Synthesis of pH-sensitive amphotericin B-poly(ethylene glycol) conjugates and study of their controlled release in vitro

Article abstract of DOI:10.1016/j.bmc.2007.03.083

New intravenous conjugates of amphotericin B (AMB) with poly(ethylene glycols) (PEG) (M = 5000, 10,000, 20,000) have been synthesized and characterised. The intermediate PEGs possess a 1,4-disubstituted benzene ring with aldehyde group at the end of the chain. The benzene ring is connected with PEG at its 4-position (with respect to the aldehyde group) by various functional groups (ether, amide, ester). Reaction of terminal aldehyde group of the substituted PEGs with AMB gave conjugates containing a pH-sensitive imine linkage, which can be presumed to exhibit antimycotic effect at sites with lowered pH value. All types of the conjugates are relatively stable in phosphate buffer at physiological conditions of pH 7.4 (37 °C), less than 5 mol% AMB being split off from them within 24 h. For a model medium of afflicted tissue was used a phosphate buffer (pH 5.5, 37 °C), in which controlled release of AMB from the conjugates takes place. The imine linkage is split to give free AMB with half-lives of 2-45 min. The rate of acid catalysed hydrolysis depends upon substitution of the benzene ring; however, it does not depend on molecular weights of the PEGs used. The conjugates with ester linkage undergo enzymatic splitting in human blood plasma and/or blood serum at pH 7.4 (37 °C) with half-lives of 2-5 h depending on molecular weights of the PEGs used (M = 5000, 10,000, 20,000). At first, the splitting of ester linkage produces the relatively stable pro-drug, that is, 4-carboxybenzylideniminoamphotericin B, which is decomposed to AMB and 4-formylbenzoic acid in a goal-directed manner only at pH 7 (t_1/2 = 2 min, pH 5.5, 37 °C). A goal-directed release of AMB is only achieved by acid catalysed hydrolysis of imine linkage, either from the polymeric conjugate or from the pro-drug released thereof. The LD₅₀ values determined in vivo (mouse) are 20.7 mg/kg and 40.5 mg/kg for the conjugates with ester linkage (M = 10,000 and 5000, respectively), which means that they are ca. 6-11 times less toxic than free AMB.

Full text of DOI:10.1016/j.bmc.2007.03.083

Bioorganic & Medicinal Chemistry 15 (2007) 4069–4076
Synthesis of pH-sensitive amphotericin B–poly(ethylene
glycol) conjugates and study of their controlled release in vitro
a,
b
c
c
*
´
´
Milosˇ Sedlak, Martin Pravda, Frantisek Staud, Lenka Kubicova,
d
d
ˇ
´ˇ
´
Katerina Tycova and Karel Ventura
a
ˇ
Department of Organic Chemistry, University of Pardubice, Cs. legiı´ 565, 532 10 Pardubice, Czech Republic
bDepartment of Inorganic and Organic Chemistry, Faculty of Pharmacy, Charles University, 500 05 Hradec Kralove, Czech Republic
cDepartment of Pharmacology and Toxicology, Faculty of Pharmacy, Charles University, 500 05 Hradec Kralove, Czech Republic
d
ˇ
Department of Analytical Chemistry, University of Pardubice, Cs. legiı´ 565, 532 10 Pardubice, Czech Republic
Received 17 October 2006; revised 26 March 2007; accepted 30 March 2007
Available online 3 April 2007
´ ´
Dedicated to Professor Jaromır Kavalek on the occasion of his 70th birthday.
Abstract—New intravenous conjugates of amphotericin B (AMB) with poly(ethylene glycols) (PEG) (M = 5000, 10,000, 20,000)
have been synthesized and characterised. The intermediate PEGs possess a 1,4-disubstituted benzene ring with aldehyde group at
the end of the chain. The benzene ring is connected with PEG at its 4-position (with respect to the aldehyde group) by various func-
tional groups (ether, amide, ester). Reaction of terminal aldehyde group of the substituted PEGs with AMB gave conjugates con-
taining a pH-sensitive imine linkage, which can be presumed to exhibit antimycotic effect at sites with lowered pH value. All types of
the conjugates are relatively stable in phosphate buffer at physiological conditions of pH 7.4 (37 ꢀC), less than 5 mol% AMB being
split off from them within 24 h. For a model medium of afflicted tissue was used a phosphate buffer (pH 5.5, 37 ꢀC), in which con-
trolled release of AMB from the conjugates takes place. The imine linkage is split to give free AMB with half-lives of 2–45 min. The
rate of acid catalysed hydrolysis depends upon substitution of the benzene ring; however, it does not depend on molecular weights of
the PEGs used. The conjugates with ester linkage undergo enzymatic splitting in human blood plasma and/or blood serum at pH 7.4
(37 ꢀC) with half-lives of 2–5 h depending on molecular weights of the PEGs used (M = 5000, 10,000, 20,000). At first, the splitting
of ester linkage produces the relatively stable pro-drug, that is, 4-carboxybenzylideniminoamphotericin B, which is decomposed to
AMB and 4-formylbenzoic acid in a goal-directed manner only at pH 7 (t1/2 = 2 min, pH 5.5, 37 ꢀC). A goal-directed release of AMB
is only achieved by acid catalysed hydrolysis of imine linkage, either from the polymeric conjugate or from the pro-drug released
thereof. The LD50 values determined in vivo (mouse) are 20.7 mg/kg and 40.5 mg/kg for the conjugates with ester linkage
(M = 10,000 and 5000, respectively), which means that they are ca. 6–11 times less toxic than free AMB.
ꢁ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
systemic fungal infections.2 Amphotericin B was iso-
lated by Trejo and Bennett3 half a century ago from
the Streptomyces nodosus found along the Orinoco
River in Tebladora (Venezuela). However, clinical appli-
cations of AMB are limited by its poor solubility in
water and by possible serious organ toxicity, particu-
larly nephrotoxicity, which appears to be dose
dependent.4 With the aim of increasing the therapeutic
index of AMB, a number of formulations and conju-
gates have been adopted, which enable solubilization
of AMB in aqueous media and also ensure continuous
dosage usually accompanied by lowered toxicity.5 The
clinically applied formulations are non-covalent com-
plexes of the type of amphotericin B lipid complex
(ABLC6), which are colloidal systems composed
of biodegradable phospholipid matrices (e.g.,
The AIDS epidemic, cancer chemotherapy, pervasive
use of antibiotics and immunosuppressive drugs used
for organ transplantation have all contributed to the
emergence of fungal pathogens1 like Candida albicans,
Aspergillus fumigatus, Cryptococcus neoformans, etc.
Polyene macrocyclic antibiotics, particularly amphoteri-
cin B (AMB), have been found to be effective and exten-
sively used drugs in treatment of both pre-systemic and
Keywords: Amphotericin B; Poly(ethylene glycol); Drug polymer
conjugates; Acid-sensitivity; Controlled release in vitro.
*
Corresponding author. Tel.: +420 466 037 015; fax: +420 466 037
068; e-mail: milos.sedlak@upce.cz
0968-0896/$ - see front matter ꢁ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.03.083

´
M. Sedlak et al. / Bioorg. Med. Chem. 15 (2007) 4069–4076
4070
dimyristoylphosphtidylcholine (DMPC)6 or modified
poly (ethylene glycols) [e.g., distearoyl-N-[(monometh-
oxy)poly(ethylene glycol)succinyl)]-phosphatidyletha-
nolamin (DSPE-PEG)]6with amphotericin B. Such lipid
formulations of amphotericin B often significantly in-
crease antifungal activity in vitro as compared with com-
mon amphotericin B.7 Our previous paper8 suggested a
synthesis of a conjugate for intravenous application in
which 50 mol% AMB is bound by covalent carbamate
bond to methoxy poly(ethylene glycol) (mPEG) and the
remaining 50 mol% is bound non-covalently in this conju-
gate. This non-covalent bond is most probably realised by
p–p interaction and hydrogen bonds between the cova-
lently bound AMB and the monomeric AMB in similar
way9 as it is in the aggregated forms of AMB. Our conju-
gate exhibits very good water solubility and an antimy-
cotic spectrum similar to that of the liposomal and
deoxycholate formulations.8 At present, we are finishing
pharmacokinetic studies in vivo of this conjugate on ani-
mal models. Our study8 was continued by the team of
R.B. Greenwald,10 who prepared conjugates of AMB
with poly(ethylene glycol) (PEG) in which AMB is bound
by means of labile carbamate bond that is decomposed in
blood plasma by elimination mechanism to give free
AMB (t1/2 = 1.5–3 h, blood plasma, rat). This labile
bond between PEG and AMB is constructed in such a
way that the first slowest step of release consists in enzy-
matic reaction in which blood hydrolases split the ester
linkage. In the second step, the pro-drug thus formed
undergoes a fast base-catalysed 1,6-benzyl elimination11
giving quinonemethide and carbamic acid of AMB. The
quinonemethide is then immediately hydrated, and the
carbamic acid of AMB is rapidly decomposed to carbon
dioxide and free AMB.10,11 This system of release of active
substance appears to be suitable for goal-directed
distribution of cytostatics containing primary amino
group, namely in the delivery systems that in their poly-
meric carrier contain a bound monoclonal antibody pos-
sessing affinity to specific glycoprotein surface of
tumour-transformed cells.12 However, this principle
cannot be applied generally to all groups of drugs. For
example, in the case of PEG-AMB conjugates this system
of 1,6-benzyl elimination results in relatively fast
splitting off of free AMB directly into the blood circula-
tion, which rapidly increases its concentration in kidneys,
and this is undesirable due to its nephrotoxicity.
The AMB released in blood undergoes subsequent distri-
bution13 between the present lipoproteins (high-density-
lipoprotein (HDL) and low-density-lipoprotein (LDL)).
Previously, it has been found out14 that the increased level
of complexes of AMB with LDL causes nephrotoxic
effects of AMB.
lead to local lowering of pH value as much as to
pH ꢀ5. This piece of knowledge has been utilised re-
cently in constructing very efficient pH-sensitive lipid
formulations (liposomes) of nystatin.16 An analogous
principle was also used for pH-sensitive spacers between
the drug and carrier that enable release of an active drug
from the carrier in a tumour tissue, either slightly acidic
extracellular fluids or, after endocytosis, in endosomes
or lysosomes of cancer cell.17–19 For the pH-sensitive
bond between amino group of mycosamine AMB and
PEG, we have selected imino group, which should be
sufficiently stable at pH values that are usual in blood
and tissues pH 7.4–7.6), for example, in blood plasma
or blood serum. On the other hand, the bond should
be very easy to hydrolyse at lowered pH value. The cho-
sen carriers are poly(ethylene glycols) (mPEG M = 5000
and PEG M = 10,000, 20,000), which can be chemically
substituted very easily20,21 and are non-toxic, non-
immunogenic, non-antigenic, that is, resistant to recog-
nition by the immunity system of organism.22 Figure 1
presents the structures of our selected conjugates:
a-methoxy-x-(4-AMB-iminomethylphenyloxy)poly(eth-
ylene glycol) (1), in which the poly(ethylene glycol) chain
is bound to the benzene ring by means of ether linkage.
This linkage should not be split by usual components of
blood plasma. The 4-position of the benzene ring of the
conjugate carries the pH-sensitive imino group
connecting the conjugate with AMB. The second sug-
gested conjugate, that is, a-methoxy-x-(4-AMB-imi-
nomethylphenylcarboxamino)-poly(ethylene glycol) (2),
contains linkages that are not split in blood plasma,
namely amide group directed to the poly(ethylene gly-
col) chain and pH-sensitive imino group directed to
AMB. The third type of our suggested conjugates in-
cludes
a-methoxy-x-(4-AMB-iminomethylphenylcar-
H
mPEG N
mPEG
CH
O
CH
CH
N-AMB
N-AMB
O
2
1
mPEG O
PEG O
N-AMB
CH
N-AMB
O
O
3a
3b-c
O
O
PEG :
mPEG:
H3C
H3CO
O
M: 10 000 (b), 20 000 (c)
M: 5 000
The aim of the present paper is synthesis and character-
isation of a new type of AMB-PEG conjugates in which
AMB is bound in such a way as to prevent splitting off
of free AMB in blood circulation during intravenous
applications: AMB should be released in a goal-directed
manner only at the site of activity of fungal pathogens.
One of the methods of achieving such goal-directed anti-
mycotic action in such cases is based on the fact that a
number of pathogens, inclusive of fungal agents, induce
such biochemical processes in the afflicted tissue15 that
OH
OH
OH
O
HO
H3C
O
OH OH
OH OH
O
CH3
CO2H
H
CH3
OH
O
O
Amphotericin B (AMB)
NH2
OH
Figure 1. Structures of PEG amphotericin B conjugates 1, 2, 3a–c.

´
M. Sedlak et al. / Bioorg. Med. Chem. 15 (2007) 4069–4076
4071
bonyloxy)poly(ethylene glycol) (3a) and a,x-bis(4-
O
O
CH
mPEG-NH2/DIC
AMB-iminomethylphenylcarbonyloxy)poly(ethylene
glycols) (3b–c), which in contrast to the previous deriv-
atives have ester linkage between PEG chain and 4-imi-
nomethylbenzoic acid that can be split by blood
hydrolases. Apart from synthesis and characterisation
of the suggested conjugates, it was also the aim of this
paper to study the behaviour of the conjugates in blood
plasma and/or blood serum and in phosphate buffers
(pH 7.4; 5.5). The paper also includes determination of
toxicity of the synthesized conjugates on mouse model
in vivo in order to enable evaluation of anticipated
advantages of the above-mentioned conjugates.
CH
mPEGHN
O
HO
O
Scheme 2. Synthesis of a-methoxy-x-(4-formylphenylcarboxami-
no)poly(ethylene glycol).
of integral intensities of hydrogen atoms of poly(ethyl-
ene glycol) chain and hydrogen atoms of aromatic sys-
tem and/or of CH@O or CONH groups showed that
the terminal functional group(s) (OSO2CH3, NH2 or
OH) of poly(ethylene glycol) chain were transformed
practically quantitatively. Infrared spectroscopy re-
vealed typical bands of the functional groups at 3225,
1693 cmꢁ1(CONH), 1701 cmꢁ1 (CHO) and 1735 cmꢁ1
for the ester functional group. The syntheses of conju-
gates 1, 2, 3a–c made use of the reaction25 of the alde-
hyde functional group present in the benzene ring of
substituted PEG with the primary amino group of
amphotericin B in the presence of activated molecular
sieves 4A and trimethyl orthoformate giving imino link-
age (Schiff’s base) (Scheme 3).
2. Results and discussion
2.1. Synthesis of substituted poly(ethylene glycols) and
conjugates
First, we synthesized substituted poly(ethylene glycols)
having a terminal benzene ring with aldehyde group.
The benzene ring is bound to the poly(ethylene glycol)
chain at 4-position with respect to the aldehyde group
by means of various bridges (ether, amide, ester)
(Fig. 1). a-Methoxy-x-(4-formylphenyloxy)poly(ethyl-
ene glycol) (M = 5000) was prepared by nucleophilic
substitution of methanesulfonic group in a-methoxy-x-
methanesuphonylpoly(ethylene glycol)23 with anion of
4-(1,3-dioxolan-2-yl)phenol.24 The reaction was carried
out in melt (65 ꢀC) using a 15-fold excess of 4-(1,3-
dioxolan-2-yl)phenol in the presence of potassium
carbonate as basic catalyst (Scheme 1). Deblockade of
aldehyde functional group was carried out by hydrolysis
in 1 M hydrochloric acid. After extraction with dichlo-
romethane, the crude product containing ca 5% free
4-hydroxybenzaldehyde was twice recrystallized from
propan-2-ol; yield 68 % pure product.
The yields of conjugates varied in the interval of
74–92%. The conjugates were characterised by means
of infrared spectroscopy; the spectra exhibited the fol-
lowing bands: 1735 cmꢁ1 (ester), 1719 cmꢁ1 (lactone),
1638 cmꢁ1 (imine). Using HPLC and GPC, we esti-
mated the purity of conjugates; the content of free
AMB was <1 mol%. Moreover, UV/vis spectroscopy
showed that the AMB content in conjugates 1, 2, 3a cor-
responds to the molar ratio of 1:1 (AMB/PEG) and in
conjugates 3b–c to 2:1 (AMB/PEG). The UV/vis spectra
of all the conjugates measured in water exhibit identical
typical maxima corresponding to the polyene system of
AMB (346, 367, 386, 409 nm). In an independent way,
we also synthesized 4-carboxybenzylideniminoampho-
tericin B (4), that is, the pro-drug formed by splitting
of ester linkage of conjugates 3 by hydrolases of blood
plasma and/or blood serum (Scheme 4). 4-Carbo-
xybenzylideniminoamphotericin B (4) was prepared by
reaction of AMB with 4-formylbenzoic acid in dimethyl-
sulfoxide at similar conditions to those used for conju-
gates 1–3; composition of the product was verified by
means of ESI-MS, showing the molecular peak
[M+H]+ m/z 1057.2. IR spectrum showed typical bands
a-Methoxy-x-(4-formylphenylcarboxamino)poly(ethyl-
ene glycol) (M = 5000) was prepared by reaction of
a-amino-x-methoxypoly(ethylene glycol) with 4-formyl-
benzoic acid, which was pre-activated with diisopropyl-
carbodimide (DIC); yield of reaction 86% (Scheme 2).
a-Methoxy-x-(4-formylphenyloxy)poly(ethylene glycol)
(M = 5000) and a,x-bis(4-formylphenylcarbonyloxy)-
poly(ethylene glycol)25 (M = 10,000) or a,x-bis(4-form-
of lactone at 1719 cmꢁ1 and imine at 1638 cmꢁ1
.
ylphenylcarbonyloxy)poly(ethylene
glycol)25
(M =
20,000) were prepared similarly in accordance with the
literature.25 The synthesized starting substituted PEGs
were characterised by means of H NMR. Comparison
2.2. Study of hydrolytic behaviour of conjugates in phosphate
buffers, human blood plasma and human blood serum
1
A key part of the research consisted in investigation of
behaviour of synthesized conjugates in various media
O
1) K2CO3
O
mPEGO
CH
HO
CH
O
O
2) mPEG-OSO2CH3
AMB/ HC(OCH3)3
(m)PEG X
CH
HCl/H2O
1, 2, 3a-c
O
molecular sieves 4A.
mPEGO
CH
O
C
O
C
O
O
X:
,
NH
,
O
Scheme 1. Synthesis of a-methoxy-x-(4-formylphenyloxy)poly(ethyl-
ene glycol).
Scheme 3. Synthesis of PEG amphotericin B conjugates 1, 2, 3a–c.

´
M. Sedlak et al. / Bioorg. Med. Chem. 15 (2007) 4069–4076
4072
enzymatic hydrolysis of ester bond
(human blood plasma or serum)
acid-catalyzed hydrolysis
(m)PEG X
CH
N-AMB
t1/2 = 2 – 45 min
t1/2 = 2 – 5 h
1 – 3a-c
aromatic substitution
and pH control
molecular weight of PEG control
O
CH
(m)PEG X
CH
O
N-AMB
4
O
acid-catalyzed hydrolysis
t1/2 = 2 min
Amphotericin B (AMB)
Scheme 4. Survey of controlled release of AMB from conjugates.
imitating conditions that control the stability of solu-
tions of the conjugates and their behaviour in organism.
First, we studied stability of all the conjugates 1–3a–c in
phosphate buffers and then in the medium of human
blood plasma and/or human blood serum. On the basis
of monitoring of time changes in UV/vis spectra, we
found out that in solutions of phosphate buffer at the
value of pH 7.4, (7 · 10ꢁ2 M, 37 ꢀC) all types of conju-
gates are relatively stable, very slow hydrolysis being
only observed in the case of imino linkage: less than
5 mol% AMB is split off from the conjugates during
24 h. All the conjugates were constructed in such a
way as to release AMB in goal-directed manner only
at spots with pathologically lowered pH value. The
model medium imitating the afflicted tissue (character-
ised by the lowered pH value) was a phosphate buffer
solution26 (pH 5.5; 7 · 10ꢁ2 M). During the hydrolysis
of conjugate 3a, we observed absorbance decrease at
346 nm (corresponding to concentration drop of conju-
gate) and absorbance increase at 367, 386 and 409 nm.
The absorption band at 409 nm is specific9 for mono-
meric AMB, and its time increase corresponds to the
rate of its formation (Fig. 2).
Figure 2. Time changes of spectrum of conjugate 3a in phosphate
buffer (pH 5.5; 7 · 10ꢁ2 M, 37 ꢀC). The inset represents the absor-
bance–time dependence during hydrolysis of imino linkage in 3a at
409 nm (t1/2 = (124 8) s).
Similar spectral record was also obtained during moni-
toring of hydrolysis of the other conjugates 1–3b–c.
Conjugates 2, 3a–c in phosphate buffer medium (pH
5.5; 7 · 10ꢁ2 M) undergo a relatively fast splitting of
imino linkage giving free AMB, the half-lives of individ-
ual conjugates being as follows: 2: t1/2 = (1.8 0.2) min;
3a: t1/2 = (2.1 0.1) min. We also found out that the
hydrolysable linkages, namely ester and imine linkage
(Fig. 1). If the same conditions (pH 5.5, 7 · 10ꢁ2 M)
were adopted when monitoring the stability of ester of
methoxy-x-(4-formylphenyloxy)poly(ethylene glycol), it
was found that the hydrolysis of ester linkage proceeds
immeasurably slowly in this case, which confirmed the
fact that in the phosphate buffer medium the splitting
of conjugates exclusively concerns their imino linkage
and gives AMB.
acid-catalysed hydrolyses of conjugates 3b: t1/2
=
(2.4 0.2) min and 3c: t1/2 = (2.2 0.4) min proceed at
comparable rates, which means that the molecular
weight of poly(ethylene glycol) carrier has virtually
no effect on these rates. Conjugate 1 with different
substitution in benzene ring exhibits a rate decrease
of acid-catalysed hydrolysis of imino linkage
(t1/2 = (45 1) min). The rate difference of hydrolyses
of conjugate 1 as compared with conjugates 2, 3a–c
corresponds to the effect of substituent at 4-position of
benzene ring upon the acid-catalysed hydrolysis of imi-
nes.27 Conjugates 1 and 2 are hydrolysed at measurable
rates only in the imine moiety, the difference between
rates of their acid-catalysed hydrolyses being approxi-
mately one order of magnitude. Conjugates 3a–c, in
contrast to conjugates 1 and 2, contain two types of
Furthermore, using HPLC, we monitored the stability
and/or decomposition of all the conjugates in media of
human blood plasma or human blood serum. Results
of the kinetic measurements indicate that conjugates 1
and 2 are relatively stable in both blood serum and
blood plasma (pH 7.4), like in phosphate buffer (pH
7.4). Conjugates 3a–c in blood serum and/or blood plas-
ma undergo enzymatic hydrolysis of their ester linkage
by action of blood hydrolases, which is indicated by
the decrease in concentration of the starting conjugate.

´
M. Sedlak et al. / Bioorg. Med. Chem. 15 (2007) 4069–4076
4073
The kinetic dependence shown in Figure 3 indicates
that the results of kinetic measurements of hydrolysis
of conjugate 3a in blood plasma and in blood serum
were mutually comparable within experimental error.
However, the determination of concentration of the
conjugates in blood plasma was loaded with larger er-
ror and worse reproducibility (Fig. 3). From the
standpoint of analysis, it appeared more appropriate
to monitor the concentration decrease of starting con-
jugates than to follow the AMB increases, since AMB
interacts with the present lipoproteins,28 which results
in concentration changes of free AMB that are not
connected with the kinetic course of decrease in the
starting conjugates.
The kinetic dependence shown in Figure 3 also allows
calculation of half-life (2.1 0.1 h) of enzymatic hydro-
lysis of conjugate 3a in human blood serum. The half-
life found by us is comparable with literature data,10,12
reporting splitting of ester linkages in poly(ethylene
glycol) (M = 5000) in rat blood serum with the half-life
values of 1–3 h. However, in our case the splitting in
blood plasma or serum produces the relatively stable
pro-drug 4, that is, AMB with bound 4-formylbenzoic
acid, which can subsequently release AMB only at the
conditions of lowered pH value, that is, at the spot of
damaged tissue. Our kinetic measurements confirmed
(phosphate buffer, pH 5.5, 7 · 10ꢁ2 M, 37 ꢀC) that the
half-life of acid-catalysed hydrolysis of pro-drug 4
(t1/2 = (2.15 0.01) min) is comparable with the half-
lives of acid-catalysed hydrolyses of conjugates 3a–c
(Scheme 4). Moreover, we found out that the half-lives
of enzymatic hydrolyses of ester linkage by blood hydro-
lases in plasma and/or serum depend on molecular
weights of individual conjugates: 3a: t1/2 = (2.1
0.2) h; 3b: t1/2 = (3.0 0.3) h; 3b: t1/2 = (5 0.2) h;
(PEG: M = 5000, 10,000, 20,000) (Figs. 3 and 4).
Figure 4. Concentration–time dependence of conjugate 3b (d)
(t1/2 = (3.0 0.3) h) and 3c (s) (t1/2 = (5.0 0.3) h) in human blood
serum (pH 7.4) at 37 ꢀC.
4 (Scheme 4). Application of PEGs with higher molecu-
lar weight results in retardation of release of pro-drug 4
and can be favourable from the point of view of prolon-
gation of the drug effect, not only for the reason of gen-
erally known12 slower excretion of PEGs with higher
molecular weight by kidneys.
2.3. Study of toxicity of conjugates 3a and 3b
Acute toxicity was tested for AMB as a control and for
two conjugates (3a and 3b).
Mice (NMRI, weight 25–30 g) were purchased from
Biotest Ltd. Konarovice, Czech Republic. For each
compound tested, the mice were divided into 6 groups
of 10 males, and the compound was administered intra-
venously via the tail vein at doses carefully determined
after a series of preliminary trials. Most deaths occurred
5 min to 24 h after iv administration of the tested
compound. When compared to control, PEGylation
increased LD50 11 times in the case of 3a conjugate
and about five times in the case of 3b conjugate (see
Table 1).
From these results it is obvious that molecular weight of
carrier controls the rate of enzymatic release of pro-drug
Both the toxicities of conjugates determined in vivo and
the results of hydrolytic studies in vitro agree with the
presumption that our suggested system of two-stage
independent release of AMB minimizes the level of free
AMB and its complexes with LDH in blood circulation
and thereby reduces the risk of toxic damage to the pa-
tient.14 In conclusion, it can be stated that the conju-
gates prepared by us represent promising candidates
for new intravenous forms of AMB suitable for further
tests on animal model.
Table 1. Effect of amphotericin PEGylation on acute iv toxicity in
mice
Compound
Control
(AMB)
3a
3b
Figure 3. Concentration–time dependence of conjugate 3a in human
blood plasma (d) (pH 7.4) and in human blood serum (s) (pH 7.4) at
37 ꢀC. The conjugate decrease represents the enzymatic hydrolysis of
ester linkage by blood hydrolases (t1/2 = 2 h).
LD50 mouse iv (mg/kg)
95% confidence intervals
(mg/kg)
3.7
3.2–4.2
40.5
38.1–43.1
20.7
15.4–32.3

´
M. Sedlak et al. / Bioorg. Med. Chem. 15 (2007) 4069–4076
4074
3. Experimental
100 RP-18e 5 lm (Merck) and eluted with mobile phase
of acetonitrile with 20 mM chelaton II. The measured
time dependences of the analyser responses (mV) were
used to calculate the pseudo-first-order observed rate
constants (kobs (hꢁ1)) similarly as in above cases.
3.1. Materials
Unless otherwise stated, all the poly(ethylene glycols),
amphotericin B, other chemicals and solvents were ob-
tained from Fluka or Aldrich and used without further
purification.
3.3. Synthesis of a-methoxy-x-(4-formylphenyloxy)poly(eth-
ylene glycol)
3.2. Analytical methods
A melt of 5 g (1 mmol) a-methoxy-x-methanesuphonyl-
poly(ethylene glycol),23 2.5 g (15 mmol) 4-(1,3-dioxolan-
2-yl)phenol25 and 2 g (15 mmol) potassium carbonate
was stirred and heated under argon atmosphere at
65 ꢀC for 48 h. After cooling, the reaction mixture was
dissolved in 200 ml dichloromethane, filtered with Celite
and evaporated on a water bath. The dry evaporation
residue was dissolved in 150 ml of 1 M hydrochloric acid
and refluxed for 4 h. After cooling, the solution was sat-
urated with sodium chloride and extracted with
5 · 60 ml dichloromethane. The combined extracts were
dried and evaporated, and the residue was twice recrys-
tallized from propan-2-ol. Yield 3.4 g product (68%). 1H
NMR (CDCl3, ppm) d: 3.36 (s; 3H, CH3-O); 3.52-4.27
(m; 460H, (CH2–CH2–O)115); 7.17 (d; J 7.7 Hz, 2H-
arom); 7.97 (d; J 7.7 Hz, 2H-arom); 9.94 (s; 1H,
CH@O). IR (KBr) m: 3439, 2945, 2887, 2741, 1701,
1467, 1360, 1343, 1280, 1201, 1149, 1110, 1061, 963,
1
3.2.1. The H NMR spectra were measured in deuterio-
chloroform (CDCl3), on a Bruker AMX 360 apparatus
and a Bruker 500 Avance apparatus.
3.2.2. Positive-ion electron-spray ionisation (ESI) mass
spectra were measured on an Esquire 3000 ion trap ana-
lyzer (Bruker Daltonics, Bremen, Germany) in the m/z
range of 50–1500.
3.2.3. The IR spectra were measured using Nicolet
Impact 400 spectrometer. The samples (0.2 mg) were
measured in KBr pellets (200 mg).
3.2.4. Gel permeation chromatography measurements of
the prepared conjugates 1–3 were performed with
HEMA–BIO columns (hydrophilic modified HEMA–
Gel, particle size 10 lm, porosity 40/100/300/1000) at
room temperature using RI detector and UV/vis detec-
tor. Redistilled water (pH 7.1) was used as the eluent.
The columns were calibrated with a series of standard
PEGs with varying molecular weights (PSS, Polymer
Standard Service GmbH, Mainz, Germany).
843, 531 cmꢁ1
.
3.4. Synthesis of a-methoxy-x-(4-formylphenylcarbox-
amino)poly(ethylene glycol)
a-Amino-x-methoxypoly(ethylene glycol) (5 g; 1 mmol),21
0.6 g (4 mmol) 4-formylbenzoic acid and 0.05 g
(0.4 mmol) 4-dimethylaminopyridine were dissolved in
50 ml dry dichloromethane, and the obtained solution
was cooled to 0 ꢀC. The reaction mixture was treated
with cold solution of 0.5 g (4 mmol) diisopropylcarbo-
diimide in 10 ml dichloromethane added drop by drop
during 10 min. Then the reaction mixture was stirred
at 0 ꢀC for another 4 h, whereupon it was slowly
warmed to room temperature. After 24 h, diisopropyl-
urea was filtered off and the filtrate was evaporated;
the residue was recrystallized from propan-2-ol. Yield
3.2.5. Study of acid-catalysed hydrolysis of imine link-
age in compounds 1–3a–c, 4 in phosphate buffers. The
kinetic measurements were performed on an HP UV/
vis 8453 Diode Array apparatus in 1 cm closable quartz
cell kept at a temperature of 37 ꢀC. The cell was charged
with 2 ml phosphate buffer (pH 7.4, 7 · 10ꢁ2 M or pH
5.5, 7 · 10ꢁ2 M), and after reaching the above-men-
tioned temperature, 10-20 ll methanolic solutions of
conjugates 1–3a–c or pro-drug 4 were injected to make
the final concentration of substrate ca 5 · 10ꢁ5 M. The
measured time dependences of absorbance (A) were used
to calculate the pseudo-first-order observed rate
constants (kobs (sꢁ1)) using the equation: kobst = ln
1
4.3 g product (86%). H NMR (CDCl3, ppm) d: 3.36
(s; 3H, CH3-O); 3.52–4.27 (m; 460H, (CH2-CH2-O)115);
4.51 (m; 2H, CH2–NH–CO); 6.48 (m; H, NH–CO);
7.96 (d; J 7.8 Hz, 2H-arom); 8.22 (d; J 7.8 Hz, 2H-
arom); 10.11 (s; 1H, CH@O). IR (KBr) m: 3439, 3225,
2945, 2887, 2741, 1701, 1693, 1467, 1360, 1343, 1280,
DA + const., where DA = (A1 ꢁ At) or (At ꢁ A ).
1
3.2.6. Study of enzymatic hydrolysis of ester linkage of
conjugates 3a–b in human blood plasma and human
blood serum by means of HPLC. A micro-test-tube
was charged with 500 ll plasma or serum sample (ob-
tained from healthy donors of plasma at transfusion sta-
tion) and 500 ll aqueous solution of conjugates 3a–c to
make the final conjugate concentration of 1 · 10ꢁ1M.
The samples were incubated at 37 ꢀC. After the incuba-
tion, they were gradually (at definite time intervals: 0
(immediately), 1, 2 to 10 h and after 24 h) stabilized with
100 ll of 4% ZnSO4Æ7H2O. The mixture was shaken for
1 min and then centrifuged at 14000 rpm for 3 min. The
supernatant (20 ll) was then introduced on a LiChro-
CARTꢂ 125 · 4 mm column packed with LiChrospherꢂ
1201, 1149, 1110, 1061, 963, 843, 531 cmꢁ1
.
3.5. General procedure of synthesis of conjugates 1, 2, 3a–c
a-Methoxy-x-(4-formylphenyloxy)poly(ethylene glycol)
(0.1 mmol; M = 5000), a-methoxy-x-(4-formylphenyl-
carboxamino)poly(ethylene glycol) (0.1 mmol; M = 5000),
a-methoxy-4-(formylphenylcarbonyloxy)poly(ethylene
glycol)25 (0.1 mmol; M = 5000), a,x-bis(4-formylphenyl-
carbonyloxy)poly(ethylene glycol)25 (0.1 mmol; M =
10,000), or a,x-bis(4-formylphenylcarbonyloxy)poly-
(ethylene glycol)25 (0.1 mmol; M = 20,000) was dissolved
in 5–10 ml trimethyl orthoformate, and the solution

´
M. Sedlak et al. / Bioorg. Med. Chem. 15 (2007) 4069–4076
4075
formed was treated with an equivalent amount (0.1 or
0.2 mmol) of saturated solution of amphotericin B in
dimethylformamide with addition of 0.2 g activated
molecular sieves 4A. The reaction mixture was stirred un-
der argon atmosphere at room temperature for several
days with exclusion of light. The end of reaction was
established for the individual cases by means of GPC. Fi-
nally, the reaction mixture was filtered, and the filtrate
was evaporated in vacuum to remove trimethyl orthofor-
mate. The distillation residue was poured in 250 ml dry
diethyl ether, and the precipitated product was collected
by filtration. The conjugates prepared were dissolved in
dichloromethane and reprecipitated with diethyl ether.
derivatives (3a, 3b). For each compound tested, the
mice were divided into 6 groups of 10 males, and the
compound was administered intravenously via the tail
vein at doses carefully determined after a series of pre-
liminary trials. Toxic signs and mortality were moni-
tored for 14 days after administration. LD50 values
and the corresponding confidence limits were deter-
mined by non-linear regression using GraphPad Prism
(version 4; GraphPad software, Inc., San Diego CA,
USA).
4. Conclusion
3.5.1. a-Methoxy-x-(4-AMB-iminomethylphenyloxy)poly-
(ethylene glycol) (1). Yield 450 mg (74%). IR (KBr) m:
3428, 2945, 2885, 2741, 2695, 1719, 1637, 1467, 1360,
In contrast to the previous idea of covalent-noncovalent
conjugates and conjugates based on 1,6-benzyl elimina-
tion, which release free AMB in blood circulation, we
have suggested such conjugates that selectively release
AMB only at the site of expected operation of fungal
pathogen, that is, tissue with lowered pH value. The
goal-directed release can only take place by acid-cataly-
sed hydrolysis of imino linkage, either in the polymeric
conjugate or in the liberated pro-drug 4, which is split
off enzymatically from the carrier by action of blood
hydrolases. The molecular weight of carrier controls
the rate of enzymatic liberation of pro-drug 4. From
the LD50 values determined in vivo on mouse model
for conjugates 3b and 3a it follows that they are ca 6–
11 times less toxic than free AMB. The above-described
system of release of AMB from conjugates 3 minimizes
the level of free AMB and hence also that of its com-
plexes with LDH in blood circulation, which results in
lowered toxicity of the conjugates as compared with free
AMB. Our suggested conjugates 3 of amphotericin B
with poly(ethylene glycols) represent a new system of
two-stage independent release of AMB from polymeric
carrier. However, its appropriateness for practical appli-
cations will need further studies on animal models and
their critical evaluation.
1343, 1280, 1201, 1149, 1110, 1061, 963, 843 cmꢁ1
.
3.5.2. a-Methoxy-x-(4-AMB- iminomethylphenylcarbox-
amino)poly(ethylene glycol) (2). Yield 500 mg (54%). IR
(KBr) m: 3439, 3225, 2945, 2887, 2741, 1719, 1693,
1638, 1467, 1360, 1343, 1280, 1201, 1149, 1110, 1061,
963, 843, 531 cmꢁ1
.
3.5.3. a-Methoxy-x-(4-AMB- iminomethylphenylcarbon-
yloxy)poly(ethylene glycol) (3a). Yield 510 mg (86%). IR
(KBr) m: 3439, 2945, 2887, 2741, 1735, 1719, 1638, 1467,
1360, 1343, 1280, 1201, 1149, 1110, 1061, 963, 843,
531 cmꢁ1
.
3.5.4. a,x-Bis(4-AMB- iminomethylphenylcarbonyloxy)-
poly(ethylene glycol) (3b). Yield 980 mg (83%). IR (KBr)
m: 3439, 2945, 2887, 2741, 1735, 1719, 1693, 1638, 1467,
1360, 1343, 1280, 1201, 1149, 1110, 1061, 963, 843,
531 cmꢁ1
.
3.5.5. a,x-Bis(4-AMB- iminomethylphenylcarbonyloxy)-
poly(ethylene glycol) (3c). Yield 2 g ( 92%). IR (KBr) m:
3439, 2945, 2887, 2741, 1735, 1719, 1693, 1638, 1467,
1360, 1343, 1280, 1201, 1149, 1110, 1061, 963, 843,
531 cmꢁ1
.
Acknowledgments
3.6. Synthesis of 4-carboxybenzylideniminoamphotericin
B (4)
The authors acknowledge the financial support from the
MSM 002 162 7501 and the Science Foundation of the
Czech Republic, Grant No. 203/06/0583.
Amphotericin B (93 mg; 0.1 mmol) and 15 mg (0.1 mmol)
4-formylbenzoic acid were dissolved in 2 ml dry dimethyl-
sulfoxide and 3 ml trimethyl orthoformate with addition
of 0.1 g activated molecular sieves 4A. After 48 h
(25 ꢀC), the reaction mixture was filtered, and the filtrate
was distilled in vacuum at the temperature of 30 ꢀC to re-
move the trimethyl orthoformate. After addition of
400 ml dry diethyl ether, the obtained solid was collected
by filtration, washed and dried in vacuum desiccator.
Yield 90 mg (84%). IR (KBr) m: 3439, 3097, 2838, 2679,
2560, 2345, 1719, 1638, 1610, 1575, 1560, 1505, 1430,
References and notes
1. Diamond, R. D. Rev. Infect. Dis. 1991, 13, 480.
2. Ceregheti, D. M.; Carreira, E. M. Synthesis 2006, 914.
3. Trejo, W.; Bennett, R. J. Bacteriol. 1962, 85, 436.
4. Heidemann, H. T.; Gerkens, J. F.; Jackson, E. K.; Branch,
R. A. Naunyn-Schmiedeberg’s Arch. Pharmacol. 1983, 324,
148.
5. Janoff, A. S.; Boni, L. T.; Poescu, M. C.; Minchey, S. R.;
Cullis, P. R.; Madden, T. D.; Tarashi, T. F.; Gruner, S. M.;
Shyamsunder, E. Proc. Natl. Acad. Sci. U.S.A. 1988, 85,
61227.
1393, 1293, 1125, 934, 885, 790, 762, 686, 532, 461 cmꢁ1
;
ESI-MS (M = 1056.2): [M+H]+ m/z 1057.2.
3.7. Acute toxicity
6. Janoff, A. S.; Perkins, W. R.; Saleton, S. L.; Swenson, C.
E. J. Liposome Res. 1993, 3, 451.
Acute toxicity was tested for amphotericin B (Bristol-
Myers Squibb) as a control and for two PEGylated
7. Moribe, K.; Mararuyama, K.; Iwatsuru, M. Int. J. Pharm.
2000, 201, 37.

´
M. Sedlak et al. / Bioorg. Med. Chem. 15 (2007) 4069–4076
4076
ˇ
8. Sedlak, M.; Buchta, V.; Kubicova, L.; Simunek, P.;
´
´
˚
Holcapek, M.; Kasparova, P. Bioorg. Med. Chem. Lett.
´ ´
18. Ulbrich, K.; Etrych, T.; Chytil, P.; Pechar, M.; Jelınkova,
ˇ
M.; Rıhova, B. Int. J. Pharm. 2004, 277, 63.
ˇ
2001, 11, 2833.
ˇ
´
´
19. Rodrigues, P. C. A.; Roth, T.; Fiebig, H. H.; Unger, C.;
´
´
9. Gaboriau, F.; Cheron, M.; Leroy, L.; Bolard, J. Biophys.
Chem. 1997, 66, 1.
Mulhaupt, R.; Kratz, F. Bioorg. Med. Chem. 2006, 14, 4110.
¨
´
20. Sedlak, M.; Antonietti, M.; Co¨lfen, H. Macromol. Chem.
Phys. 1998, 199, 247.
10. Conover, C. D.; Zhao, H.; Clifford, B.; Longley, C. B.;
Shum, K. L.; Greenwald, R. B. Bioconjugate Chem. 2003,
14, 661.
11. Wakselman, M. Nouv. J. Chim. 1983, 7, 439.
12. Grrenwald, R. B.; Choe, Y. H.; McGurie, J.; Conover, C.
D. Adv. Drug Delivery Rev. 2003, 55, 217.
13. Sivak, O.; Lau, B.; Patankar, N.; Wasan, K. M. Pharm.
Res. 2004, 21, 2336.
14. (a) Clements, J. S.; Peacock, J. E. Am. J. Med. 1990, 88,
22; (b) Wasan, K. M.; Wong, J. C.; Corr, T.; Pritcharrd, S.
Cancer Chemother. Pharmacol. 2006, 57, 120.
15. Mellman, I.; Fuchs, R.; Helenius, A. Ann. Rev. Biochem.
1986, 55, 773.
´
21. Sedlak, M. Collect. Czech. Chem. Commun. 2005, 70, 269.
22. Harris, J. M. In Poly(ethylene glycol) Chemistry: Bio-
technical and Biomedical Application; Harris, J. M., Ed.;
Plenum Press: New York, 1992; pp 1–13.
23. Sieber, F.; Ventworth, P.; Janda, K. D. Molecules 2000, 5,
1018.
24. Gopinath, R.; Haque, S. J.; Patel, B. K. J. Org. Chem.
2002, 67, 5842.
25. Shou, W. G.; Yang, Y. Y.; Wang, Y. G. Synthesis 2005, 530.
26. Diwu, Z.; Chen, C. S.; Stass, S. A.; Mixon, A. G. Chem.
Biol. 1999, 6, 411.
27. Cordes, E. H.; Jencks, W. P. J. Am. Chem. Soc. 1963, 85,
2843.
28. Bekersky, I.; Fielding, R. M.; Dressler, D. E.; Lee, J. W.;
Buell, D. N.; Walsh, T. J. Antimicrob. Agents Chemother.
2002, 46, 834.
16. Nasti, T. H.; Khan, M. A.; Owais, M. J. Antimicrob.
Chemother. 2006, 57, 349.
ˇ
17. Ulbrich, K.; Subr, V. Adv. Drug Delivery Rev. 2004, 56,
1023.