Kinetics of degradation of 4-imidazolidinone prodrug types obtained from reacting prilocaine with formaldehyde and acetaldehyde

Article abstract of DOI:10.1016/S0928-0987(03)00198-2

The kinetics of decomposition of 4-imidazolidinone prodrug types obtained by reacting prilocaine (I) with formaldehyde and acetaldehyde has been studied in aqueous solution in the pH range 1-7.4 at 60 and 37°C, respectively. At pH<5 the hydrolysis of the derivative derived from formaldehyde (II) to yield I obeyed apparent first-order kinetics. At higher pH, the decomposition reactions proceeded to an equilibrium and the reactions could be described by first- and second-order reversible kinetics. A plot of the logarithm of the apparent first-order rate constants for hydrolysis of II against pH resulted in a sigmoidal-shaped pH-rate profile characteristic for the hydrolysis of many N-Mannich bases. A half-life at pH 7.4 (60°C) of 6.9h for compound II was calculated. Compared to II the 4-imidazolidinone derived from acetaldehyde (III) exhibited enhanced instability in aqueous buffer solutions. The decomposition was followed at 37°C monitoring the decrease in concentration of intact (III). At acidic pH the reactions displayed strict first-order kinetics and the disappearance of III was accompanied by a concomitant formation of I. At pH 7.4, the rate data also applied reasonably well to first-order kinetics despite the observation that small amounts of III was formed at pH 7.4 from a solution containing equimolar concentrations of acetaldehyde and prilocaine (10 ^-4M). In case of III, a bell-shaped pH-rate profile was obtained by plotting the logarithm of the pseudo-first-order rate constants against pH indicating the involvement of a kinetically significant intermediate in the reaction pathway and a change of the rate-limiting step in the overall reaction with pH. For the stability studies performed at pH 6.9 and 7.4 product analysis revealed that parallel to formation of (I) an unknown compound (X) emerged. Compared to III, compound X is hydrolysed to give I at a slower rate (t _50%=30h at 37°C). Based on LC-MS data it is suggested that (X) is an isomeric form of III, which may exist in four diastereomeric forms. Thus, at physiological pH an initial relatively fast regeneration of I from III is to be expected followed by a slower drug activation resulting from hydrolysis of the isomeric form of III.

Full text of DOI:10.1016/S0928-0987(03)00198-2

European Journal of Pharmaceutical Sciences 20 (2003) 233–240
Kinetics of degradation of 4-imidazolidinone prodrug types obtained
from reacting prilocaine with formaldehyde and acetaldehyde
a
a
b
a,∗
Susan Weng Larsen , Martin Sidenius , Michael Ankersen , Claus Larsen
a
Department of Pharmaceutics, The Danish University of Pharmaceutical Sciences, Universitetsparken 2, DK-2100 Copenhagen, Denmark
b
Novo Nordisk A/S, Novo Nordisk Park, DK-2760 Måløv, Denmark
Received 3 March 2003; received in revised form 24 June 2003; accepted 11 July 2003
Abstract
The kinetics of decomposition of 4-imidazolidinone prodrug types obtained by reacting prilocaine (I) with formaldehyde and acetaldehyde
◦
has been studied in aqueous solution in the pH range 1–7.4 at 60 and 37 C, respectively. At pH < 5 the hydrolysis of the derivative derived from
formaldehyde (II) to yield I obeyed apparent first-order kinetics. At higher pH, the decomposition reactions proceeded to an equilibrium and
the reactions could be described by first- and second-order reversible kinetics. A plot of the logarithm of the apparent first-order rate constants
for hydrolysis of II against pH resulted in a sigmoidal-shaped pH-rate profile characteristic for the hydrolysis of many N-Mannich bases. A
◦
half-life at pH 7.4 (60 C) of 6.9 h for compound II was calculated. Compared to II the 4-imidazolidinone derived from acetaldehyde (III)
◦
exhibited enhanced instability in aqueous buffer solutions. The decomposition was followed at 37 C monitoring the decrease in concentration
of intact (III). At acidic pH the reactions displayed strict first-order kinetics and the disappearance of III was accompanied by a concomitant
formation of I. At pH 7.4, the rate data also applied reasonably well to first-order kinetics despite the observation that small amounts of III was
−
4
formed at pH 7.4 from a solution containing equimolar concentrations of acetaldehyde and prilocaine (10 M). In case of III, a bell-shaped
pH-rate profile was obtained by plotting the logarithm of the pseudo-first-order rate constants against pH indicating the involvement of a
kinetically significant intermediate in the reaction pathway and a change of the rate-limiting step in the overall reaction with pH. For the
stability studies performed at pH 6.9 and 7.4 product analysis revealed that parallel to formation of (I) an unknown compound (X) emerged.
◦
Compared to III, compound X is hydrolysed to give I at a slower rate (t50% = 30 h at 37 C). Based on LC–MS data it is suggested that (X)
is an isomeric form of III, which may exist in four diastereomeric forms. Thus, at physiological pH an initial relatively fast regeneration of I
from III is to be expected followed by a slower drug activation resulting from hydrolysis of the isomeric form of III.
©
2003 Elsevier B.V. All rights reserved.
Keywords: Prodrugs; 4-Imidazolidinones; Cyclic N-Mannich base; Prilocaine; Kinetics of decomposition; Isomeric form
1
. Introduction
(Rasmussen and Bundgaard, 1991; Bak et al., 1999). The
-aminoamide side chain in ampicillin is shown to re-
␣
Transient modification of physicochemical properties
act easily with acetone to yield the clinically used cyclic
N-Mannich base prodrug hetacillin from which the parent
antibiotic is readily regenerated in aqueous solution (Tsuji
and Yamana, 1974). In addition, 4-imidazolidinones have
been derived from carbonyl compounds and ␣-aminoanilides
(Katsura, 1971a; Nelson et al., 1973).
In postoperative pain control a need exists for drug de-
livery systems providing pain relief for 24 h or longer. To
this end, various formulation approaches comprising local
anaesthetics have been investigated (Kohane et al., 2000,
and references cited therein). In a recent paper (Larsen et al.
of amines and NH-acidic compounds such as amides and
imides is possible through formation of prodrugs of the
N-Mannich base type (Bundgaard, 1985). 4-Imidazolidi-
nones formed by condensation of a compound containing
an ␣-aminoamide moiety with aldehydes or ketones can
be considered as cyclic N-Mannich bases (Klixbüll and
Bundgaard, 1984). Bioreversible 4-imidazolidinone deriva-
tives of various peptides have been designed to enhance
lipophilicity and stability against enzymatic degradation
2002a) bupivacaine plasma profiles were determined after
∗
s.c. administration of a vegetable oil solution of bupiva-
caine base and aqueous Marcain , respectively, revealing
Corresponding author. Tel.: +45-3530-6466; fax: +45-3530-6012.
®
E-mail address: csl@dfh.dk (C. Larsen).
0
928-0987/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0928-0987(03)00198-2

2
34
S.W. Larsen et al. / European Journal of Pharmaceutical Sciences 20 (2003) 233–240
Chemicals used for preparation of buffer solutions and
mobile phases were of analytical or chromatographic grade.
H
N
N
H
*
2.2. Synthesis of 4-imidazolidinones of prilocaine
O
The free base of prilocaine was obtained from prilocaine
I
hydrochloride as previously described (Larsen et al., 2002b).
O
N
◦
◦
mp 37–39 C (mp 37–38 C according to Löfgren and Tegner
*
(
1960)).
N
The 4-imidazolidinones of racemic prilocaine were syn-
thesised by procedures previously reported (Nelson et al.,
973) with minor modifications. Briefly, to a solution of
1
II
prilocaine (1.1 g, 5 mmol) in toluene (20 ml) 37% aqueous
formaldehyde (10 ml) was added. The reaction mixture was
refluxed for 3 h and cooled. The organic layer was washed
with saturated NaCl, dried (Na2SO4) and filtered. The sol-
vent was removed in vacuo yielding an oily residue. In
case of III, prilocaine (1.7 g, 7.7 mmol) was dissolved in
toluene (20 ml) and acetaldehyde (5 ml) was added. The re-
action mixture was stirred at ambient temperature for 5 h.
II and III were isolated in the form of their hydrochlo-
ride salts by taking up the oil residues in ethanolic hydro-
gen chloride and precipitation of the salts by addition of
diisopropylether.
O
N
*
N
III
Fig. 1. Cyclic N-Mannich bases of prilocaine (I) derived from formalde-
hyde (II) and acetaldehyde (III).
◦
significantly slower drug release from the former formula-
tion. In case of oil solutions diminishing drug release rates
are to be expected with increasing lipophilicity of the solute
molecule (Hirano et al., 1982; Larsen et al., 2000; Larsen
et al., 2001a). The polar local anaesthetic agent prilocaine
II, HCl: mp 151 C (hot stage microscopy). Anal. Calcd.
(C H N O, HCl): C, 62.56; H, 7.87; N, 10.42. Found:
14
20
2
C, 62.75; H, 8.04; N, 10.46. MS: m/z = 233.0 (m + 1).
1
H NMR (300 MHz; (CD ) SO) δ: 7.36–7.29 (4H, m), 4.98
3 2
(2H, br s), 4.31 (1H, br s), 3.77 (1H, br s), 3.16 (1H, br s),
2.24 (3H, s), 1.88–1.69 (2H, m), 1.61 (3H, d, J = 6.3 Hz),
0.95 (3H, t, J = 7.2 Hz).
(
Fig. 1, compound I) contains an ␣-aminoanilide structure
allowing formation of 4-imidazolidinone derivatives poten-
tially attractive in design of depot oil injectables. Desirable
lipophilicities of such derivatives might be accomplished
by proper selection of the carbonyl component and from
the fact that N-Mannich base formation leads to a signif-
icant decrease in the basicity of the reacting amino group
◦
III, HCl: mp 163.5–164.0 C (hot stage microscopy).
Anal. Calcd. (C H N O, HCl): C, 63.70; H, 8.20; N,
22
2
15
9.91. Found: C, 63.64; H, 8.20; N, 9.85. MS: m/z = 247.1
1
(m + 1). H NMR (300 MHz; (CD ) SO) δ: 7.35–7.30 (4H,
3 2
br s), 5.48 (1H, br s), 4.45 (1H, br s), 3.67 (1H, br s), 2.93
(1H, br s), 2.22 (3H s), 1.99–1.38 (8H, m), 0.96 (3H, t,
J = 7.5 Hz).
(
3
Bundgaard, 1991). 4-Imidazolidinones are weak bases (pKa
–4) and are thus unionised at physiological pH. However,
the feasibility to function as prodrug candidates requires that
parent prilocaine is liberated from the derivatives at suit-
able rates in aqueous solution. The present study was un-
dertaken to investigate the kinetics of degradation of cyclic
N-Mannich bases formed by condensing racemic prilocaine
with formaldehyde and acetaldehyde (Fig. 1) in aqueous
buffer solutions.
2.3. Kinetic measurements
2.3.1. Hydrolysis in aqueous buffer
The decomposition of the cyclic N-Mannich bases II and
◦
III was studied at 60 and 37 ± 0.5 C, respectively, in aque-
ous buffer solutions. The buffers used were phosphate (pH
2.5–3.6 and 6.0–7.4) and acetate (pH 4.0–5.4) and a to-
tal buffer concentration of 0.02 M was employed except for
the experiments where potential buffer effects on degrada-
tion rates were studied. At pH below 2.5 hydrochloric acid
was used. A constant ionic strength of 0.5 was maintained
for each buffer by adding a calculated amount of potas-
sium chloride. The reactions were in most cases initiated by
2
. Materials and methods
2
.1. Materials
Prilocaine hydrochloride was obtained from Sigma
®
−2
Chemi cal Co., USA. Fractionated coconut oil (Viscoleo )
and acetaldehyde were purchased from P. Broste A/S, Den-
mark and Merck, Germany, respectively. Semicarbazide hy-
drochloride was obtained from Fluka Chemie, Switzerland.
adding 100 ␮l of a 10 M stock solution of the compounds
in ethanol (II) or acetonitrile (III) to 9.9 ml of preheated
−4
buffer solution resulting in a final concentration of 10 M.
The reaction mixtures were kept at constant temperature in

S.W. Larsen et al. / European Journal of Pharmaceutical Sciences 20 (2003) 233–240
235
O
kobs or kf
H
N
N
N
+ HCHO
kr
N
H
O
II
I
Fig. 2. Decomposition of compound II and formation of compound II from prilocaine (I) and formaldehyde in aqueous solution in the pH range 2.1–7.4
◦
(
60 C). kobs and kf refer to the observed pseudo-first-order rate constants for hydrolysis of compound II. kr is the pseudo-second-order rate constant for
formation of compound II (kr is negligible at pH below 5).
a water bath and at appropriate time intervals samples were
withdrawn and analysed immediately by HPLC.
a
2.4. pK determination
For the hydrolysis reactions of compound II and III
obeying apparent first-order kinetics the degradation re-
actions were followed for at least two half-lives and the
pseudo-first-order rate constants were calculated from the
slopes of the logarithm of intact 4-imidazolidinone against
time. In the case where the degradation reactions applied to
first- and second-order reversible kinetics (above pH 5 for
compound II (Fig. 2)) the pseudo-first-order rate constant
The pK values of prilocaine and compound II were deter-
a
mined as the mean of two independent potentiometric titra-
◦
tions at 25 C according to Albert and Serjeant (1971). pK
(prilocaine): 7.93± 0.03 (8.02 (Strichartz et al., 1990)). pK
(compound II): 3.89 ± 0.01.
a
a
2
.5. Determination of partition coefficients
for the hydrolysis reaction, k , was derived from Eq. (1)
according to Capellos and Bielski (1972):
f
Apparent partition coefficients (Papp) of compounds I and
®
II between Viscoleo and 0.02 M phosphate buffer pH 7.40
were determined at ambient temperature from experiments
done in duplicate. Equilibrium was attained after rotating the
mixtures for 4 h. Drug concentration in the aqueous phase
ꢀ[
II]0[I]eq + [I]t[II]eq
ꢁ
2[II]0 − [I]eq
ln
=
× kft
(1)
[
II]0([I]eq − [I]t)
[I]eq
−
4
where [II]0 and [II]eq are the initial concentration and the
concentration at equilibrium of compound II, respectively.
before (10 M) and after partition was measured by HPLC
allowing apparent partition coefficients to be calculated ac-
cording to Eq. (3).
[I]t and [I]eq refer to the concentrations of parent prilocaine
at time t and at equilibrium, respectively. At various pH val-
ues, slopes of the linear plots given by Eq. (1) were calcu-
ꢂ
ꢃ ꢂ
ꢃ
C − Cw
Vw
Vo
i
Papp =
(3)
lated allowing the k values to be obtained from:
Cw
f
[
I]eq
kf = slope2
(2)
where Ci and Cw represent the drug concentration before
and after partition, respectively. Vw and Vo are the volumes
of the aqueous and the oil phase.
[II]0 − [I]eq
In a few cases, the reversible kinetics of compound II at
pH ≥ 5 was transformed into apparent first-order kinet-
ics by trapping the formaldehyde formed with semicar-
bazide and following the decrease in the concentration of
intact 4-imidazolidinone. Semicarbazide was included in
the buffers at a concentration of 10 M ([II]0 approx-
imately 10 M). The rate constant, k , was determined
from the slopes of linear plots of the logarithm of residual
-imidazolidinone against time.
2.6. Analyses
1H NMR spectra were recorded on a Varian Gemini 2000
(300 MHz) instrument in (CD3)2SO solution.
−
2
−
4
High-performance liquid chromatography was carried
out using a Merck-Hitachi L-6000 pump connected to a
Merck-Hitachi L-4000 UV detector. A Rheodyne 7125 in-
jection valve equipped with a 20 ␮l sample loop was used.
Reversed phase chromatography was done using an Inertsil
ODS-2 (250 × 4.6 mm, 5 ␮m particles) column equipped
with an Inertsil ODS-2 precolumn (Chrompack). The mo-
bile phases consisted of acetonitrile and 0.1% phosphoric
acid in demineralised water which were made 1 mM with
respect to triethylamine. Simultaneous determination of the
respective 4-imidazolidinone and prilocaine was obtained
by variation of the content of acetonitrile in the mobile phase
f
4
2
8
.3.2. Hydrolysis in human plasma
The stability of compounds II and III was investigated in
0% human plasma pH 7.4. The reactions were initiated by
adding 50 ␮l of a stock solution in acetonitrile to 5 ml of
preheated plasma to give an initial concentration of 10 M.
−
4
◦
The reaction mixtures were kept in a water bath at 37± 0.5 C
and at appropriate time intervals 250 ␮l samples were taken
and deproteinised with 750 ␮l of methanol. The mixture was
immediately vortexed and after centrifugation for 3 min at
−1
(20–40 vol.%). The flow rate was set at 1 ml min and the
column effluent was monitored at 220 nm. Quantitation of
compounds I–III was done from peak height measurements
1
3,000 rpm 20 ␮l of the supernatant was analysed by HPLC
for remaining 4-imidazolidinone.

2
36
S.W. Larsen et al. / European Journal of Pharmaceutical Sciences 20 (2003) 233–240
in relation to those of standards chromatographed under the
same conditions.
100
80
Analysis of degradation products of the decomposition of
◦
compound III was performed at 37 ± 0.5 C in aqueous so-
lution at pH 7.40 and in methanol using a HPLC system
consisting of a Merck-Hitachi L-7100 pump connected to a
Merck-Hitachi L-7400 UV detector. The sampling procedure
was performed automatically by a Merck-Hitachi L-7200 au-
tosampler while the reaction mixtures were kept at constant
temperature using a Merck L-7200 Peltier Sample Cooler.
Isomerisation of compound III was studied using a
LC–MS method consisting of a HP1100 MSD equipped
with binary pump, column compartment, diode array de-
tector and a single quadrupole masspectrometer detector.
Electrospray ionisation was performed in positive ion made
60
4
0
0
0
2
0
100
200
time (min)
300
400
−
4
Fig. 3. Degradation of compound II (10 M) ( ) and concomitant for-
mation of prilocaine ( ) in 0.02 M phosphate buffer solution pH 7.40
◦
at 60 ± 0.5 C and µ = 0.5. Formation of compound II ( ) from sto-
◦
mode. The analysis was performed at 40 C on Waters
−4
chiometric amounts of prilocaine and formaldehyde (10 M) in 0.02 M
Xterra MS C-18 × 3 mm column. The mobile phases con-
◦
phosphate buffer solution pH 7.40 at 60 ± 0.5 C and µ = 0.5.
sisted of mobile phase A: 0.01% TFA and mobile phase B:
0
1
1
.01% TFA in acetonitrile. A linear gradient from 5% B to
00% B over 7.5 min was performed using a flow rate of
containing equimolar amounts of prilocaine and formalde-
−
4
−
1
hyde (10 M). In the latter case a ratio [II]/[I] of 0.83 at
.5 ml min . After the diode array detector the flow was
o
−
1
equilibrium was calculated. In the decomposition studies
divided with 0.5 ml min to the MS. UV detection was
monitored at 210 nm and MS scanning range from 100 to
applying to reversible kinetics the equilibrium concentra-
tion of intact compound II increased as a function of pH
in the pH range 5.4–7.4 possibly reflecting that the reactant
besides formaldehyde in the formation of compound II is
unionised prilocaine. At a few pH values, the formaldehyde
formed from degradation of compound II was trapped by
1
000 amu.
3
. Results and discussion
−2
adding large excess of semicarbazide (10 M) to the buffer
solutions enabling conversion of reversible degradation ki-
netics into apparent first-order kinetics (Fig. 4). The lack of
an induction period in the observed pseudo-first-order plot
in Fig. 4 implies that trapping of formaldehyde was fast rel-
ative to formaldehyde formation. A similar observation has
been reported in stability assessment of other formaldehyde
releasing prodrug types (Bundgaard and Johansen, 1980).
Under these experimental conditions, estimation of the rate
3
.1. Kinetics of hydrolysis
The kinetics of hydrolysis of the 4-imidazolidinones II
and III was studied in aqueous buffer solutions over the
pH range 1–7.4 at 60 and 37 C, respectively. Simultaneous
◦
disappearance of the derivatives and appearance of parent
prilocaine were monitored by HPLC. From kinetic runs em-
ploying buffer concentrations in the range 0.02–0.1 M it was
observed that the decomposition rates were not subject to
significant general acid–base catalysis. Similar degradation
rates of compounds II and III in buffer pH 7.4 and human
plasma were found suggesting that the derivatives, like other
constant k was done from slopes of linear plots of the loga-
f
rithm of intact compound II versus time. The pH dependent
4
-imidazolidinones tested (Bundgaard, 1991), were not sus-
5
ceptible to undergo enzyme catalysed decomposition.
At constant pH and temperature the hydrolysis of com-
pound II to yield prilocaine and formaldehyde displayed
strict first-order kinetics at pH below approximately 5. At
higher pH, the decomposition reactions proceeded to an
equilibrium (Fig. 2). To illustrate the existence of reversible
kinetics at pH above 5, two experiments were run in 0.02 M
phosphate buffer pH 7.40. In the first experiment, compound
II was added to the buffer (initial concentration of 10 M)
and the concentrations of intact compound II and formed
prilocaine were monitored as a function of time. Rela-
tive to [II]0, the equilibrium concentrations of compounds
I and II amounted to 18 and 81%, respectively (Fig. 3).
Likewise, the time courses for the two compounds were de-
termined under identical conditions from a buffer solution
4
.5
4
−
4
3.5
0
100
200
300
time (min)
400
500
600
−4
Fig. 4. Degradation of compound II (10 M) in 0.02 M phosphate buffer
◦
solution pH 7.40 at 60 ± 0.5 C and µ = 0.5 ( ) and in buffer pH 7.40
−
2
containing 10 M semicarbazide ( ).

S.W. Larsen et al. / European Journal of Pharmaceutical Sciences 20 (2003) 233–240
237
Table 1
-
-
-
0.8
1.2
1.6
Observed pseudo-first-order rate constants (kobs or kf ) for hydrolysis
of compound II, calculated pseudo-second-order rate constants (kr) for
formation of compound II and equilibrium constants (K) for the reversible
◦
reactions according to Fig. 2 in aqueous buffer solution at 60 ± 0.5 C
and µ = 0.5
pH
kobs or kf
kr
K
−
1
3
−1 −1
−5
(h
)
(10 M
h
)
(10 M)
2
2
3
3
3
4
4
4
5
5
6
6
6
7
7
.15
0.007
0.011
0.029
0.047
0.066
0.065
0.079
0.104
0.096
0.115
0.110
0.109
0.107
0.100
0.105
-
2
.52
.00
.51
.99
.02
.50
.99
.00
.40
.00
.42
.80
.40
.40
a
-2.4
1
2
3
4
5
6
7
8
pH
a
Fig. 5. pH-rate profile for the hydrolysis of compound II in aqueous buffer
b
◦
solution at 60± 0.5 C and µ = 0.5. (kobs or kf : observed pseudo-first-order
0.91
2.42
5.42
13.5
25.6
12.6
4.54
2.01
0.79
0.39
−
1
rate constant (h )). The full line has been calculated from Eq. (6).
characteristic for the hydrolysis of a number of N-Mannich
base derivatives (Katsura, 1971b; Klixbüll and Bundgaard,
984; Bak et al., 1999). The observed pH dependence of the
apparent first-order rate constants can be accounted for by
assuming spontaneous decomposition of protonated (BH )
a
1
a
−2
Buffer containing 10 M semicarbazide.
Calculated according to pseudo-first-order kinetics.
b
+
and unionised (B) N-Mannich base with the overall rate
equation given by:
pseudo-first-order rate constants (kobs or k ) for hydrolysis
of compound II in aqueous buffer are presented in Table 1.
f
a
+
Ka
H
The almost identical estimates of k at pH 7.40 calculated ac-
kobs = k1
+ k2
a
+
H
(6)
f
a
H
+
+ Ka
+ Ka
cording to reversible kinetics (Eqs. (1) and (2)) or first-order
kinetics indicate that addition of semicarbazide did not affect
the reaction rate significantly. At pH 4 and 5, degradation of
compound II was done in pure buffer and in buffer contain-
where Ka is the ionisation constant of compound II and aH+
refers to the hydrogen ion activity. k1 and k2 are the appar-
ent first-order rate constants for spontaneous degradation
of protonated and unionised N-Mannich base, respectively.
The full line in Fig. 5 was constructed from Eq. (6). With k2
−2
ing 10 M semicarbazide. At both pH values comparable
pseudo-first-order rate constants were obtained suggesting
that at pH below 5 the reverse reaction depicted in Fig. 2
for practical purposes is negligible and that the degradation
of compound II adequately can be described by first-order
kinetics under the experimental conditions employed. The
pH-dependent equilibrium constants, K, presented in Table 1
were calculated from the determined concentrations at
equilibrium of compound II ([II]eq) and prilocaine ([I]eq):
−
1
amounting to 0.108 h the best fit was obtained by using
−
3
−1
−3.7
. The magnitude of
k1 = 4.5 × 10
h
and Ka = 10
the kinetically determined ionisation constant is close to that
−3.89
calculated from potentiometric titration (10
). Thus, pKa
of the 4-imidazolidinone is lowered about 4 units in compar-
ison to that of the amino group in prilocaine. As is apparent
from Fig. 5, reasonable agreement between calculated and
experimentally determined rate constants is observed sug-
gesting that Eq. (6) adequately describes the degradation
kinetics.
2
[
I]
eq
K = [
II]eq
(4)
In comparison to compound II the 4-imidazolidinone
formed from prilocaine and acetaldehyde (compound III)
exhibited enhanced instability in aqueous buffer in the pH
and the apparent second-order rate constants for formation
of compound II from prilocaine and formaldehyde, kr, were
calculated from:
◦
range 1.1–7.4. The decomposition was followed at 37 C
k
f
monitoring the decrease in concentration of intact com-
pound III. The reactions displayed strict first-order kinetics
for about three half-lives. Formation of 4-imidazolidinone
after addition of equimolar amounts of prilocaine and ac-
kr =
(5)
K
The stability of 4-imidazolidinones is influenced by substi-
tution in the ring system (Katsura et al., 1971) and this find-
ing may explain the observed enhanced hydrolytic lability
of compound II compared to that of the N −1 unsubstituted
−
4
etaldehyde (10 M) to the reaction solutions was not ob-
served at acidic pH ensuring lack reversible kinetics in this
pH interval. Although the similar experiment performed
at pH 7.4 revealed formation of a small amount of the
cyclic N-Mannich base the data were treated according to
first-order kinetics. The calculated pseudo-first-order rate
3
-phenyl-4-imidazolidinone (Katsura, 1971b). A plot of the
logarithm of the apparent first-order rate constants (kobs
or k ) for hydrolysis of compound II against pH resulted
f
in a sigmoidal-shaped pH-rate profile (Fig. 5) which is

2
38
S.W. Larsen et al. / European Journal of Pharmaceutical Sciences 20 (2003) 233–240
Table 2
7
6
5
4
3
2
1
00000
00000
00000
00000
00000
00000
00000
0
Observed pseudo-first-order rate constants (kobs) and half-lives (t1/2) for
◦
the degradation of compound III in aqueous buffer solution at 37± 0.5 C
and µ = 0.5
pH
kobs (h 1)
−
t1/2 (h)
1
2
3
3
4
4
5
6
6
7
.14
0.076
0.157
0.273
0.708
1.02
1.42
1.70
1.38
0.573
0.492
9.1
4.4
2.5
1.0
0.7
0.5
0.4
0.5
1.2
1.4
.64
.00
.57
.00
.50
.00
.00
.90
.40
0
20
40
60
80
time (hours)
Fig. 7. Rearrangement of compound III ( ) to compound X ( ) in
methanolic solution at ambient temperature.
constants for degradation of compound III are presented
in Table 2 and the bell-shaped pH-rate profile obtained by
plotting the logarithm of the pseudo-first-order rate con-
stants against pH is depicted in Fig. 6. Whereas the influ-
ence of pH on the hydrolysis rates of N-Mannich bases in
most cases are characterised by S-shaped pH-rate profiles,
bell-shaped profiles for hydrolysis of hetacillin (Tsuji and
Yamana, 1974) and other 4-imidazolidinones derived from
peptides (Klixbüll and Bundgaard, 1984; Rasmussen and
Bundgaard, 1991) have also been reported. The bell-shaped
pH-rate profile for compound III cannot be interpreted in
terms of Eq. (6). Most likely, this shape indicates the in-
volvement of a kinetically significant intermediate in the
reaction pathway and a change of the rate-determining step
in the overall reaction with pH. Like compound III the hy-
drolysis of hetacillin shows a bell-shaped pH-rate profile.
In the interconversion between hetacillin and ampicillin the
existence of a Schiff base intermediate has been reported
the stability studies performed at pH 6.9 and 7.4, however,
product analysis (HPLC) revealed that parallel to forma-
tion of prilocaine an unknown compound (compound X)
emerged. The latter compound eluted with a greater reten-
tion time compared to that of compound III. At pH 7.4 also
compound X was unstable and degraded to yield prilocaine.
Reversed phase chromatography of a freshly prepared solu-
tion of compound III in methanol resulted in the appearance
of only one peak. On storage, in this methanolic solution
compound III rearranged as function of time to give a chem-
ical substance, the retention time of which was identical to
that of compound X formed in the degradation study per-
formed in aqueous buffer pH 7.4 (Fig. 7). LC–MS analysis
revealed that the peaks corresponding to compounds III and
X, respectively, possessed identical molecular masses with
m/z = 247.1 (m+1). In addition, the fragmentation patterns
of the two peaks were identical. These observations together
with the fact that prilocaine is formed from degradation
of compound X at weakly alkaline pH strongly suggest
that unknown compound X constitutes an isomeric form
of compound III. It has previously been reported that the
cyclic N-Mannich base hetacillin in alkaline solution par-
allel to release ampicillin is transformed into epihetacillin,
which is degraded more slowly than hetacillin (Tsuji and
Yamana, 1974). The 4-imidazolidinone ring of the cyclic
N-Mannich base formed from reacting racemic prilocaine
with acetaldehyde contains two asymmetric carbon atoms
(Durbin and Rydon, 1970). The latter type of profile has
been recognised in Schiff base formation and hydrolysis
(Cordes and Jencks, 1962).
Decomposition of compound III at acidic pH was ac-
companied by concomitant formation of prilocaine. For
0.4
0
-
-
-
0.4
0.8
1.2
(
C-2 and C-5) (Fig. 1). Thus, four diastereomeric forms
might exist which in reversed phase chromatography are
expected to elute in the form of two peaks containing the
respective enantiomeric pairs. The fact that compound III in
RP-HPLC appears as only one peak containing one or both
of the species constituting one of the enantiomeric pairs
indicates that in the synthesis the formation of compound
III is kinetically rather than thermodynamically controlled.
For the decomposition of compound III at pH 7.4 the rel-
ative time courses for disappearance and formation of the
1
2
3
4
5
6
7
8
pH
Fig. 6. pH-rate profile for decomposition of compound III in aqueous
buffer solution at 37 ± 0.5 C and µ = 0.5.
◦

S.W. Larsen et al. / European Journal of Pharmaceutical Sciences 20 (2003) 233–240
239
7
6
5
4
3
2
1
00000
00000
00000
00000
00000
00000
00000
0
References
Albert, A., Serjeant, E., 1971. The Determination of Ionisation Constants,
second ed. Chapmann & Hall, London.
Bak, A., Fich, M., Larsen, B.D., Frokjær, S., Friis, G.J., 1999. N-Terminal
4-imidazolidinone prodrugs of Leu-enkephalin: synthesis, chemical
and enzymatic stability studies. Eur. J. Pharm. Sci. 7, 317–323.
Bundgaard, H., 1985. Design of prodrugs: bioreversible derivatives for
various functional groups and chemical entities. In: Bundgaard, H.
(Ed.), Design of Prodrugs. Elsevier, Amsterdam, pp. 1–92.
Bundgaard, H., 1991. Novel chemical approaches in prodrug design.
Drugs Future 16, 443–458.
Bundgaard, H., Johansen, M., 1980. Prodrugs as drug deliv-
ery systems. VIII. Bioreversible derivatization of hydantoins by
N-hydroxymethylation. Int. J. Pharm. 5, 67–77.
0
20
40
60
80
100
time (hours)
Capellos, C., Bielski, B.H., 1972. Kinetic Systems. Wiley/Interscience,
New York, pp. 41–43.
−
4
Fig. 8. Decomposition of compound III (10 M) in 0.02 M phosphate
◦
buffer solution pH 7.40 at 37 ± 0.5 C and µ = 0.5. The time courses
Cordes, E.H., Jencks, W.P., 1962. On the mechanism of Schiff base
formation and hydrolysis. J. Am. Chem. Soc. 84, 832–837.
Durbin, A.K., Rydon, H.N., 1970. The equilibrium between the antibiotics
hetacillin and ampicillin in solution. Chem. Commun., 1249–1250.
Hansch, C., Leo, A., 1979. Substituent Constants for Correlation Analysis
in Chemistry and Biology. Wiley, New York.
for disappearance and formation of various species (compound III ( ),
compound X ( ) and prilocaine ( )).
various species (compounds III, X and prilocaine) are pre-
sented in Fig. 8. Treating the terminal part of the descending
curvature representing degradation of compound X ac-
Hirano, K., Ichihashi, T., Yamada, H., 1982. Studies on the absorption of
practically water-insoluble drugs following injection. V. Subcutaneous
absorption in rats from solutions in water immiscible oils. J. Pharm.
Sci. 71, 495–500.
−
1
cording to first-order kinetics a kobs value of 0.023 h
Katsura, Y., 1971a. Pharmaceutical studies of imidazolidinone derivatives.
was calculated. Thus, at physiological pH an initial, rela-
tively fast regeneration of parent prilocaine from compound
III is to be expected followed by a slower drug activa-
tion resulting from hydrolysis of the isomeric form of
compound III.
1. Stability of 1,2-dimethyl-3-phenyl-4-imidazolidinone. J. Pharm. Soc.
Jpn. 91, 12–18.
Katsura, Y., 1971b. Pharmaceutical studies of imidazolidinone derivatives.
III. Stability of 3-phenyl-4-imidazolidinone. J. Pharm. Soc. Jpn. 91,
26–32.
Katsura, Y., Sugiyama, M., Ono, Y., 1971. Pharmaceutical studies of imi-
dazolidinone derivatives. IV. Consideration on degradation mechanism
of imidazolidinone derivatives. J. Pharm. Soc. Jpn. 91, 33–38.
3
.2. Apparent partition coefficients between Viscoleo®
Klixbüll, U., Bundgaard, H., 1984. Prodrugs as drug delivery systems.
XXX. 4-Imidazolidinones as potential bioreversible derivatives for the
and buffer pH 7.4
␣
-aminoamide moiety in peptides. Int. J. Pharm. 20, 273–284.
Apparent partition coefficients of prilocaine (log Papp
.8) and compound II (log Papp 1.21) between Viscoleo
Kohane, D.S., Lipp, M., Kinney, R.C., Lotan, N., Langer, R., 2000.
Sciatic nerve blockade with lipid–protein–sugar particles containing
bupivacaine. Pharm. Res. 17, 1243–1249.
Larsen, D.B., Fredholt, K., Larsen, C., 2000. Assessment of rate of drug
release from oil vehicle using a rotating dialysis cell. Eur. J. Pharm.
Sci. 11, 223–229.
Larsen, D.B., Fredholt, K., Larsen, C., 2001a. Addition of hydrogen bond
donating excipients to oil solution: effect on in vitro drug release rate
and viscosity. Eur. J. Pharm. Sci. 13, 403–410.
Larsen, S.W., Thomsen, A.E., Rinvar, E., Friis, G.J., Larsen, C., 2001b.
Effect of drug lipophilicity on in vitro release rate from oil vehicles
using nicotinic acid esters as model prodrug derivatives. Int. J. Pharm.
®
0
and aqueous buffer pH 7.4 were obtained from experiments
done in duplicate at ambient temperature. Due to instabil-
ity in aqueous solution, this parameter was not determined
for compound III. However, estimation of lipophilicities of
cyclic N-Mannich base derivatives of prilocaine obtained
by introduction of alkyl substituents at C-2 position of
the imidazolidinone ring might be feasible by using the
hydrophobic substituent constants for the octanol–water
partitioning system (Hansch and Leo, 1979) since com-
parable ␲(CH2) values has been found for oil–water
and octanol–water partitioning systems (Larsen et al.,
216, 83–93.
Larsen, D.B., Joergensen, S., Olsen, N.V., Hansen, S.H., Larsen, C., 2002a.
In vivo release of bupivacaine from subcutaneously administered oily
solutions. Comparison with in vitro release. J. Control. Release 81,
2
001b).
145–154.
Larsen, D.B., Parshad, H., Fredholt, K., Larsen, C., 2002b. Characteristics
of drug substances in oily solutions. Drug release rate, partitioning
and solubility. Int. J. Pharm. 232, 107–117.
Acknowledgements
Löfgren, N., Tegner, C., 1960. Studies on local anesthetics. XX. Synthesis
of some ␣-monoalkylamino-2-methylpropionanilides. A new useful
local anesthetic. Acta Chem. Scand. 14, 486–490.
This study has been supported by The Centre for Drug
Design and Transport. Dr. Bente Froelund, Department of
Medicinal Chemistry, is acknowledged for providing the H
NMR data. Jørgen Olsen is acknowledged for helpful dis-
cussions in the area of LC–MS analysis.
Nelson, S.D., Breck, G.D., Trager, W.F., 1973. In vivo metabolite con-
1
1
3
densations. Formation of N -ethyl-2-methyl-N -(2,6-dimethylphenyl)-
-imidazolidinone from the reaction of a metabolite of alcohol with
a metabolite of lidocaine. J. Med. Chem. 16, 1106–1112.
4

2
40
S.W. Larsen et al. / European Journal of Pharmaceutical Sciences 20 (2003) 233–240
Rasmussen, G.J., Bundgaard, H., 1991. Prodrug of peptides. 10. Pro-
tection of di- and tripeptides against aminopeptidase by formation
of bioreversible 4-imidazolidinone derivatives. Int. J. Pharm. 71, 45–
octanol:buffer partition coefficients and pKa values of clinically used
drugs. Anesth. Analg. 71, 158–170.
Tsuji, A., Yamana, T., 1974. Kinetic approach to the development in
␤-lactam antibiotics. II. Prodrug. (1). Simultaneous determination of
hetacillin and ampicillin, and its application to the stability of hetacillin
in aqueous solution. Chem. Pharm. Bull. 22, 2434–2443.
5
3.
Strichartz, G.R., Sanchez, V., Arthur, R., Chafetz, R., Martin, D.,
990. Fundamental properties of local anesthetics. II. Measured
1