN; (d) D1-OH, DMAP, ACN, 63–80%; (e) CDI, DMF, 11 (wherein, D1-
OH = metronidazole), 43%; (f) for deacetylation: 0.1 N NaOH, THF, 0 °C, 75%; (g) for removal of chloroacetyl group: NH3, methanol, 0 °C, 90%; (h) 10% TFA in CH2Cl2, 67%; (i)
diphosgene, TEA, CH2Cl2; (j) CDI, DMF; (k) 4-nitrophenyl chloroformate, CH2Cl2, 53%; (l) D2-OH, DMAP, diisopropylethylamine or TEA, CH2Cl2, 44–67%; (m) D2-NHR,
diisopropylethylamine, DMF or CH2Cl2, 22–51%.
in presence of DMAP yielded the corresponding intermediates 21a
and 21b. The intermediate 21a was treated with 0.1 N sodium
hydroxide to yield the intermediate alcohol 23a. Treatment of
21b with ammonia in methanol yielded the intermediate alcohol
23b. In an alternative method, treatment of 11 with trityl-DMAP
complex yielded the mono-trityl derivative, which was then trea-
ted with N,N0-disuccinimidyl carbonate (DSC) to yield the reactive
intermediate 20c, which was further treated with zidovudine in
presence of DMAP to afford the intermediate 21c. The intermediate
21c was treated with 10% TFA in DCM to yield the intermediate
alcohol 23c. In another alternative method, metronidazole was
treated with CDI to yield the reactive intermediate 22, which
was then treated with 11 to afford the intermediate alcohol 23d.
The intermediates 23a–d were treated with diphosgene or CDI or
4-nitrophenyl chloroformate to yield the reactive intermediates
24 (wherein, X = chloride or imidazole or 4-nitrophenoxy groups),
which were then treated with appropriate hydroxyl or amino-
containing drugs to yield the corresponding mutual prodrugs.
Thus, reaction of appropriate intermediate 24 with lamivudine
and metronidazole in presence of DMAP and triethylamine or
diisopropylethylamine yielded the mutual prodrugs 8a and 8b,
respectively. Similarly, treatment of appropriate intermediate 24
with gabapentin ethyl ester and norfloxacin yielded the mutual
prodrugs 9a and 9b, respectively.
pine) with the diol 11 yielded the corresponding intermediate
alcohol 25a–d. Reaction of 25d (D1-CONH2 = OAc-licarbazepine)
with CDI yielded the corresponding reactive intermediate 26,
which was further treated with gabapentin in aqueous bicarbonate
to yield the mutual prodrug 4d. Treatment of intermediates 25a–c
with appropriate acyl isocyanate intermediates 17e–g (wherein,
D2-CONH2 = valpromide,
nicotinamide
and
oxcarbazepine,
respectively) yielded the corresponding mutual prodrugs 10a–c.
2.3. Metabolic stability of mutual prodrugs in biological fluids
We have selected a few representative mutual prodrugs con-
taining all 4-types of linkages (i.e., carbonate, ester, imide and ure-
thane) and determined their metabolic stability in biologically
relevant fluids such as simulated gastric fluid (SGF) and human
plasma. Thus, we have incubated bis-urethane-containing prodrug
2d, ester and carbonate-containing prodrug 6a, ester and imide-
containing prodrugs 7d and 7e, and bis-carbonate-containing pro-
drug 8a along with metabolically unstable parent drug aspirin in
SGF at pH 1.2 and at 37 °C for stated time and determined stability
or decomposition by HPLC. The prodrugs 2d (bis-urethane-con-
taining), 6a (ester and carbonate-containing), 7e (ester and
imide-containing) and 8a (bis-carbonate-containing) remained
stable in SGF for 180 min (study period). However, aspirin-naprox-
en mutual prodrug 7d, which contains an ‘‘imide’’ linkage between
aspirin and the linker and an ‘‘ester’’ linkage between naproxen
and the linker, degraded quickly to yield salicylic acid (SA) and a
major metabolite M6 (Table 1 and Scheme 6). We have determined
the structure of the metabolite M6 as 2-((2-((acetylcarbamoyl)
oxy)ethyl)disulfanyl)ethyl 2-(6-methoxynaphthalen-2-yl)propan-
oate by LCMS. Contrarily, the other ester and imide-containing
prodrug 7e, which contains an ‘‘ester’’ linkage between ketoprofen
Synthesis of mutual prodrugs 4d and 10a–c was performed as
shown in Scheme 5. Firstly, urea/amide/carbamate-containing
drugs (i.e., carbamazepine, levetiracetam, valpromide, nicotin-
amide, oxcarbazepine, and O-acetyl licarbazepine) were treated
with oxalyl chloride under refluxing conditions to yield the corre-
sponding acyl isocyanate intermediates 17a–b and 17e–h.
Treatment of the intermediates 17a–b, 17e and 17h (D1-CONH2 =
carbamazepine or levetiracetam or valpromide or OAc-licarbaze-

A.K. Jain et al. / Bioorganic Chemistry 49 (2013) 40–48
45
and the linker and an ‘‘imide’’ linkage between ibuprofen and the
linker, showed good stability in SGF under similar conditions. Even
the free drug aspirin exhibited considerable stability in SGF by
undergoing about 15% decomposition in 180 min.
We have proposed a rational mechanism for the formation of
the metabolite M6 and SA from the prodrug 7d as shown in
Scheme 6. Thus, the prodrug 7d could undergo O- to- N-acetyl
transfer via a probable transient intermediate TI1, which could
break down into the metabolite M6 and o-oxoketene (TI2) [14].
The highly reactive o-oxoketene (M7A), which can exist in equilib-
rium with the lactone form TI3 [14,15], could rapidly undergo
hydration to give a six-membered cyclic transition state intermedi-
ate TS1, which could then achieve aromatization via rearrange-
ment to yield SA [15]. The highly strained lactone form TI3 (i.e.,
benzodioxetane] could as well undergo hydration to yield SA. We
have also confirmed the structure of the major metabolite M6 by
its synthesis as shown in Scheme 6.
Similarly, we have checked metabolic stability of bis-urethane-
containing prodrug 2d, ester and carbonate-containing prodrug 6a,
ester and imide-containing prodrug 7e, and bis-carbonate-contain-
ing prodrug 8a in human plasma at 37 °C for 60 min and their sta-
bility or decomposition was monitored by HPLC (Table 2). As
anticipated, the bis-urethane-containing prodrug 2d showed good
stability in human plasma. To our surprise, even the ester and
imide-containing prodrug 7e also exhibited good stability in hu-
man plasma during the study period. However, as expected, the
other two mutual prodrugs 2d and 8a, which carry carbonate link-
ages, underwent fast decomposition (half-life: <15 min) and re-
leased significant amounts of their constituent drugs (Table 2).
2.4. Mechanistic studies
Our proposition of a probable mechanism of drug release from
mutual prodrugs containing disulfide linker as shown in Scheme 1
was based on earlier reports by us [8–11], as well as others [12]
where the formation of metabolites ethylene monothiolcarbonate
(M3) and ethylene sulfide (i.e., thiirane, M4) was rationally pro-
posed in a similar kind of sulfhydryl-assisted cleavage of other
types of prodrugs containing disulfide linkers. However, our efforts
to isolate and identify the metabolites M3 and M4 were not suc-
cessful at that time. Now, we have done careful mechanistic stud-
ies on sulfhydryl-assisted cleavage of a few representative mutual
prodrugs and tried to isolate or identify the said metabolites. First,
we were unsuccessful in identifying the molecular masses of the
extra metabolites seen in HPLC chromatograms of decomposition
mixture at various time points by LCMS as we could not capture
their molecular ion peaks which may be due to the low molecular
masses of metabolites M3 and M4. As a practical solution, we have
synthesized the metabolite ethylene monothiolcarbonate (M3) by
a known procedure [16] and purchased the thiirane (metabolite
M4) from commercial sources. We have then conducted
sulfhydryl-assisted decomposition studies on a few representative
mutual prodrugs and observed formation of the proposed metabo-
lites M3 and M4, which were further confirmed by co-injection
with authentic samples of M3 and M4 (see Fig. S3, Supporting
Information). These results unequivocally provided some credence
to the proposed mechanism of drug release from the reported mu-
tual prodrugs.
Scheme 5. Synthesis of mutual prodrugs 4d and 10a–c. Reagents and Conditions:
(a) (COCl)2, DCE or CH2Cl2, reflux; (b) 11 (Scheme 2), CH2Cl2, RT, 25–55%; (c) 25a–c,
17e–g, DCE or CH2Cl2, RT, 55–75%; (d) 25d, CDI, THF, RT; (e) D2NHR, NaHCO3 aq.,
50%.
Table 1
Stability of aspirin, 2d, 6a, 7d, 7e and 8a in simulated gastric fluid (SGF).a,b
Incubation time (min)
Aspirin
7d
% Asp
% SA
% 7d
% M6c
% Nap
% SA
0
5
10
15
30
60
120
180
100
100
100
100
97
94
89
85
0
0
0
0
3
6
11
15
93
55
49
46
44
46
36
37
4
38
46
46
50
46
56
56
0
0
0
0
0
0
0
0
2
42
50
54
56
54
64
64
We have selected bis-carbonate-containing mutual prodrug 8a,
ester and carbonate-containing mutual prodrug 6b, ester and
imide-containing mutual prodrug 7e, and bis-urethane-containing
mutual prodrug 2d and subjected these compounds to decomposi-
tion in the presence of a sulfhydryl-containing species such as
dithiothreitol (DTT) [17] under biologically relevant pH (ꢁ7.4)
and estimated their drug release profiles using HPLC (Table 3).
Half-life (t1/2
)
>180 min
<10 min
a
The prodrugs 2d, 6a, 7e and 8a remained stable for 180 min (study time) in SGF
at 37 °C.
b
Values (Mol%) given are an average from triplicate experiments.
c
Structure of the metabolite M6 (tR = 8.7 min) is confirmed by LCMS as well as
synthesis (Scheme 6). SA = Salicylic acid; Asp = Aspirin; Nap = Naproxen.

46
A.K. Jain et al. / Bioorganic Chemistry 49 (2013) 40–48
Scheme 6. A plausible mechanism of cleavage of imide-containing mutual prodrug of aspirin and naproxen 7d in SGF and synthesis of metabolite M6 for confirmation of its
structure. Reagents and conditions: (a) silver cyanate, toluene, 80 °C, 2 h; (b) 18a (see Scheme 3), CH2Cl2, 50%.
slightly differ in their UV spectrum (see Fig. S4, Supporting Infor-
mation). As seen in the Spectrum D in Fig. S4, the UV spectrum
of the peak supposedly corresponding to the remaining prodrug
7e shows very slight difference (i.e., having only very slight wedge)
to that seen in Spectrum F of Fig. S4, which is corresponding to the
pure ketoprofen dimer (M7), thereby indicating the presence of
higher concentration of the metabolite M7. After 30 min of treat-
ment with DTT, the UV spectrum of the peak corresponding to
the prodrug 7e shows a steep curve (see Spectrum E in Fig. S4)
which is almost identical to that of the ketoprofen dimer (M7)
(see Spectrum F in Fig. S4). This UV data therefore confirms that
the 70% of prodrug 7e seen in Table 3 after 30 min treatment of
7e with DTT is indeed corresponding to the metabolite M7 and
the real half-life of the prodrug 7e could be <30 min. After 3 h of
incubation, HPLC chromatogram showed nearly 90% of ibuprofen
amide, about 30% of ketoprofen and about 41% of ketoprofen dimer
(M7) based on its UV spectrum. After 7 h of incubation, HPLC anal-
ysis of the mixture revealed that only 4% of M7 remained and
nearly 70% of ketoprofen and about 80% of ibuprofen amide were
released. Based on the identified metabolites, we have logically
proposed a plausible mechanism of sulfhydryl-assisted decomposi-
tion of the ester and imide-containing mutual prodrug 7e to re-
lease the ibuprofen amide, the metabolites M3 (i.e., ethylene
monothiolcarbonate), M4 (i.e., thiirane), M7 (i.e., the ketoprofen di-
mer) and finally the free drug ketoprofen as shown in Scheme 7. An
important observation in this particular study is the formation of
ibuprofen amide instead of the free drug ibuprofen from the
imide-containing prodrug 7e. In fact, this result is in line with
our earlier observation where an imide-containing NO-flurbiprofen
prodrug released flurbiprofen amide in vitro, but released only ibu-
profen in vivo when administered orally to rats [10]. We therefore
expect that the imide-containing mutual prodrug 7e would also re-
lease the free drug ibuprofen in vivo.
Table 2
Stability of mutual prodrugs 2d, 6a, 7e and 8a in human plasma.a,b
Incubation time (min)
Prodrug 6a
Prodrug 8a
% 6a
% Nap
% Par
% 8a
% Lam
% Zid
0
15
30
60
100
22
15
2
0
38
46
68
0
76
88
100
0
0
0
0
92
92
94
0
100
100
100
100
Half-life (t1/2
)
<15 min
<15 min
a
Values (Mol%) given are average of triplicate experiments.
The prodrugs 2d and 7e remained stable in human plasma for 60 min (study
b
time) at 37 °C. Nap = Naproxen; Par = Paracetamol; Lam = Lamivudine;
Zid = Zidovudine.
Thus, the bis-carbonate-containing mutual prodrug 8a decom-
posed quickly (t1/2 < 5 min) on treatment with DTT at pH ꢁ 7.4 to
release significant amounts of free drugs lamivudine (88%) and
zidovudine (60%). The ester and carbonate-containing mutual pro-
drug 6b also decomposed quickly (t1/2 < 5 min) under similar con-
ditions to release significant amount of paracetamol (70%) and only
a slight amount (ꢁ4%) of ketoprofen after 60 min. However, there
appeared a major metabolite M7 (24%), which was identified as
the ketopropfen dimer by co-injection with an authentic sample
of the ketoprofen dimer, which was synthesized as shown in
Scheme 7. In this case, as mentioned above, we were successful
in the identification of peaks corresponding to the expected metab-
olites ethylene monothiolcarbonate (M3) and ethylene sulfide (i.e.,
thiirane, M4) in the HPLC chromatogram after co-injection of
authentic samples of M3 or M4 with the decomposition mixture
after 2 h of treatment with DTT at pH ꢁ 7.4 (see Fig. S3, Supporting
Information).
Interestingly, when ester and imide-containing mutual prodrug
7e was treated with DTT at pH ꢁ 7.4, significant amount (70%) of
ibuprofen amide and only 5% of ketoprofen were released after
60 min of reaction time (Table 3). However, there remained 44%
of a peak whose UV spectrum is identical to ketoprofen, but dif-
fered in retention times. Based on UV spectrum and retention time,
we have figured out that the ketoprofen dimer (i.e., metabolite M7)
and mutual prodrug 7e shared same retention time! However, they
When the bis-urethane mutual prodrug 2d was similarly trea-
ted with DTT at pH ꢁ 7.4, only 48% of the prodrug remained after
5 min (i.e., t1/2 < 5 min) and nearly quantitative amount (ꢁ40%) of
olanzapine was released. However, there was no release of fluoxe-
tine. However, there appeared three unidentified metabolites, M8
(20%, its UV matches with olanzapine), M9 (12%, its UV matches
with fluoxetine) and M10 (20%, its UV matches with fluoxetine).

A.K. Jain et al. / Bioorganic Chemistry 49 (2013) 40–48
47
Scheme 7. A plausible mechanism of drug release via thiol-assisted cleavage of mutual prodrug 7e and synthesis of the metabolite M7 for confirmation of its structure and its
further thiol-assisted cleavage to release the free drug; DTT = dithiothreitol (reduced).
Table 3
Thiol (DTT)-assisted cleavage of mutual prodrugs 2d, 6b, 7e and 8aa.
Incu-bation time (min) Prodrug 8a
Prodrug 6bb
Prodrug 7e
Prodrug 2d
% 8a % Lam % Zido % 6b % Keto % Par % M7c % 7ed % Keto % Ibu Am % 2d % Ola % Flu % M8e % M9f % M10g
0
5
15
30
60
120
180
240
300
420
20 h
Half-life (t1/2
100
33
–
17
4
–
–
–
0
78
–
82
88
–
–
–
–
–
0
56
–
56
60
–
–
–
–
–
100
50
–
52
44
10
–
0
0
3
–
3
4
55
–
55
68
–
0
27
–
33
70
86
–
76
80
–
0
23
–
24
24
13
–
12
1
–
100
70
–
70
44
42
41
–
0
1
–
2
5
23
30
–
67
70
–
0
27
–
33
70
86
90
–
100
–
48
–
40
26
–
23
0
0
–
40
–
52
80
–
100
100
–
0
–
0
–
0
–
–
15
48
–
0
–
20
–
18
12
–
0
0
–
0
0
–
12
–
6
18
–
69
48
–
0
–
20
–
30
33
–
14
2
–
–
–
0
–
–
4
4
76
80
–
–
0
–
3
–
–
–
–
–
–
100
75
0
)
<5 min
<5 min
<60 min
<15 min
a
Values (Mol%) given are average of triplicate experiments conducted at pH ꢁ 7.4.
b
The metabolites M3 (tR = 2.1 min) and M4 (tR = 3.13 min) were identified as ethylene monothiolcarbonate and ethylene sulfide (i.e., thiirane), respectively, by co-injection
with authentic samples of these compounds (see Fig. S1 in Supporting Information).
c
The metabolite M7 is identified as ketoprofen dimer by LCMS and it was further confirmed by synthesis (Scheme 7).
The prodrug 7e and the metabolite M7 co-eluted and it is observed by PDA UV absorbance (see Fig. S2 in Supporting Information). Hence, the given Mol% values
d
correspond to both the compounds. Based on UV spectrum, it is observed that the% of 7e is higher at initial time points and the% of metabolite M7 is higher at later time points
as 7e degrades slowly to generate ibuprofen amide (Ibu Am) and the metabolite M7, which eventually undergoes thiol-mediated cleavage to release free ketoprofen (Keto)
(Scheme 7).
e
The M8 (tR = 5.7 min) is a olanzapine-containing (determined by UV) unidentified metabolite, which degraded with time to release free olanzapine (Ola).
The M9 (tR = 9.7 min) is a fluoxetine-containing (determined by UV) unidentified metabolite, which degraded with time to release free fluoxetine (Flu).
The M10 (tR = 11.3 min) is a fluoxetine-containing (determined by UV) unidentified metabolite, which degraded with time to release free fluoxetine. Nap = Naproxen;
f
g
Par = Paracetamol; Lam = Lamivudine; Zido = Zidovudine; – = not done.
Although the mutual prodrug 2d and the metabolites M8 and M10
slowly decomposed with time, the amount of the metabolite M9
increased during that time. Thus, at 4 h incubation time, the mix-
ture showed 23% of the mutual prodrug 2d, 100% of olanzapine,

48
A.K. Jain et al. / Bioorganic Chemistry 49 (2013) 40–48
15% of fluoxetine, 0% of M8, 69% of M9 and 14% of M10. By the fifth
hour, there remained 0% of 2d, 100% of olanzapine, 48% of fluoxe-
tine, 0% of M8, 48% of M9 and 2% of M10. By the twentieth hour,
the M9 also disappeared and the mixture indicated 100% of olanza-
pine, 75% of fluoxetine and only 3% of M10 (Table 3).
thanks are also due to our Analytical Department for providing
spectral and analytical data. Particularly, we thank Narendra Rout
and Rajanikanta Sahu for LCMS data. We also thank our patents
department for their timely professional help.
The above sulfhydryl-assisted decomposition data clearly indi-
cate that the drug release from the reported mutual prodrugs de-
pends not only on the type of linkage between the linker and
drug but also on the structural features of the constituent drug.
However, the above data would help us to design mutual prodrugs
for a particular medical need such as for achieving simultaneous or
controlled/sustained or immediate/faster release of constituent
drugs in vivo.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bioorg.2013.
06.007.
References
[1] G. Singh, P.D. Sharma, Indian J. Pharm. Sci. 56 (1994) 69–79.
[2] D. Bhosle, S. Bharambe, N. Gairola, S.S. Dhaneshwar, Indian J. Pharm. Sci. 68
(2006) 286–294 (and relevant references cited therein).
3. Conclusions
[3] N. Das, M. Dhanawat, B. Dash, R.C. Nagarwal, S.K. Shrivastava, Eur. J. Pharm. Sci.
41 (2010) 571–588 (and relevant references cited therein).
[4] S. Ohlan, S. Nanda, D.P. Pathak, M. Jagia, IJPSR 2 (2011) 719–729 (and relevant
references cited therein).
[5] I. Niculescu-Duvaz, A. Cambanis, E. Tarnauceanu, J. Med. Chem. 10 (1967) 172–
174.
[6] B. Baltzer, E. Binderup, W. von Daehne, W.O. Godtfredsen, K. Hansen, B.
Nielsen, H. Sorensen, S. Vangedal, J. Antibiotics 33 (1980) 1183–1192.
[7] Structures of all the qualified drugs used in this study are presented in Fig. S2,
Supporting Information.
[8] A. Satyam, Bioorg. Med. Chem. Lett. 18 (2008) 3196–3199.
[9] K.V.S. Nemmani, S.V. Mali, N. Borhade, A.R. Pathan, M. Karwa, V. Pamidiboina,
S.P. Senthilkumar, M. Gund, A.K. Jain, N.K. Mangu, N.P. Dubash, D.C. Desai, S.
Sharma, A. Satyam, Bioorg. Med. Chem. Lett. 19 (2009) 5297–5301.
[10] N. Borhade, A.R. Pathan, S. Halder, M. Karwa, M. Dhiman, V. Pamidiboina, M.
Gund, J.J. Deshattiwar, S.V. Mali, N.J. Deshmukh, S.P. Senthilkumar, P. Gaikwad,
S.G. Tipparam, J. Mudgal, M. Dutta, A.U. Burhan, G. Thakre, A. Sharma, S.
Deshpande, D.C. Desai, N.P. Dubash, A.K. Jain, S. Sharma, K.V.S. Nemmani, A.
Satyam, Chem. Pharm. Bull. 60 (2011) 465–481.
In summary, we have reported a few representative examples of
novel mutual prodrugs containing nine distinct types of self-imm-
olative drug-releasable disulfide linkers with urethane, ester, car-
bonate, or imide linkages between the linker and any two amine/
amide/urea (primary or secondary) or carboxyl or hydroxyl
(including phenolic)-containing drugs. We have also reported drug
release profiles of a few representative mutual prodrugs in biolog-
ical fluids such as simulated gastric fluid and plasma. We have also
proposed plausible mechanisms of drug release from these mutual
prodrugs. Based on suggested sulfhydryl-assisted cleavage of mu-
tual prodrugs, we have also conducted a few mechanistic studies
and characterized a few important metabolites in support of the
proposed mechanism of drug release from the reported mutual
prodrugs. We believe that the reported prodrug technology would
pave the way for successful design of mutual prodrugs for specific
medical needs such as achieving simultaneous or sustained or
immediate release of constituent drugs from the mutual prodrugs
in vivo. Our future work would revolve around application of this
prodrug technology to the design, synthesis and evaluation of mu-
tual drugs in a few important therapeutic areas where the use of
combination or multi/polypill therapy is a standard treatment
option.
[11] A. Satyam, Prodrugs containing novel bio-cleavable linkers, US Patent
7,932,294 B2, 2011 (Originally published as US2006046967, 2006);
A. Satyam, Chem Abstr 144 (2006) 273755.
[12] (Intermediacy of the metabolite ethylene monothiolcarbonate [M1] in
a
similar sulfhydryl-assisted cleavage of prodrugs containing releasable
disulfide linkers has been proposed by others also: see) L.R. Jones, E.A. Goun,
R. Shinde, J.B. Rothbard, C.H. Contag, P.A. Wender, J. Am. Chem. Soc. 128 (2006)
6526–6527.
[13] A.J.L. Cooper, J.T. Pinto, Biochem. Pharmacol. 69 (2005) 209–220.
[14] C. Wentrup, W. Heilmayer, G. Kollenz, Synthesis (1994) 1219–1248.
[15] R.C.-Y. Liu, J. Lusztyk, M.A. McAllister, T.T. Tidwell, B.D. Wagner, J. Am. Chem.
Soc. 120 (1998) 6247–6251.
[16] D.D. Reynlods, J. Am. Chem. Soc. 79 (1957) 4951–4952.
Acknowledgments
[17] In order to achieve quantitative release of the constituent drugs, larger
quantities of cysteine or GSH had to be added several times during the in vitro
drug release studies as most of the added cysteine or GSH had quickly
undergone auto oxidation to their respective disulfide species (Cystine or
GSSG) under the reaction conditions. Since it is just a mechanistic study
involving a thiolate anion (RSꢀ), we have used DTT in place of either cysteine
or GSH as it could be used in limited quantity to achieve near quantitative
release of constituent drugs.
We sincerely thank Piramal Healthcare management for their
support and encouragement. We also sincerely thank all of our
Special Projects team members (Particularly, Gajanan Thakre,
Milan Dutta, Shubhada Deshpande, Vijaya Nadar, Dnyaneshwar
Pacharne and Santosh Goud) for their technical help. Our special