Towards an efficient prodrug of the alkylating metabolite monomethyltriazene: Synthesis and stability of N-acylamino acid derivatives of triazenes

Article abstract of DOI:10.1016/j.ejmech.2008.06.022

A series of 3-[α-(acylamino)acyl]-1-aryl-3-methyltriazenes 6a-l, potential cytotoxic triazene prodrugs, were synthesised by coupling 1-aryl-3-methyltriazenes to N-acylamino acids. Their hydrolysis was studied in isotonic pH 7.4 phosphate buffer and in hum

Full text of DOI:10.1016/j.ejmech.2008.06.022

Available online at www.sciencedirect.com
European Journal of Medicinal Chemistry 44 (2009) 1049e1056
http://www.elsevier.com/locate/ejmech
Original article
Towards an efficient prodrug of the alkylating metabolite
monomethyltriazene: Synthesis and stability of N-acylamino acid
derivatives of triazenes
a
a
b
Maria de Jesus Perry a, , Emılia Carvalho , Eduarda Rosa , Jim Iley
*
´
a iMed.UL, CECF, Faculty of Pharmacy, University of Lisbon, Avenida Professor Gama Pinto, 1649-019 Lisbon, Portugal
b Chemistry Department, The Open University, Milton Keynes, MK7 6AA, UK
Received 10 March 2008; received in revised form 19 June 2008; accepted 20 June 2008
Available online 28 June 2008
Abstract
A series of 3-[a-(acylamino)acyl]-1-aryl-3-methyltriazenes 6ael, potential cytotoxic triazene prodrugs, were synthesised by coupling 1-aryl-
3-methyltriazenes to N-acylamino acids. Their hydrolysis was studied in isotonic pH 7.4 phosphate buffer and in human plasma, while hydro-
lysis of the derivative 6a was studied in more depth across a range of pH values. Prodrugs 6ael hydrolyse by cleavage of the triazene acyl group
to afford the corresponding monomethyltriazenes. Studies in human plasma demonstrate that acylation of the a-amino group of the amino acid
carrier is an effective means of reducing the chemical reactivity of the a-aminoacyl derivatives while retaining a rapid rate of enzymatic hydro-
lysis. These derivatives displayed log P values that suggest they should be well absorbed through biological membranes.
Ó 2008 Elsevier Masson SAS. All rights reserved.
Keywords: Triazene; Prodrug; Aminoacyl; Antitumour drugs; Kinetics; Enzymes and enzyme reactions
1. Introduction
towards hydrolysis in acidic and neutral solutions by the
strong electron-withdrawing inductive effect of the protonated
amino group [4].
For many years dacarbazine 1 has been one of the most
widely used drugs to treat malignant melanoma [5].
Prodrug formation is an important means of enhancing drug
efficiency. A major requirement of the prodrug derivative is
that it must be converted rapidly, or at a controlled rate, to
the active therapeutic agent in vivo while at the same time
be sufficiently stable in vitro such that a stable pharmaceutical
product may be developed. This is often achieved using the
double prodrug concept via molecules that make use of an en-
zymatic release mechanism followed by a spontaneous chem-
ical reaction [1]. Another important aspect of prodrug design
is the need for the carrier moiety to be non-toxic [2]. To this
end, biologically compatible compounds, particularly a-amino
acids, have been used as carriers for drugs that contain either
the hydroxyl functional group (forming esters) or a carboxyl
or amine group (forming amides) [3]. Esters of a-amino acids
are chemically very unstable, the ester group being activated
H3C
H
N
N
N
N
H3C
N
H2N
O
1
The biological action of dacarbazine and, in general, the
anticancer 1-aryl-3,3-dimethyltriazenes (2, in Fig. 1) is a con-
sequence of their capacity to alkylate DNA [6].
These compounds suffer metabolic oxidation by cyto-
chrome P450 enzymes to give hydroxymethyltriazenes 3,
which, by loss of formaldehyde, generate the cytotoxic mono-
methyltriazenes 4 (Fig. 1) [7e9]. These are known alkylating
agents, capable of methylating DNA and RNA [10,11]. How-
ever, dimethyltriazenes are poorly metabolised by humans
* Corresponding author. Tel.: þ351 217 946 400; fax: þ351 217 946 470.
E-mail address: mjprocha@ff.ul.pt (M.deJ. Perry).
0223-5234/$ - see front matter Ó 2008 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2008.06.022

1050
M.deJ. Perry et al. / European Journal of Medicinal Chemistry 44 (2009) 1049e1056
X
P450
displayed significant chemical stability in isotonic phosphate
X
buffer yet rapid hydrolysis to monomethyltriazene in plasma
[27]. To demonstrate the potential of this type of compound
as possible bioreversible prodrugs of the anticancer monome-
thyltriazene system, we here report the synthesis of a series of
3-[a-(acylamino)acyl]-1-aryl-3-methyltriazenes, 6ael, a deter-
mination of their log P values and an examination of their sta-
bility both in isotonic pH 7.4 phosphate buffer and in plasma.
H3C
N
H3C
N
N
N
N
N
CH3
CH2OH
2
3
-HCHO
X
X
plasma
O
X
H3C
N
O
H
N
N
N
H
N
N
N
R
R'
N
N
N
CH3
O
R
CH3
6
4
6
6a R = CH3
6b R = H
R' = CH3
R' = CH3
R' = CH3
X = CN
X
X = CN
H3C
N
6c R = CH2Ph
X = CN
N
N
6d R = CH(CH3)2 R' = CH3
X = CN
H
6e R = CH3
6f R = CH3
6g R = CH3
6h R = CH3
6i R = CH3
6j R = CH3
6k R = CH3
6l R = CH3
R' = CH2CH3
X = CN
R' = CH2CH2CH3
R' = Ph
X = CN
X
X = CN
DNA
+
CH3
+
N
CH3DNA
N N
+
2
X = COCH3
X = COOEt
X = CONH2
X = Br
R' = CH3
R' = CH3
R' = CH3
R' = CH3
H2N
Fig. 1. Metabolism pathways for 1-aryl-3,3-dimethyltriazenes and 3-amino-
acyl-1-aryl-3-methyltriazenes.
R' = CH3
X = CH3
[12]. Consequently, with the aim of finding suitable prodrugs
that by-pass the need for this oxidative metabolism, alongside
hydroxymethyltriazenes 3 themselves [13], a range of deriva-
tives of 3 and 4, including alkyl ether, aryl ether, alkane-
2. Results and discussion
thioether, and acyl ester derivatives of
3 and acyl,
2.1. Synthesis
acyloxymethylcarbamate and ureido derivatives of 4, have
been synthesised by several groups, including ourselves
[14e25]. Indeed, one such compound, imidazo[3,4-e]tetra-
zin-5-one temozolomide 5 [18], has received FDA approval
for the treatment of a specific type of brain cancer, anaplastic
astrocytoma (http://www.fda.gov/cder/da/da0899.htm).
CH3
The a-(acylamino)acyl derivatives, 6ael, were readily syn-
thesised by the conventional coupling reaction of the appropri-
ate 1-aryl-3-methyltriazene 4 with an (S )-N-acylamino acid
using dicyclohexylcarbodiimide activation. The 1H NMR
spectra of 6ael display a typical triazene N-methyl singlet
in the range d 3.40e3.53 ppm. For 6d the two methyl groups
of the isopropyl side-chain are diastereotopic and appear as
two doublets.
N
O
N
N
N
N
2.2. Kinetic studies
H2N
O
The hydrolysis of the a-(acylamino)acyltriazenes 6ael to
give the corresponding 3-methyltriazenes 4 (Fig. 1) was stud-
ied in isotonic phosphate buffer and in 80% human plasma
containing 20% isotonic phosphate buffer. The hydrolysis of
6a was also studied in more detail across the pH range 1e
11 in aqueous buffers. Under the reaction conditions the
monomethyltriazenes are unstable, hydrolysing further to the
corresponding anilines. These reactions were easily followed
either by HPLC (monitoring both the loss of starting material
and the formation of products) or by UV spectroscopy (mon-
itoring the loss of starting material). For compound 6g the
5
Given that the incidence of malignant melanoma is on the
increase and is expected to continue to rise [5] there is a need
to develop improved pharmacological entities. This, together
with the need to employ carrier moieties that are non-toxic,
has directed our interests towards triazenes linked to amino
acids via the triazene N-3 nitrogen atom and the acyl group
of the amino acid [26,27]. These triazene amide derivatives
were found to have reactivities very similar to those of amino
acid esters. However, an N-acetylated alanyl derivative

M.deJ. Perry et al. / European Journal of Medicinal Chemistry 44 (2009) 1049e1056
1051
concomitant quantitative formation of N-benzoyl-(S )-alanine
7 (Scheme 1) was also observed (Fig. 2).
water to the base catalyst as being integral to the rate-limiting
hydrolytic step.
Thus, taken together these studies reveal that in solutions at
a physiological pH of 7.4 the hydrolysis of aminoacyltriazenes
comprises three components: pH-independent, base-catalysed
and buffer-catalysed reactions. Consequently, general conclu-
sions about the factors affecting the relative magnitudes of
the t1/2 values for the compounds in Table 1 ought to be
made with caution, although in pH 7.4 phosphate buffer the
major reaction is the hydroxide ion catalysed process. Never-
theless, certain features are apparent. First, compounds 6a,hel
reveal that the reaction is influenced by the substituent, -X,
present in the aryl group, electron-attracting substituents in-
creasing the rate of hydrolysis (these compounds afford a Ham-
mett correlation with a r value of þ0.81; r2 ¼ 0.99; n ¼ 6).
Second, the structure of the a-amino acid unit is also impor-
tant, the glycine derivative 6b being the most reactive and
the sterically hindered valine derivative 6d the most stable.
Third, the structure of the N-acyl group (compounds 6a,ee
g) has only a small effect on the rate of hydrolysis.
2.2.1. Hydrolysis in aqueous buffers
Half-lives for the hydrolysis of triazenes 6ael in isotonic
phosphate buffer are given in Table 1. Inspection of Table 1
shows that these compounds decompose at physiological pH
with half-lives ranging from 7.6 to 84 h.
To better understand the chemistry of these potential pro-
drug derivatives, compound 6a was studied in more detail.
Pseudo-first-order rate constants, kobs, for its hydrolysis were
determined in aqueous buffers (using several buffer concentra-
tions at each pH), in deuterated buffers, and in NaOH and
H2SO4 solutions (Table 2). The values of kobs depend upon
both the pH of the solution and the buffer concentration. Using
the intercepts of plots of kobs versus buffer concentration, to-
gether with kobs values determined in NaOH and H2SO4 solu-
tions, a pH-rate profile can be constructed (Fig. 3), which
shows that there are regions of acid- and base-catalysis, as
well as a pH-independent process. The compound is most sta-
ble between pH 3 and 5.5. At pH 7.4 the most important reac-
tion is that catalysed by hydroxide ion (by about a factor of 20).
Further, comparison of the second-order rate constant for hy-
droxide-catalysed reaction, 5.7 Mꢀ1 sꢀ1, with that for the cor-
2.2.2. Hydrolysis in human plasma
Blood serum and plasma are known to contain a range of
enzymes that catalyse the hydrolysis of esters and amides
[30]. Consequently, we were particularly interested in examin-
ing the hydrolysis of the a-(acylamino)acyltriazenes 6ael in
human plasma. The results in Table 1 clearly show that, in
plasma, all the compounds studied liberate the corresponding
monomethyltriazenes; that they do so at rates that are 3e15
times faster than in phosphate buffer also implies that the a-
(acylamino)acyltriazenes are substrates for plasma enzymes.
This is in contradistinction to the more reactive parent amino-
acyltriazenes, which were found not to be substrates for
plasma enzymes [26].
Although the prodrugs liberate the cytotoxic monomethyl-
triazene more rapidly in plasma than in phosphate buffer,
Fig. 5 reveals a correlation (albeit with an r2 correlation coef-
ficient of 0.67) between the half-lives for the two processes.
The implication is that similar electronic and steric effects op-
erate in both cases.
responding deuteroxide-catalysed reaction, 7.8 Mꢀ1 sꢀ1
affords a solvent deuterium isotope effect of 0.73. This value
identifies hydrolysis proceeding by hydroxide ion acting as
a nucleophilic entity rather than as a general base [28].
,
The rate constants for catalysis of the hydrolysis reaction
by the buffer species, kB, were determined from the slopes
of plots of kobs versus buffer concentration. Fig0. 4 shows that
the correlation between kB and the buffer pKa (i.e. the pKa
corrected for ionic strength [29]) is linear, giving a Brønsted
b value of 0.7. Significantly, the sterically hindered base
0
2,2,4,4-tetramethylpiperidine, pKa 11.21, exhibits a catalytic
rate constant kB of 0.41 Mꢀ1 sꢀ1, essentially not different
from that of the non-sterically hindered parent 0piperidine, kB
0.42 Mꢀ1 sꢀ1, which has an almost identical pKa of 11.26. To-
gether with the Brønsted plot, these data imply that, in contrast
to hydroxide ion, the buffer materials exert catalysis by acting
as general bases rather than nucleophiles. This was corrobo-
rated by the deuterium isotope effect for the buffer catalysed
process, which was determined for pyridine, a relatively
weak catalyst, and piperidine, a relatively efficient catalyst.
The data are shown in Table 3; the rates are more rapid in
H2O compared to D2O, which identifies proton transfer from
We observed previously that, despite formally being amide
derivatives, 3-acyltriazenes have a reactivity similar to that of
esters [15,21]. Since substrate lipophilicity is one of the main
features that influence esterase activity, lower enzyme reactiv-
ity being associated with increasing lipophilicity [31], we ex-
amined the effect of substrate lipophilicity on the rate of
H
N
O
H
N
O
N
Ar
H
N
Ar
+
N
N
OH
N
N
O
Me
6g
Me
O
Me
Me
7
4
H2N Ar
Scheme 1.

1052
M.deJ. Perry et al. / European Journal of Medicinal Chemistry 44 (2009) 1049e1056
the anticancer monomethyltriazenes. The compounds are suit-
100
ably stable in phosphate buffer (t1/2; ca. 8e84 h) while, in
plasma, those containing an electron withdrawing group in
the triazene moiety together with a non-bulky group (acylami-
no)acyl moiety decompose over 1e2 h. By way of compari-
son, temozolomide, the recently introduced clinically
approved triazene, is likewise a chemically-activated prodrug
that has a half-life in both pH 7.4 phosphate buffer and plasma
of ca. 2 h [32]. Thus, the acylation of the a-amino group seems
to be an effective means of reducing the chemical reactivity of
the a-aminoacyl derivatives while retaining a rapid rate of en-
zymatic hydrolysis. Usually this kind of tunable dual reactivity
is only achieved using the double prodrug strategy [33], but
that involves formaldehyde liberation over which there toxic-
ity concerns [34e36]. Additionally, these a-(acylamino)acyl
derivatives display log P values that suggest they should be
well absorbed by biological membranes.
80
60
40
20
0
0
5
10
15
20
25
t / h
Fig. 2. Time course for the hydrolysis of 6g in 80% human plasma at 37 ꢁC:
B, 6g; -, N-benzoyl-(S )-alanine; ꢂ, 4; , corresponding aniline.
6
4. Experimental protocols
monomethyltriazene liberation from the prodrugs 6ael. The
partition coefficients between octanol and pH 7.4 phosphate
buffer at 25 ꢁC for compounds 6ael (Table 1) give values of
log P that cover a wide range of lipophilicities (0.69e3.04).
However, Fig. 6 reveals no correlation between lipophilicity
and plasma stability and we infer that prodrug stability de-
pends only on those factors that govern the chemical reactivity
of the aminoacyltriazene unit.
Warning: All triazenes used in this study should be consid-
ered as mutagenic and/or carcinogenic and appropriate care
should be taken to handle them safely.
Melting points were determined using a Kofler camera
Bock-Monoscop ‘‘M’’ and are uncorrected. IR spectra were
recorded as KBr discs using a PerkineElmer 1310 spectropho-
1
tometer. H NMR spectra were recorded in CDCl3 solutions
using a Jeol JNM LA-300 spectrometer; chemical shifts, d,
are reported as parts per million from Me4Si, and coupling
constants, J, in hertz. Mass spectra were recorded using
a VG Mass Lab 20e250 spectrometer. High-performance liq-
uid chromatography (HPLC) was performed using a Merck-
Hitachi system consisting of an isochrom LCL-7100 pump,
a Lichrospher 100 RP-8 (5 mm) column, an UV L-7400 detec-
tor and a D-2500 integrator. Elemental analyses were obtained
from Medac Ltd., Brunel Science Park, Englefield Green, Eg-
ham, Surrey, UK. All chemicals were reagent grade except
those for kinetic studies and HPLC, which were of analytical
or LiChrosolv (Merck) grade. 1-Aryl-3-methyltriazenes were
synthesised by previously published methods [37]. The (S )-
N-acetylamino acids were purchased from Sigma. N-Propa-
noyl- and N-butanoylalanine were synthesised by the reaction
of tert-butyl (S )-2-aminopropanoate with the corresponding
propanoyl and butanoyl chlorides followed by the conven-
tional deprotection of the tert-butyl ester using trifluoroacetic
acid [38].
3. Conclusions
a-(Acylamino)acyltriazenes 6ael have two amide func-
tional groups susceptible to hydrolysis, the N-acyl group and
the triazene amide. Our results e dependence on the triazene
aryl substituent, reduced reactivity with bulky substituents at
the a-carbon of the amino acid, minimal effect of the N-acyl
group, and isolation of the N-acylamino acid hydrolysis co-
product e point to reaction proceeding via hydrolysis of the
triazene amide (Fig. 7) and therefore that the a-(acylami-
no)acyl derivatives provide a potential prodrug system for
Table 1
Half-lives, t1/2, for the triazenes 6ael in both pH 7.7 isotonic phosphate buffer
and 80% human plasma at 37 ꢁC, together with log P values
b
t1/2 (h)
log Pexp
log Pcalc
pH 7.7 buffera
Plasmab
6a
6b
6c
6d
6e
6f
6g
6h
6i
9.8 ꢃ 0.4
7.6 ꢃ 0.1
1.0 ꢃ 0.1
1.2 ꢃ 0.2
2.4 ꢃ 0.3
5.7 ꢃ 1.1
1.3 ꢃ 0.2
1.1 ꢃ 0.2
1.6 ꢃ 0.3
2.3 ꢃ 0.3
4.6 ꢃ 0.5
7.6 ꢃ 0.6
5.7 ꢃ 0.9
15.1 ꢃ 1.2
1.41 ꢃ 0.01
1.1 ꢃ 0.03
2.76 ꢃ 0.05
2.18 ꢃ 0.04
1.87 ꢃ 0.06
2.28 ꢃ 0.02
3.04 ꢃ 0.03
1.51 ꢃ 0.09
2.52 ꢃ 0.04
0.69 ꢃ 0.06
2.8 ꢃ 0.04
2.45 ꢃ 0.02
1.93 ꢃ 0.51
1.58 ꢃ 0.50
3.86 ꢃ 0.52
2.81 ꢃ 0.51
2.46 ꢃ 0.51
2.99 ꢃ 0.51
3.84 ꢃ 0.51
1.81 ꢃ 0.49
2.87 ꢃ 0.48
0.89 ꢃ 0.48
3.01 ꢃ 0.54
2.45 ꢃ 0.46
13.9 ꢃ 0.0
84.3 ꢃ 1.8
10.1 ꢃ 0.3
9.1 ꢃ 0.2
4.1. General procedure for the synthesis of a-
(acylamino)acyltriazenes (6ael)
N,N0-Dicyclohexylcarbodiimide (2.4 mmol) was added to
a solution of the N-acylamino acid (1.9 mmol) in CH2Cl2
(10 ml). After 30 min at room temperature, triethylamine
(0.19 mmol) and the appropriate 1-aryl-3-methyltriazene
(1.9 mmol) in CH2Cl2 (5 ml) were added. After 48 h at room
temperature, N,N0-dicyclohexylurea by-product was removed
by filtration and the solvent removed under reduced pressure.
9.2 ꢃ 0.0
13.9 ꢃ 0.4
15.7 ꢃ 0.6
18.4 ꢃ 0.2
25.5 ꢃ 0.2
46.5 ꢃ 1.5
6j
6k
6l
a
Values represent the means of at least 2 determinations ꢃ SD.
b
Values represent the means of at least 3 determinations ꢃ SD.

M.deJ. Perry et al. / European Journal of Medicinal Chemistry 44 (2009) 1049e1056
1053
Table 2
Pseudo-first-order hydrolysis rate constants for 6a in aqueous buffers at 25 ꢁC
Buffer
H2SO4
Buffer/M
pH (pHcalc
)
k
obs/10ꢀ5 sꢀ1
Buffer
Buffer/M
pH (pHcalc
)
k
obs/10ꢀ5 sꢀ1
0.0946
0.0085
(0.72)
(1.78)
0.102
0.013
Morpholinee
0.0010
0.0040
0.0080
0.010
7.53
8.67
8.71
8.60
0.711
0.808
0.841
0.869
Formatea
0.0020
0.0200
0.0400
0.1000
3.72
3.43
3.41
3.44
0.0055
0.0057
0.0058
0.0068
Piperazinef
Piperidineg
0.01
0.02
0.08
0.16
0.20
9.21
9.31
9.39
9.47
9.36
12.6
15.3
19.8
26.2
27.0
Pyridineb
0.01
0.05
0.1
5.73
5.78
5.78
5.72
5.70
0.0097
0.0126
0.0232
0.0400
0.0625
0.2
0.4
0.01
0.05
0.10
0.30
0.40
10.84
11.11
11.14
11.15
11.17
477
830
908
Phosphatec
Imidazoled
0.0063
0.0127
0.0422
0.8440
6.62
6.62
6.66
6.70
0.0036
0.0041
0.0045
0.0048
2127
2481
2,2,4,4-Tetramethylpiperidineh
0.01
0.02
0.03
0.04
10.86
11.06
11.12
11.14
532
775
0.01
0.08
0.16
0.20
0.30
6.90
7.01
7.06
7.02
7.03
0.0532
0.1085
0.171
0.191
0.238
995
1156
NaOH
0.0006
0.0011
0.0013
0.0018
0.0024
0.0035
(10.80)
(11.03)
(11.12)
(11.26)
(11.38)
(11.55)
454
710
1016
1265
1536
2107
a
0
0
0
0
0
0
0
0
pKa 3.61.
b
c
d
e
f
pKa 5.09.
pKa 7.35.
pKa 7.09.
pKa 8.47.
pKa 9.95.
g
h
pKa 11.26.
pKa 11.21.
The crude residue was subjected to chromatography (silica gel
60, 70e230 mesh ASTM, Merck) using one of the following
systems: hexane:diethyl ether (for 6f), CH2Cl2:ethyl acetate
(for 6e,g), ethyl acetate (for 6aed,h,i,k,l) and ethyl acetate:-
MeOH (for 6j). The isolated N-(acylamino)acyltriazenes
were subsequently recrystallised from hexane/ethyl acetate
or hexane/CH2Cl2.
4.68 (2H, d, J ¼ 5.0 Hz, COCH2N), 6.37 (1H, br s, NH ),
7.66e7.71 and 7.72e7.76 (4H, AA0XX0, Jortho ¼ 9.0 Hz, Ar);
m/z (EIMS) 259 (Mþ, 8%), 130 (70), 102 (100) and 100
(23). Found: C, 55.1; H, 5.3; N, 26.9; C12H13N5O requires
C, 55.59; H, 5.05; N, 27.01%.
4.1.3. 3-(2-(Acetylamino)-3-phenylpropanoyl)-1-(4-cyano
phenyl)-3-methyltriazene (6c)
4.1.1. 3-(2-(Acetylamino)propanoyl)-1-(4-cyanophenyl)-3-
methyltriazene (6a)
Yield 58%; m.p. 178e179 ꢁC; nmax/cmꢀ1 3281, 2221, 1715,
1641; dH 2.04 (3H, s, Ac), 3.12 (2H, m, CH2Ph), 3.40 (3H, s,
NMe), 5.97 (1H, m, CH ), 6.31 (1H, br d, J ¼ 8.4 Hz, NH ),
7.00e7.15 (5H, m, Ph), 7.56e7.61 and 7.70e7.75 (4H,
AA0XX0, Jortho ¼ 8.7 Hz, Ar); m/z (EIMS) 349 (Mþ, 6%), 148
(11), 102 (41) and 190 (50). Found: C, 65.2; H, 5.5; N, 20.4;
C19H19N5O requires C, 65.32; H, 5.48; N, 20.04%.
Yield 34%; m.p. 203e204 ꢁC; nmax/cmꢀ1 3326, 3238,
2224, 1719, 1645; dH 1.47 (3H, d, J ¼ 7.6 Hz, MeCH), 2.07
(3H, s, Ac), 3.49 (3H, s, NMe), 5.64 (1H, m, CH ), 6.37 (1H,
br d, J ¼ 7.5 Hz, NH ), 7.67e7.71 and 7.72e7.77 (4H,
AA0XX0, Jortho ¼ 8.7 Hz, Ar); m/z (EIMS) 274 (MHþ, 20%),
130 (7), 102 (20) and 114 (23). Found: C, 57.1; H, 5.6; N,
25.6; C13H15N5O2 requires C, 57.13; H, 5.53; N, 25.63%.
4.1.4. 3-(2-(Acetylamino)-3-methylbutanoyl)-1-(4-cyano
phenyl)-3-methyltriazene (6d)
4.1.2. 3-(2-(Acetylamino)acetyl)-1-(4-cyanophenyl)-3-
methyltriazene (6b)
Yield 56%; m.p. 172e173 ꢁC; nmax/cmꢀ1 3321, 3279,
2232, 1716, 1650; dH 2.13 (3H, s, Ac), 3.51 (3H, s, NMe),
Yield 35%; m.p. 185 ꢁC; nmax/cmꢀ1 3371, 2222, 1700,
1670; dH 0.92 and 1.01 (each 3H, 2 ꢂ d, J ¼ 6.8 Hz,
2 ꢂ Me), 2.09 (3H, s, Ac), 2.16 (1H, m, Me2CH ), 3.49 (3H,
s, NMe), 5.71 (1H, m, COCHN), 6.19 (1H, br d, J ¼ 9.2 Hz,

1054
M.deJ. Perry et al. / European Journal of Medicinal Chemistry 44 (2009) 1049e1056
-1
-2
-3
-4
-5
-6
-7
-8
Table 3
Solvent isotope effects for the hydrolysis of 6a in pyridine and piperidine
buffers
Buffer
pH/pD
kB/10ꢀ6 sꢀ1
kHB/kDB
Pyridine
H2O
D2O
5.74
6.31
1.75
0.86
2.03
Piperidine
H2O
D2O
10.84
12.02
419 785
202 406
2.08
NMe), 5.64 (1H, quin, J ¼ 7.0 Hz, CH ), 6.30 (1H, br d,
J ¼ 7.0 Hz, NH ), 7.67e7.71 and 7.72e7.76 (4H, AA0XX0,
Jortho ¼ 9.0 Hz, Ar); m/z (LCMS) 288 (MHþ, 5%), 130 (100),
102 (40) and 128 (7). Found: C, 58.5; H, 6.0; N, 23.9;
C14H17N5O2 requires C, 58.52; H, 5.96; N, 24.37%.
0
2
4
6
pH
8
10
12
4.1.6. 3-(2-(Butanoylamino)propanoyl)-1-(4-cyanophenyl)-
3-methyltriazene (6f)
Yield 71%; m.p. 160e161 ꢁC; nmax/cmꢀ1 3326, 3242,
2223, 1722, 1644; dH 0.95 (3H, t, J ¼ 7.0 Hz, MeCH2), 1.44
(3H, d, J ¼ 7.0 Hz, MeCH), 1.67 (2H, m, CH2Me), 2.22 (2H,
t, J ¼ 7.0 Hz, COCH2), 3.46 (3H, s, NMe), 5.64 (1H, quin,
J ¼ 7.0 Hz, CH ), 6.30 (1H, br d, J ¼ 7.0 Hz, NH ), 7.66e
7.71 and 7.72e7.76 (4H, AA0XX0, Jortho ¼ 8.8 Hz, Ar); m/z
(LCMS) 302 (MHþ, 5%), 130 (38) and 142 (15). Found: C,
59.7; H, 6.4; N, 23.0; C15H19N5O2 requires C, 59.79; H,
6.36; N, 23.24%.
Fig. 3. pH-rate profile for compound 6a. The points are experimental, thþe
curve is the best fit using kOH 5.7 Mꢀ1 sꢀ1, kH O 4.4 ꢂ 10ꢀ8 sꢀ1 and kH
2
5.17 ꢂ 10ꢀ5 Mꢀ1 sꢀ1
.
NH ), 7.72e7.74 and 7.77e7.79 (4H, AA0XX0, Jortho ¼ 6.0 Hz,
Ar); m/z (EIMS) 142 (39), 130 (17), 114 (59), 102 (21), 72
(100) and 43 (50). Found: C, 59.5; H, 6.4; N, 23.3;
C15H19N5O2 requires C, 59.79; H, 6.36; N, 23.24%.
4.1.5. 1-(4-Cyanophenyl)-3-methyl-3-(2-(propanoylamino)
propanoyl)triazene (6e)
Yield 51%; m.p. 189 ꢁC; nmax/cmꢀ1 3324, 3241, 2223,
1721, 1647; dH 1.16 (3H, t, J ¼ 7.6 Hz, Me), 1.43 (3H, d,
J ¼ 7.0 Hz, Me), 2.27 (2H, q, J ¼ 7.7 Hz, CH2), 3.46 (3H, s,
4.1.7. 3-(2-(Benzoylamino)propanoyl)-1-(4-cyanophenyl)-3-
methyltriazene (6g)
Yield 21%; m.p. 196e197 ꢁC; nmax/cmꢀ1 3340, 2226, 1717,
1640; dH 1.59 (3H, d, J ¼ 7.0 Hz, MeCH), 3.53 (3H, s, NMe),
5.86 (1H, m, CH ), 7.09 (1H, br d, J ¼ 8.0 Hz, NH ), 7.46e
7.57 and 7.84e7.86 (5H, m, Ph), 7.75e7.80 (4H, AA0XX0,
Jortho ¼ 6.6 Hz, Ar); m/z (FABꢀ MS) 334 (MeHꢀ), (EIMS)
130 (12%), 102 (38) and 176 (12). Found: C, 64.5; H, 5.1; N,
20.8; C18H17N5O2 requires C, 64.47; H, 5.11; N, 20.88%.
3
2
1
0
3
2.5
2
-1
-2
-3
-4
-5
-6
-7
-8
1.5
2.6
2.8
3
3.2
3.4
3.6
3.8
0
2
4
6
8
10
12
14
16
18
log t1/2 (phosphate buffer)
pKa'
Fig. 5. Correlation of the plasma half-lives with the phosphate buffer half-lives
for compounds 6ael.
Fig. 4. Brønsted plot for compound 6a.

M.deJ. Perry et al. / European Journal of Medicinal Chemistry 44 (2009) 1049e1056
1055
3
2,5
2
1.93 (3H, s, Ac), 3.35 (3H, s, NMe), 5.50 (1H, quin.,
J ¼ 7.7 Hz, CH ), 6.85 (1H, br d, J ¼ 8.0 Hz, NH), 7.54e
7.59 and 7.83e7.88 (4H, AA0XX0, Jortho ¼ 9.0 Hz, Ar); m/z
(EIMS) 292 (MHþ, 23%), 148 (100), 120 (25) and 114 (3).
Found: C, 53.0; H, 5.9; N, 23.7; C13H17N5O3 requires C,
53.60; H, 5.88; N, 24.04%.
4.1.11. 3-(2-(Acetylamino)propanoyl)-1-(4-bromophenyl)-
3-methyltriazene (6k)
Yield 30%; m.p 192e194 ꢁC; nmax/cmꢀ1 3322, 1667, 1626;
d 1.43 (3H, d, J ¼ 7.0 Hz, MeCH), 2.03 (3H, s, Ac), 3.42 (3H,
s, NMe), 5.59 (1H, quin, J ¼ 7.0 Hz, CH ), 6.41 (1H, br d,
J ¼ 7.7 Hz, NH ), 7.47e7.53 and 7.54e7.59 (4H, AA0BB0,
Jortho ¼ 9.0 Hz, Ar); m/z (LCMS) 328/326 (MHþ, 17%), 185/
183 (30), 157/155 (20) and 114 (8). Found: C, 44.2; H, 4.7;
N, 17.0; C12H15N4O2Br requires C, 44.05; H, 4.62; N, 17.12%.
1,5
0
0,5
1
1,5
2
2,5
3
3,5
log P
Fig. 6. Plot of the plasma half-lives versus lipophilicity for compounds 6ael.
4.1.12. 3-(2-(Acetylamino)propanoyl)-1-(4-tolyl)-3-
methyltriazene (6l)
4.1.8. 3-(2-(Acetylamino)propanoyl)-1-(4-acetylphenyl)-3-
methyltriazene (6h)
Yield 31%; m.p. 136e137 ꢁC; nmax/cmꢀ1 3345, 1687,
1673; dH 1.44 (3H, d, J ¼ 7.0 Hz, MeCH), 2.03 (3H, s, Ac),
2.38 (3H, s, MePh), 3.42 (3H, s, NMe), 5.60 (1H, quin,
J ¼ 7.0 Hz, CH ), 6.47 (1H, br d, J ¼ 7.1 Hz, NH ), 7.18e
7.29 and 7.52e7.54 (4H, AA0XX0, Jortho ¼ 8.4 Hz, Ar); m/z
(EIMS) 262 (Mþ, 6%), 119 (67) and 91 (100). Found: C,
59.9; H, 7.0; N, 20.9; C13H18N4O2 requires C, 59.53; H,
6.92; N, 21.36%.
Yield 73%; m.p. 175e174 ꢁC; nmax/cmꢀ1 3344, 1692,
1671, 1628; dH 1.45 (3H, d, J ¼ 7.0 Hz, MeCH), 2.04 (3H, s,
Ac), 2.62 (3H, s, PhAc), 3.46 (3H, s, NMe), 5.62 (1H, quin.,
J ¼ 7.1 Hz, CH ), 6.39 (1H, br d, J ¼ 7.5 Hz, NH), 7.61e
7.70 and 7.97e8.01 (4H, AA0XX0, Jortho ¼ 8.8 Hz, Ar); m/z
(EIMS) 291 (MHþ, 49%), 147 (20) and 119 (50). Found: C,
57.9; H, 6.4; N, 18.8; C14H18N4O3 requires C, 57.92; H,
6.25; N, 19.30%.
4.1.9. 3-(2-(Acetylamino)propanoyl)-1-(4-ethoxycarbonyl
phenyl)-3-methyltriazene (6i)
4.2. Experimental procedure for hydrolysis studies
Yield 48%; m.p. 142 ꢁC; nmax/cmꢀ1 3530, 1717, 1648; dH
1.40 (3H, t, J ¼ 7.0 Hz, MeCH2), 1.45 (3H, d, J ¼ 7.0 Hz,
MeCH), 2.04 (3H, s, Ac), 3.46 (3H, s, NMe), 4.38 (2H, q,
J ¼ 7.0 Hz, CH2O), 5.62 (1H, quin, J ¼ 7.1 Hz, CH ), 6.39
(1H, br d, J ¼ 7.5 Hz, NH), 7.63e7.68 and 8.09e8.14 (4H,
AA0XX0, Jortho ¼ 8.8 Hz, Ar); m/z (EIMS) 321 (MHþ, 10%),
177 (100) and 149 (50). Found: C, 56.2; H, 6.4; N, 17.2;
C15H20N4O4 requires C, 56.24; H, 6.29; N, 17.49%.
Plasma was derived from seven different healthy individ-
uals, pooled, and kept at ꢀ18 ꢁC until required. The 3-(acyla-
mino)acyltriazenes were incubated at 37 ꢁC in 80% human
plasma diluted with 0.066 M pH 7.4 isotonic phosphate buffer.
At appropriate intervals samples of the plasma reactions were
withdrawn, diluted with acetonitrile (0.4 ml) and centrifuged
at 15 000 rpm for 6 min. The clear supernatant was analysed
by HPLC. The 3-(acylamino)acyltriazenes were also incubated
at 37 ꢁC in 0.066 M pH 7.7 isotonic phosphate buffer. Samples
of these reaction mixtures were analysed directly by HPLC us-
ing a mobile phase of acetonitrile:water (20:80 to 50:50) and
a detector wavelength of 290 nm. Values of t1/2 were calcu-
lated from the decrease in substrate concentration using
4.1.10. 3-(2-(Acetylamino)propanoyl)-3-methyl-1-(4-amino
carbonylphenyl) triazene (6j)
Yield 35%; m.p. 209e210 ꢁC; nmax/cmꢀ1 3419, 3351,
3303, 1694, 1664, 1619; dH 1.34 (3H, d, J ¼ 6.9 Hz, MeCH),
-
H
N
O
H
N
O
N
Ar
N
Ar
N
N
N
N
OH
O
R
Me
O
R
Me
(b )
(a)
O
-
H
H
HO
H
N
O
B
H
N
Ar
+
-
N
N
O
Me
O
R
Fig. 7. Pathways of the hydroxide (a) and general base (b) catalysed hydrolyses of triazene prodrugs 6ael.

1056
M.deJ. Perry et al. / European Journal of Medicinal Chemistry 44 (2009) 1049e1056
[10] J.L. Skibba, D.D. Beal, G. Ramirez, G.T. Bryan, Cancer Res. 30 (1970)
147e150.
Enzfitter version C, 1987 (Elsevier-Biosoft). For 6g samples of
the reaction mixtures were also analysed by HPLC using pH
3.5 phosphate buffer containing 16% acetonitrile as eluant
and a detector wavelength of 250 nm, this allowed the quanti-
fication of N-benzoylalanine.
[11] F.W. Kruger, R. Preussman, N. Niepelt, Biochem. Pharmacol. 20 (1971)
¨
529e533.
[12] C.J. Rutty, D.R. Newell, R.B. Vincent, G. Abel, P.M. Goddard,
S.J. Harland, et al., Br. J. Cancer 48 (1983) 140.
[13] A. Katsoulas, Z. Rachid, F. Brahimi, J. McNamee, B.J. Jean-Claude,
Leukemia Res. 29 (2005) 693e700.
[14] L.M. Cameron, R.J. LaFrance, C.M. Hemens, K. Vaughan,
R. Rajaraman, D.C. Chubb, et al., Drug Design 1 (1985) 27e36.
[15] E. Carvalho, J. Iley, E. Rosa, J. Chem. Soc., Perkin Trans. 2 (1993)
865e870.
4.3. Experimental procedure for partition coefficients
For compounds 6ael, partition coefficients were deter-
mined in octanol-pH 7.4 phosphate buffer at 25 ꢁC. Each
phase was mutually saturated before the experiment. The com-
pounds were dissolved in octanol and the octanol-pH 7.4 mix-
tures were shaken for 30 min to reach an equilibrium
distribution; each phase was analysed separately by the
HPLC method described above. The partition coefficients,
[16] E. Carvalho, A.P. Francisco, J. Iley, E. Rosa, Bioorg. Med. Chem. 8
(2000) 1719e1725.
[17] E. Carvalho, A.P. Francisco, J. Iley, E. Rosa, Eur. J. Org. Chem. (2005)
2056e2063.
[18] A.S. Clark, B. Deans, M.F.G. Stevens, M.J. Tisdale, R.T. Wheelhouse,
B.J. Denny, et al., J. Med. Chem. 38 (1995) 1493e1504.
[19] J. Iley, E. Rosa, L. Fernandes, J. Chem. Res. (1987) (S) 264, (M)
2216e2229.
Pexp, were obtained from the ratio of the peak area in octanol
to the peak area in buffer. Calculated values of P, expressed as
log Pcalc, were obtained using the software Chemsketch ver-
sion 3.60 from Advanced Chemistry Development Inc. (Ad-
vanced Chemistry Development Inc., 90 Adelaide St. West,
Suite 702, Toronto, Ontario ONM5H 2L3, Canada). Calcu-
lated log P values were typically ca. 0.5 (ꢃ0.2) units too large
and therefore overestimated lipophilicity.
[20] J. Iley, R. Moreira, E. Rosa, J. Chem. Soc., Perkin Trans. 2 (1987)
1503e1508.
[21] J. Iley, G. Ruecroft, E. Carvalho, E. Rosa, J. Chem. Res. (1989) (S)
162e163, (M) 1264e1273.
[22] J. Iley, R. Moreira, E. Rosa, J. Chem. Soc., Perkin Trans. 1 (1991)
3241e3247.
[23] H.W. Manning, L.M. Cameron, R.J. LaFrance, K. Vaughan,
R. Rajaraman, Drug Design 1 (1985) 37e43.
[24] M.P. Merrin, D.L. Hooper, R.J. LaFrance, R. Snooks, K. Vaughan, Can. J.
Chem. 70 (1992) 144e150.
Acknowlegments
[25] M.J. Wanner, M. Koch, G.-J. Koomen, J. Med. Chem. 47 (2004)
6875e6883.
This work was supported by JNICT and PRAXIS XXI,
project no 2/2.1/SAU/1404/95.
[26] E. Carvalho, J. Iley, M.J. Perry, E. Rosa, Pharm. Res. 15 (1998)
931e935.
[27] E. Carvalho, J. Iley, M.J. Perry, E. Rosa, J. Chem. Soc., Perkin Trans. 2
(1998) 2375e2380.
References
[28] H. Maskill (Ed.), Structure and Reactivity in Organic Chemistry, Oxford
University Press, Oxford, 1999 (Chapter 4).
[1] H. Bundgaard (Ed.), Design of Prodrugs, Elsevier, Amsterdam, 1985.
[2] D. Bhosle, S. Bharambe, N. Gairola, S.S. Dhaneshwar, Indian J. Pharm.
Sci. 68 (2006) 286e294.
[29] C.W. Davies, J. Chem. Soc. (1938) 2093e2098.
[30] F.J. Leinweber, Drug Metab. Rev. 18 (1987) 379e439.
´ ´
[31] T.L. Huang, A. Szekacs, T. Uematsu, E. Kuwano, A. Parkinson,
[3] B. Testa, J.M. Mayer, Hydrolysis in Drug and Prodrug Metabolism, Wi-
ley-VCH, Zurich, Switzerland, 2003.
¨
B.D. Hammock, Pharm. Res. 10 (1993) 639e648.
[32] H.S. Friedman, T. Kerby, H. Clavert, Clin. Cancer Res. 6 (2000) 2585e
2597.
[4] H. Bundgaard, Design and application of prodrugs, in: P. Krogsgaard-
Larsen, H. Bundgaard (Eds.), A Textbook of Drug Design and Develop-
ment, Harwood Academic Publishers, Philadelphia, 1991, pp. 113e192.
[5] G.L. Cohen, C.I. Falkson, Drugs 55 (1998) 791e799.
[6] L. Meer, R.C. Janzer, P. Kleihues, G.F. Kolar, Biochem. Pharmacol. 35
(1986) 3243e3247.
[33] H. Bundgaard, U. Klixbull, E. Falch, Int. J. Pharm. 29 (1986) 19e28.
[34] J. Alexander U.S. Patent 5684018; 1997.
[35] K. Beaumont, R. Webster, I. Gardner, K. Dack, Curr. Drug Metabol. 4
(2003) 461e485.
[36] A. Patnaik, E.K. Rowinsky, M.A. Villalona, L.A. Hammond,
C.D. Britten, L.L. Siu, et al., Clin. Cancer Res. 8 (2002) 2142e2148.
[37] S.C. Cheng, J. Iley, J. Chem. Res. (1983) (S) 320e321.
[38] M.W. Williams, G.T. Young, J. Chem. Soc. (1963) 881e889.
[7] M.F.G. Stevens, A. Gescher, C.P. Turnbull, Biochem. Pharmacol. 28
(1979) 769e776.
[8] G.F. Kolar, R. Carubelli, Cancer Lett. 7 (1979) 209e214.
[9] G.F. Kolar, M. Maurer, M. Wildschutte, Cancer Lett. 10 (1980) 235e241.