Synthesis of poly(vinyl alcohol)-doxorubicin conjugates containing cis-aconityl acid-cleavable bond and its isomer dependent doxorubicin release

Article abstract of DOI:10.1248/bpb.31.103

Aconityl-doxorubicin (ADOX) was synthesized by the modified method of Shen and Ryser. Two isomers of cis-ADOX (cis-configuration) and trans-ADOX (trans-configuration) were generated in the reaction of DOX and cis-aconitic anhydride. These products were separated completely by using HPLC and analyzed by TOF-MS spectroscopy and ¹H- and ¹³C-NMR experiments. The yields of cis-ADOX and trans-ADOX were 36.3 and 44.8%, respectively. The free γ-carboxylic group of ADOX molecule was coupled to poly(vinyl alcohol) (PVA) via ethylenediamine spacer, resulting the macromolecular conjugates of PVA-cis-ADOX and PVA-trans-ADOX, respectively. The DOX content of the conjugates estimated by the hydrolysis method detected the aglycone of DOX which can be estimated as the PVA-bound DOX selectively was 4.4 w/w% which was similar to 4.6 w/w% by the ordinary UV method. Both PVA-cis-ADOX and PVA-trans-ADOX were very stable at neutral pH, but the release of DOX was increased markedly under acidic conditions. Half-life of the release of DOX from PVA-cis-ADOX at pH 5.0 was 3 h which was 4.7-fold shorter than that from PVA-trans-ADOX (14 h). The cytotoxicities of PVA-cis-ADOX and PVA-trans-ADOX were evaluated by using J774.1 cells employing a [³H]uridine incorporation assay as a measure of RNA synthesis. A significant difference in antitumor activity between PVA-cis-ADOX and PVA-trans-ADOX was observed where the former was much active than the later. It was suggested that the conjugate enters the cells and reaches the lysosomal/endosomal compartment, and that the aconityl spacer releases DOX from the conjugate in the acidic compartment of lysosomes/endosomes due to the participation of a free carboxylic group.

Full text of DOI:10.1248/bpb.31.103

January 2008
Biol. Pharm. Bull. 31(1) 103—110 (2008)
103
Synthesis of Poly(vinyl alcohol)–Doxorubicin Conjugates Containing
cis-Aconityl Acid-Cleavable Bond and Its Isomer Dependent
Doxorubicin Release
b
Atsufumi KAKINOKI,a Yoshiharu KANEO,*,a Yuka IKEDA,a Tetsuro TANAKA,a and Kahee FUJITA
a Laboratory of Biopharmaceutics, Faculty of Pharmacy and Pharmaceutical Sciences, Fukuyama University; Fukuyama,
Hiroshima 729–0292, Japan: and b Department of Molecular Medicinal Sciences, Graduate School of Biomedical
Sciences, Nagasaki University; Nagasaki 852–8521, Japan. Received August 20, 2007; accepted October 11, 2007
Aconityl-doxorubicin (ADOX) was synthesized by the modified method of Shen and Ryser. Two isomers of
cis-ADOX (cis-configuration) and trans-ADOX (trans-configuration) were generated in the reaction of DOX and
cis-aconitic anhydride. These products were separated completely by using HPLC and analyzed by TOF-MS
spectroscopy and 1H- and 13C-NMR experiments. The yields of cis-ADOX and trans-ADOX were 36.3 and 44.8%,
respectively. The free g-carboxylic group of ADOX molecule was coupled to poly(vinyl alcohol) (PVA) via ethyl-
enediamine spacer, resulting the macromolecular conjugates of PVA–cis-ADOX and PVA–trans-ADOX, respec-
tively. The DOX content of the conjugates estimated by the hydrolysis method detected the aglycone of DOX
which can be estimated as the PVA-bound DOX selectively was 4.4 w/w% which was similar to 4.6 w/w% by the
ordinary UV method. Both PVA–cis-ADOX and PVA–trans-ADOX were very stable at neutral pH, but the release
of DOX was increased markedly under acidic conditions. Half-life of the release of DOX from PVA–cis-ADOX at
pH 5.0 was 3 h which was 4.7-fold shorter than that from PVA–trans-ADOX (14 h). The cytotoxicities of PVA–cis-
ADOX and PVA–trans-ADOX were evaluated by using J774.1 cells employing a [3H]uridine incorporation assay
as a measure of RNA synthesis. A significant difference in antitumor activity between PVA–cis-ADOX and
PVA–trans-ADOX was observed where the former was much active than the later. It was suggested that the con-
jugate enters the cells and reaches the lysosomal/endosomal compartment, and that the aconityl spacer releases
DOX from the conjugate in the acidic compartment of lysosomes/endosomes due to the participation of a free
carboxylic group.
Key words poly(vinyl alcohol); doxorubicin; acid-sensitive spacer; macromolecular prodrug; cytotoxicity; cellular uptake
Doxorubicin (DOX) is effective antineoplastic agent and were performed to test the pharmacologic activity of these
widely used in the therapy of cancer.1—3) It is used clinically conjugates in vivo against L1210 leukemia in DBA2 mice.
to treat a variety of tumors, but this anthracycline drug is
Poly(vinyl alcohol) (PVA) is a polymer which is synthe-
limited by its toxic dose-related side effect, such as cumula- sized by polymerizing not a vinyl alcohol monomer but a
tive cardiotoxicity, myelosuppression, nephrotoxicity, and ex- vinyl acetate monomer. This monomer is polymerized in to
travasation.4)
poly(vinyl acetate) and then hydrolyzed to produce PVA.
The consequence of attachment of low molecular weight PVA’s biocompatibility makes it an excellent material for use
drugs to macromolecular carriers alters their rate of excretion in medical applications such as soft contact lenses. Recently,
from the body, changes their toxicity and immunogenicity, PVA has been used for long-term implants, including a bioar-
and limits their uptake by cells via endocytosis, thus provid- tificial pancreas, artificial cartilage, nonadhesive film, and
ing the opportunity to direct the drug to the particular cell esophagus or scleral buckling material.14) PVA can be avail-
type where its activity is needed.5) In addition, these macro- able various molecular weight according to the usage. PVA
molecular conjugates can accumulate in solid tumors due to also provide a potential targetable drug delivery system.15)
the enhanced microvasculature of tumor tissue.6,7) This phe-
nomenon has been termed enhanced permeability and reten- selective delivery of antineoplastic agents to tumor tissue. In-
tion in relation to tumor targeting (EPR-phenomenon).8)
corporating acid-sensitive bonds between the drug and the
Drug–polymer conjugates are potential candidates for the
In recent years the use of soluble synthetic polymers as polymer is an attractive approach because it ensures effective
drug delivery systems has been received increasing attention. release of the polymer-bound drug at the tumor site. This re-
N-(2-Hydroxypropyl)methacrylamide (HPMA) copolymers lease is either extracellular, resulting from the slightly acidic
containing a tetrapeptide side chain (Gly-Phe-Leu-Gly) ter- condition in tumor tissue, or intracellular, in acidic lyso-
minated in DOX and, optionally, galactosamine, as a target- somes/endosomes after cellular uptake of drug-polymer con-
ing moiety,9) were recently approved in the United Kingdom jugate.16)
for phase I/phase II clinical trials to treat primary hepatoma.
Such delivery systems can be usefully distinguished into
HPMA copolymer prodrugs produced increased life spans those formulations where the anthracyclines is bound to a
and an increased number of long-term survivors10—13) de- carrier molecule using a covalent linkage that is stable in the
pending on the structure of the conjugate, timing of adminis- bloodstream, being cleaved only after internalization into
tration, and number of doses. Similarly, HPMA copolymers cells. In this paper we report on the synthesis of the
containing the anthracycline antibiotics, Daunorubicin poly(vinyl alcohol)–doxorubicin conjugates, their release ex-
(DNR)12,13) and DOX11) were modified with fucosylamine, periment, cellular uptake, and cytotoxicity.
and it was shown that they interact with the receptor on
L1210 cells in vitro and in vivo. Subsequent experiments
∗ To whom correspondence should be addressed. e-mail: kaneo@fupharm.fukuyama-u.ac.jp
© 2008 Pharmaceutical Society of Japan

104
Vol. 31, No. 1
MATERIALS AND METHODS
132.0 (ꢀCH), 135.4 (ꢀC), 37.8 (CH2), 171.9 (CH2COOH).
Preparation of PVA–ADOX Conjugates Ethylenedi-
Materials Doxorubicin hydrochloride and fluorescein amine residues were introduced to the hydroxyl groups of
isothiocyanate isomer I (FITC) were obtained from Wako PVA molecules by the 1,1ꢃ-carbonyldiimidazole (CDI) acti-
Pure Chemical, Osaka, Japan. cis-Aconitic anhydride was vation method.18—20) Four milliliters of dimethyl sulfoxide
obtained from Aldrich, Milwaukee, WI, U.S.A. PVA (80 K, (DMSO) containing CDI (74 mg) was added to 400 mg of
MWꢀ80520) was kindly supplied by Japan Vam & Poval PVA dissolved in 60 ml of DMSO, followed by stirring for
Co., Ltd., Osaka, Japan. All other chemicals and reagents 1 h at room temperature. After several precipitations in bu-
were of the highest grade commercially available.
tanol to remove unreacted reagents, the fraction of CDI-acti-
Synthesis of Aconityl-doxorubicin Aconityl-doxoru- vated PVA was dried in vacuo. Then, ethylenediamine (2 g)
bicin (ADOX) was synthesized by the modified method of was added to the CDI-activated PVA (200 mg) dissolved in
Shen and Ryser.17) Briefly, doxorubicin hydrochloride 50 ml of DMSO and stirred for 48 h at 50 °C. After several
(10 mg) was dissolved in distilled water (10 ml) and cooled precipitations in butanol to remove unreacted reagents, the
on ice. cis-Aconitic anhydride (10 mg) was dissolved in p- fraction of PVA-ethylenediamine was dried in vacuo. The
dioxane (6 ml) and added slowly to the solution with stirring, free amino groups of ethylenediamine spacers introduced
and the pH was immediately adjusted to 9.0 by careful addi- into the PVA molecule were measured by the 2,4,6-trinitro-
tion of 0.5 M NaOH. After 15 min, the pH was adjusted to 7.0 benzenesulfonic acid (TNBS) method.21—23)
and continued for another 1 h. The reaction mixture was ana-
One hundred microliters of DMSO containing 1-ethyl-3-
lyzed and purified by using HPLC. Analysis of ADOX was (3-dimethylaminopropyl)-carbodiimide hydrochloride (EDC)
carried out using a Shimadzu liquid chromatographic system (0.27 mg) and N-hydroxysuccinimide (NHS) (0.16 mg) were
(LC-6A, Kyoto, Japan) with a variable-wavelength UV detec- added to 1 mg of the ADOX-isomers dissolved in 0.1 ml of
tor (SPD-6A) operated at 200 nm. A 4.6ꢁ150 mm, 5-mm DMSO, respectively, followed by stirring for 2 h at room
particle size, C18 reversed-phase column (Cosmosil 5C18, temperature. Then, PVA-ethylenediamine (10 mg) dissolved
Nacalai, Kyoto, Japan) was used at ambient temperature. The in 0.2 ml DMSO was added to the activated-ADOX-isomers
mobile phase was (NH4)2CO3 (3 w/w%) : methanol : acetoni- and stirred for 4 h at room temperature. The reaction mixture
trileꢀ50 : 45 : 5, v/v/v. The injection volume was 20 ml, and was dialyzed against distilled water for 48 h at 4 °C. Excess
the flow rate was 1.0 ml/min.
reagents and low molecular by-products of the reaction were
Purification of ADOX was achieved by HPLC where a removed by this process. The PVA–ADOX conjugates were
250ꢁ20 mm, 5-mm particle size, C18 reversed-phase column obtained from lyophilization of the final solution. Characteri-
(YMC-PackPro C18RS, YMC, Kyoto, Japan) was used at zations of the conjugates were carried out using a high-per-
ambient temperature. The mobile phase was (NH4)2CO3 formance size exclusion chromatography (HPSEC) described
(3 w/w%) : methanolꢀ48 : 52, v/v. The injection volume was later.
1 ml, and the flow rate was 4.0 ml/min. Fractions containing
ADOX-isomers were collected and lyophilized.
DOX Content of the Conjugates The conjugates were
dissolved in distilled water, and the absorbance at 476 nm
The TOF-MS spectrum of ADOX displayed the expected was measured. The DOX content of the conjugates was esti-
pseudomolecular ions at m/z 722 (MꢂNa) and 738 (MꢂK). mated using the calibration curve of doxorubicin standard
Its structure was confirmed by one- and two-dimensional (UV method).
1
1
1
NMR (1H, 13C, H, H COSY, H, 13C HMQC). DOX moi-
The DOX content of the conjugates was estimated by the
eties of cis-ADOX and trans-ADOX: H-NMR (500 MHz, modified method of Seymour.24) The DOX of the conjugates
DMSO-d6) d: 7.91 (1H, d, H-1), 7.91 (1H, t, H-2), 7.64 (1H, were hydrolyzed with 5 M HCl for 10 min at 85 °C. Then, 5 M
t, H-3), 3.99 (3H, s, 4-OMe), 4.94 (1H, t, H-7), 2.12 and 2.18 NaOH was added to neutralize the solution. The hydrolysis
(1ꢂ1H, dd, H2-8), 4.85 (1H, t, 9-OH), 4.59 (2H, d, 9- products were measured by HPLC described later. The
COCH2), 2.89 and 2.99 (1ꢂ1H, dd, H2-10), 5.29 (1H, d, H- amount of DOX was calculated from the quantity of agly-
lꢃ), 1.70 and 1.90 (1ꢂ1H, dd and dt, H2-2ꢃ), 3.95 (cis- cone, hydrolyzed product of DOX (hydrolysis method).
1
ADOX: 1H, m, H-3ꢃ)/4.00 (trans-ADOX: 1H, m, H-3ꢃ), 3.60
In Vitro Release Experiment The release of DOX from
(1H, d, H-4ꢃ), 4.85 (1H, d, 4ꢃ-OH), 4.18 (1H, q, H-5ꢃ), 1.16 the conjugates was determined in a 0.1 M citrate buffer solu-
(3H, d, H3-6ꢃ). 13C-NMR (125 MHz, DMSO-d6) d: 118.9 (C- tion (pH 5.0, 6.0, 7.0, mꢀ0.3) at 37 °C. The experiment was
1), 136.1 (C-2), 119.6 (C-3), 160.7 (C-4), 56.5 (4-OMe), initiated by the addition of the stock solution to a preheated
120.0 (C-4a), 186.4 (C-5), 110.7 (C-5a), 156.0 (C-6), 133.9 buffer solution to give a concentration of 1.0 mg/ml of
(C-6a), 69.8 (C-7), 36.6 (C-8), 74.9 (C-9), 213.7 (9-CO), PVA–ADOX conjugates, respectively. At fixed time intervals,
32.0 (C-10), 134.6 (C-10a), 154.4 (C-11), 110.6 (C-11a), the amounts of DOX released were determined by using a
186.4 (C-12), 135.4 (C-12a), 100.2 (C-1ꢃ), 29.4 (C-2ꢃ), 45.1 HPLC method described later.
(cis-ADOX: C-3ꢃ)/45.7 (trans-ADOX: C-3ꢃ), 67.7 (C-4ꢃ),
Preparation of FITC-Labeled PVA FITC-labeled PVA
66.6 (C-5ꢃ), 16.9 (C-6ꢃ). Aconityl moiety of cis-ADOX: (F-PVA) was prepared by the modified method of deBelder
1H-NMR (500 MHz, DMSO-d6) d: 5.70 (1H, s, ꢀCH), 3.03 and Granath.25) PVA (300 mg) was dissolved in DMSO (8 ml)
and 3.01 (1ꢂ1H, d, CH2), 8.85 (1H, d, NH). 13C-NMR containing 1 drops of pyridine. FITC-I (50 mg) was added,
(125 MHz, DMSO-d6) d: 164.4 (CONH and COOH), 127.2 followed by dibutyltin dilaurate (20 mg), and the mixture was
(ꢀCH), 138.8 (ꢀC), 43.9 (CH2), 171.1 (CH2COOH). trans- heated for 2 h at 95 °C. After several precipitations in butanol
ADOX: 1H-NMR (500 MHz, DMSO-d6) d: 6.42 (1H, s, to remove free dye, the fraction of F-PVA was dried in vacuo
ꢀCH), 3.18 and 3.19 (1ꢂ1H, CH2), 7.57 (1H, d, NH). 13C- at 80 °C. The F-PVA was further purified by size-exclusion
NMR (125 MHz, DMSO-d6) d: 166.8 (CONH and COOH), chromatography on Sephadex G-25, and then freeze-dried.

January 2008
105
The F-PVA sample was dissolved in 25 mM borate buffer (pH to each tube as an internal standard, vortex-mixed, and then
9.0), and the absorbance at 495 nm was measured. The fluo- 33% silver nitrate (0.1 ml) was added. Six milliliters of ace-
rescein content of F-PVA was estimated using the calibration tonitrile was added as an extract medium, samples were vor-
curve of fluorescein sodium (uranine).
tex-mixed three times during a 30 min period. The upper
Preparation of [125I]-Labeled PVA ([125I]-PVA) PVA- layer was carefully removed from the silver chloride precipi-
ethylenediamine (200 mg) was labeled with 0.25 mCi of tate. The samples were evaporated to dryness and redissolved
[
125I]iodine by using Bolton and Hunter reagent for protein in mobile phase (0.1 ml). The amount of DOX was deter-
iodination (GE Healthcare Bio-Sciences, Tokyo, Japan). Un- mined by HPLC described later. The amount of protein in the
reacted [125I] was removed by chromatography on a PD-10 aliquot was determined by the method of Lowry et al.26)
column (Amersham Pharmacia Biotech).
In Vitro Cytotoxicity J774.1 cells (5ꢁ103) were placed
Cells Mouse macrophage-like tumor J774.1, mouse in a 96-well culture plate and cultured in humidified air with
leukemia P388, mouse hepatoma MH134-TC and human his- 5% CO2 at 37 °C for 24 h. Subsequently, the cells were incu-
tiocytic lymphoma U937 were kindly provided by Institute bated with various drug concentrations for 5 h. Control wells
of Development, Aging and Cancer, Tohoku University were identical, except that the test compound was absent.
(Miyagi, Japan). Cells were typically kept in continuous Then, the wells were washed with fresh medium and incu-
logarithmic growth at 37 °C in a humidified atmosphere in bated for 48 h. After incubation, [3H]uridine incorporation
5% CO2–95% air in RPMI 1640 medium (Nacalai, Kyoto, assay was carried out. Two microCi of [3H]uridine was added
Japan), supplemented with 10% heat-inactivated fetal bovine to the wells, which were incubated for 2 h. After incubation,
serum (FBS) (Gibco, Tokyo, Japan) and 50 U/ml penicillin the wells were washed with PBS (pH7.4) then washed with
and 50 mg/ml streptomycin (Dainippon Sumitomo Pharma ice-cold 10% trichloroacetic acid three times. To the washed
Co., Ltd., Tokyo, Japan). The number of viable cells was de- wells, 0.5 M NaOH (150 ml) was added and allowed to stand
termined by the trypan blue exclusion method by using a at room temperature for 30 min. A solubilized aliquot of
Burker-Turk hematocytometer 3 d after incubation.
0.1 ml was neutralized with HCl and added 2 ml of liquid
Fluorescence Microscopic Examination The cells scintillator (Clear-sol I, Nacalai, Kyoto, Japan). The radio-
(8ꢁ105) were placed in a 35 mm culture dish (Iwaki, Funa- activity was determined with a liquid scintillation counter
bashi, Japan) and incubated in humidified air with 5% CO2 at (LSC-1100, Aloka, Tokyo, Japan).
37 °C for 24 h. F-PVA was added to the dishes, which were
Analytical Methods The amount of DOX and its hy-
incubated at 37 °C for 5 h. The concentration of F-PVA in the drolyzed product was determined by HPLC. Chromatography
medium was adjusted to be 500 mg/ml. After incubation, the was carried out using a Shimadzu liquid chromatographic
dishes were washed three times with phosphate-buffered system (LC-6A, Kyoto, Japan) with a variable-wavelength
saline (PBS) (pH 7.4). Specimens were examined with a fluorescent detector (RF-10AXL, Shimadzu). The excitation
Nikon transmitted light fluorescence microscope.
and emission wavelength were set at 470 nm and 560 nm, re-
Uptake of [125I]-PVA For uptake experiments, [125I]- spectively. A 4.6ꢁ150 mm, 5-mm particle size, C18 reversed-
PVA was used as the substrate against J774.1 and P388 cells. phase column (Cosmosil 5C18, Nacalai, Kyoto, Japan) was
The cells (5ꢁ103) were placed in a 96-well culture plate used at ambient temperature. The mobile phase was 34%
(Nunc, Denmark) and cultured in humidified air with 5% acetonitrile in 1% triethylamine adjusted to pH 4.0 with
CO2 at 37 °C for 24 h. The medium containing [125I]-PVA formic acid. The injection volume was 20 ml, and the flow
was added to the wells, which were incubated at 37 °C. The rate was 1.0 ml/min. For every experimental sample the con-
concentration of [125I]-PVA in the medium was adjusted to be tent of DOX and its hydrolyzed product was calculated by
10 mg/ml. After incubation, the wells were washed with PBS measuring the relevant peak area and calibrating against the
(pH 7.4) then washed with ice-cold 10% trichloroacetic acid corresponding peak area derived from the daunorubicin in-
three times. To the washed wells, 0.1 M NaOH (150 ml) was ternal standard.
added and allowed to stand at room temperature for 30 min.
High-performance size-exclusion chromatography (HPSEC)
The radioactivity of a solubilized aliquot of 100 ml was deter- was carried out using a Shimadzu liquid chromatographic
mined with a gamma counter (ARC-380CL, Aloka, Tokyo, system (LC-9A, Kyoto, Japan) equipped with a variable-
Japan). The amount of protein in the aliquot was determined wavelength UV detector (MCPD-3600, Otsuka, Osaka,
by the method of Lowry et al.26)
Japan) and a differential refractometer (RI-8000, Tosho). A
Uptake of PVA–ADOX Conjugates The cells (4ꢁ105) 7.8ꢁ300 mm, TSKge1 G4000PWXL column (Tosoh) was
were placed in a 35 mm culture dish (Iwaki, Funabashi, used at 40 °C. The mobile phase was 0.2 M NaCl in 0.05 M
Japan) and cultured in humidified air with 5% CO2 at 37 °C phosphate buffer, pH 7.0. The injection volume was 100 ml,
for 24 h. PVA–ADOX conjugates and free DOX were added and the flow rate was 1.0 ml/min.
to the dishes, which were incubated at 37 °C for 6 h. The
drug concentration for the microscopic examination and RESULTS
the intracellular release experiment were 20 mg/ml and
125 mg/ml in DOX equivalent, respectively. After incubation,
Synthesis of Aconityl-doxorubicin Aconityl-doxorubi-
the dishes were washed with PBS (pH 7.4) three times. Spec- cin (ADOX) was synthesized by the modified method of
imens were observed with a Nikon transmitted light fluores- Shen and Ryser.17) The amino group of the sugar ring of
cence microscope. Subsequently, the amount of intracellular doxorubicin (DOX) was reacted with cis-aconitic anhydride
release of DOX from the conjugates was determined by the to form a-unsaturated amide bond. Two isomers of trans-
modified method of Seymour.24) The cells were homogenized ADOX (trans-configuration) and cis-ADOX (cis-configura-
in 0.5 ml of PBS (pH 7.4). Daunorubicin (0.5 mg) was added tion) were generated (Fig. 1).

106
Vol. 31, No. 1
Fig. 1. Synthetic Pathway of Aconityl-doxorubicin (ADOX)
and then reacted with PVA-ethylenediamine. The DOX con-
tent of the conjugates estimated by the UV method and the
hydrolysis method were 4.6 and 4.4 w/w%, respectively, cor-
responding to 6.1 and 6.4 mol of DOX/mol of PVA.
HPSEC Analysis The size exclusion chromatography of
the conjugates was also performed by HPLC on a TSK gel
G4000PWXL column. PVA–cis-ADOX and PVA–trans-
ADOX were spectrophotometrically detected at 476 nm,
whereas PVA was analyzed by a differential refractometer.
Elution peaks of the conjugates and PVA were detected simi-
larly, suggesting that the ADOX-substitution has no effect on
the molecular weight of PVA (Fig. 4).
Release of DOX from PVA–ADOX Conjugates The re-
lease of DOX from the conjugates was determined in a 0.1 M
citrate buffer solution (pH 5.0, 6.0, 7.0, mꢀ0.3) at 37 °C (Fig.
5). Both PVA–cis-ADOX and PVA–trans-ADOX were very
stable at neutral pH, but the release of DOX was increased
markedly under acidic conditions. Half-life of the release of
DOX from PVA–cis-ADOX at pH 5.0 was 3 h which was
4.7-fold shorter than that from PVA–trans-ADOX (14 h).
Fig. 2. HPLC Chromatogram of ADOX
Chromatography was carried out using a HPLC system equipped with a variable-
wavelength fluorescent detector. The excitation and emission wavelength were set at
470 nm and 560 nm, respectively. A 4.6ꢁ150 mm, 5-mm particle size, C18 reversed-
phase column was used at ambient temperature. The mobile phase was (NH4)2CO3
(3 w/w%) : methanol : acetonitrileꢀ50 : 45 : 5, v/v/v. The injection volume was 20 ml,
and the flow rate was 1.0 ml/min. Peaks: 1, trans-ADOX; 2, cis-ADOX; 3, DOX.
A representative HPLC chromatogram demonstrated the Furthermore, the amount of DOX released from PVA–cis-
separation of trans-ADOX, cis-ADOX and unmodified DOX ADOX at pH 6.0 was approximately 3-fold higher than that
as shown in Fig. 2. The retention times were 10.2, 12.2 and released from PVA–trans-ADOX.
21.0 min, respectively, and were highly reproducible. Frac-
Uptake of PVA Fluorescence microscopic examination
tions corresponding to the isomers were collected and ana- of the cellular uptake of F-PVA was carried out using various
lyzed by TOF-MS spectrum. Their structure were confirmed kinds of cells (Fig. 6). It was found that MH134-TC and
1
1
by one- and two-dimensional NMR (1H, 13C, H, H COSY, U937 cells internalized F-PVA marginally. In contrast,
1H, 13C HMQC). The yields of trans-ADOX and cis-ADOX J774.1 and P388 cells accumulated F-PVA very efficiently.
were 44.8 and 36.3%, respectively.
Preparation of PVA–ADOX Conjugates ADOX was accumulated in J774.1 and P388 cells at 37 °C. The level of
In addition, Fig. 7 shows that [125I]-PVA were gradually
bound to PVA according to the synthetic scheme shown in
[
125I]-PVA accumulated in the J774.1 cells was three times
Fig. 3. The overall synthetic pathway involved two steps. The higher than that in the P388 cells. The uptake of [125I]-PVA
first was the preparation of PVA-ethylenediamine and the by these cells was saturated in 5—6 h.
second was the binding of the ADOX and PVA-ethylenedi-
amine.
Uptake of PVA–ADOX Conjugates To compare the
cellular uptake and intracellular distribution of DOX with
Hydroxyl group of PVA was activated by CDI. Then, a 10- those of PVA–cis-ADOX, J774.1 cells were incubated in the
fold weight excess of ethylenediamine was reacted with CDI- medium which contained the drug preparations at 37 °C for
activated PVA. An excess of the ethylenediamine was used in 6 h. The fluorescence of DOX was observed with a conven-
order to prevent cross-linking and cyclisation of the PVA tional fluorescence microscopy. As shown in Fig. 8, DOX
chains. The number of free amino groups of the PVA-ethyl- was internalized in the cells very efficiently. PVA–cis-ADOX
enediamine was estimated by the TNBS method.21—23) One was also internalized and the fluorescence of DOX was ob-
mole of PVA showed the color intensity of 25 mol of free served in the nuclear region.
amino group. Free g-carboxylic group of ADOX molecule
Subsequently, the amount of intracellular release of DOX
was activated with an equimolar amount of EDC and NHS, from the conjugates was determined by the modified method

January 2008
107
Fig. 3. Synthetic Pathway of PVA–ADOX Conjugates
¥¥¥
( ), PVA–cis-ADOX
Fig. 4. HPSEC Chromatograms of PVA
(– – –) and PVA–trans-ADOX (——)
HPSEC was carried out on a TSKge1 G4000PWXL column (7.8ꢁ300 mm) with
0.2 M NaCl in 0.05 M phosphate buffer, pH 7.0. PVA–cis-ADOX and PVA–trans-ADOX
were spectrophotometrically detected at 476 nm, whereas PVA was analyzed by a dif-
ferential refractometer.
Fig. 6. Fluorescence Microscopic Examination of the Cellular Distribu-
tion of FITC-Labeled PVA in J774.1 (A), P388 (B), MH134-TC (C) and
U937 (D)
Cells (8ꢁ105) were incubated with F-PVA (500 mg/ml) at 37 °C for 5 h. After incuba-
tion, the cells were washed three times with PBS (pH 7.4) and then examined with a
Nikon transmitted light fluorescence microscope of 400 magnifications.
ADOX and PVA–trans-ADOX were evaluated in comparison
with that of DOX by using J774.1 cells employing a [3H]uri-
dine incorporation assay as a measure of RNA synthesis (Fig.
10). As is evident, the cytotoxicity effects of these com-
pounds were concentration dependent. PVA–cis-ADOX is
more active than PVA–trans-ADOX, but less active than
DOX where the IC50 values for PVA–cis-ADOX and DOX
Fig. 5. Effect of pH on the Release of DOX from PVA–cis-ADOX (A)
and PVA–trans-ADOX (B)
The release of DOX from the conjugates was determined in 0.1 M citrate buffer solu-
tion (mꢀ0.3) of pH 5.0 (ꢀ), 6.0 (ꢁ), and 7.0 (ꢂ) at 37 °C.
of Seymour.24) J774.1 cells were incubated in the medium were 18 mg eq DOX/ml and 0.48 mg/ml, respectively.
which contained PVA–cis-ADOX and PVA–trans-ADOX at PVA–trans-ADOX are only marginally active, the IC50 value
37 °C for 6 h. Figure 9 shows that the amount of DOX re- was not evaluated at the concentrations tested. Significant
leased from PVA–cis-ADOX was approximately 2-fold difference in cytotoxicity was observed between the two con-
higher than that from PVA–trans-ADOX.
jugates.
In Vitro Cytotoxicity The cytotoxicity of PVA–cis-

108
Vol. 31, No. 1
Fig. 7. Cellular Uptake of [125I]-PVA in J774.1 (ꢀ) and P388 (ꢁ)
Cells were incubated with [125I]-PVA (10 mg/ml) at 37 °C. After incubation, the cells
were washed and solubilized to measure the cell-accumulated radioactivity.
Fig. 9. Intracellular Release of DOX from PVA–cis-ADOX (Close Col-
umn) and PVA–trans-ADOX (Open Column) in J774.1
Cells (4ꢁ105) were incubated with PVA–ADOX conjugates (125 mg/ml in DOX
equivalent) at 37 °C for 6 h. After incubation, the cells were washed and homogenized.
The amount of DOX was estimated by using a HPLC equipped with a variable-wave-
length fluorescent detector. The excitation and emission wavelength were set at 470 nm
and 560 nm, respectively. Values are given as meansꢄS.E. (nꢀ3). The statistical signif-
icance was determined based on Student’s t-test (∗ pꢅ0.05).
Fig. 10. Effects of PVA–cis-ADOX (ꢁ), PVA–trans-ADOX (ꢂ) and DOX
(ꢀ) on the Proliferation of J774.1 Cells
Fig. 8. Fluorescence Microscopic Examination of the Cellular Distribu-
tion of PVA–cis-ADOX (A) and DOX (B) in J774.1
Cells (4ꢁ105) were incubated with PVA–cis-ADOX (20 mg/ml in DOX equivalent)
at 37 °C for 6 h. After incubation, the cells were washed three times with PBS (pH 7.4)
and then examined with a Nikon transmitted light fluorescence microscope of 400 mag-
nifications.
Cells (5ꢁ103) were incubated with various drug concentrations for 5 h and then
washed with fresh medium and incubated for another 48 h. After incubation, the cells
were incubated with [3H]uridine for 2 h and harvested. Results are shown as percent of
the cellular uptake of [3H]uridine uptake.
aconitic anhydride consist of Z-substituted double bond (in
respect of N–CꢀO and COOH substituents), whereas the
products of the secondary propene-1,2,3-tricarboxylic acid-
1,3-anhydride could serve either as an isomer where the
DISCUSSION
The first acid-sensitive drug–polymer conjugate was syn- COOH and amide substituents are E arranged or an isomer
thesized by Shen and Ryser.17) These authors coupled with E located carboxylic groups.
daunorubicin to aminoethyl polyacrylamide polymer beads
We found that two isomers of cis-ADOX (cis-configura-
or poly-D-lysine using the aconityl spacer. Thereafter, a num- tion) and trans-ADOX (trans-configuration) were generated
ber of daunorubicin conjugates with monoclonal antibodies in the reaction of DOX and cis-aconitic anhydride as shown
and synthetic polymers have been developed with this in Fig. 1. These products were separated completely by using
method.16,27) Apart from daunorubicin, anthracycline doxo- HPLC (Fig. 2) and were analyzed by TOF-MS spectroscopy
rubicin has been also bound to monoclonal antibodies using and one- and two-dimensional NMR (1H, 13C, 1H, 1H COSY,
the cis-aconityl spacer.28)
1H, 13C HMQC). Two isomers of cis-ADOX and trans-
On the route of acylation of DOX with cis-aconitic anhy- ADOX, 36.3% and 44.8% yields, were chief products in this
dride four isomeric amide products should be expected, reaction. Similar findings were reported by Remenyi et al. in
considering the spontaneous inter-conversion of cis-aconitic the case of DNR. They demonstrated that two isomers of
anhydride to propene-1,2,3-tricarboxylic acid-1,3-anhy- aconityl-daunorubicin (ADNR) belong to the cis- and trans-
dride.29—31) Acylated products derived from the primer isomers of the a-monoamide of ADNR, respectively, by MS-

January 2008
109
spectroscopy and 1H- and 13C-NMR experiments.32)
ADOX were evaluated by using J774.1 cells employing a
ADOX was bound to PVA according to the synthetic [3H]uridine incorporation assay as a measure of RNA synthe-
scheme shown in Fig. 3. It has been suggested that most of sis. A significant difference in antitumor activity between
the free g-carboxylic group of the aconityl spacer molecule PVA–cis-ADOX and PVA–trans-ADOX was observed where
is coupled to macromolecular moiety.17) In this case the free the former was much active than the later (Fig. 10). It was
g-carboxylic group may also participate to the coupling to suggested that the conjugate enters the cells and reaches the
PVA via ethylenediamine spacer, resulting the macromolecu- lysosomal/endosomal compartment, and that the aconityl
lar conjugates of PVA–cis-ADOX and PVA–trans-ADOX, spacer releases DOX from the conjugate in the acidic com-
respectively. The conjugates were generated from the combi- partment of lysosomes/endosomes due to the participation of
nation of PVA and DOX for the first time. The contents of a free carboxylic group.
DOX of the conjugates were estimated by the separated pro-
In conclusion, it is possible to exploit the pH difference
cedures with the ordinary UV method and the hydrolysis between the lysosome/endosome and the extracellular envi-
method.24) The glycosidic bond between the daunosamine ronment, as Shen and Ryser17) first demonstrated, in the de-
ring and the aglycone moiety is hydrolysed relatively easily, velopment of the acid sensitive macromolecular prodrug,
releasing the free aglycone.33) The hydrolysis method de- PVA–cis-ADOX. Due to a different configuration between
tected the aglycone of DOX which can be estimated as the PVA–cis-ADOX and PVA–trans-ADOX, release rates of
PVA-bound DOX selectively. The DOX content estimated by DOX from the conjugates were very different. A higher ef-
this method was 4.4 w/w% which was similar to 4.6 w/w% fect can be expected by using PVA–cis-ADOX in the treat-
by the UV method. These results indicated that the DOX was ment of cancer.
linked approximately 24% of the ethylenediamine spacer
which was introduced to PVA, corresponding 6.1—6.4 mol
of DOX/26 mol of ethylenediamine/1 mol of PVA.
Acknowledgements This work was supported in part by
Grants-in-Aid (No. 13672406) for Scientific Research (C)
In the release experiments, both PVA–cis-ADOX and from the Ministry of Education, Culture, Sports, Science and
PVA–trans-ADOX were very stable at the neutral pH, Technology of Japan (to Y.K.). The authors are also grateful
whereas acidic conditions accelerated the release of DOX to Mr. Hiroshi Noguchi of Japan Vam & Poval Co., Ltd.,
from the conjugates. Furthermore, it was found that half-life Osaka, Japan.
of the release of DOX from PVA–cis-ADOX was signifi-
cantly shorter than that observed with PVA–trans-ADOX REFERENCES
under the same conditions as shown in Fig. 5. Release mech-
anism was considered to be due to the participation of a free
carboxylic group in the acidic condition and the cis-transfor-
1) Young R. C., Ozols R. F., Myers C. E., N. Engl. J. Med., 305, 139—
153 (1981).
2) Dorr R. T., Von Hoff D. D., “Cancer Chemotherapy Handbook,” 2nd
mation was much advantageous to catalyze the intramolecu-
lar hydrolysis of the amide bond.
ed., Appleton and Lange, Norwalk, 1994.
3) Gewirtz D. A., Biochem. Pharmacol., 57, 727—741 (1999).
4) Myers C. E., Chabner B. A., “Cancer Chemotherapy—Principles and
Practice,” Lippincott, Philadelphia, 1990, pp. 356—381.
5) Ringsdorf H., J. Polym. Sci. Polym. Symp., 51, 135—153 (1975).
6) Seymour L. W., Crit. Rev. Ther. Drug Carrier Syst., 9, 135—187
(1992).
7) Yuan F., Dellian M., Fukumura D., Leunig M., Berk D. A., Torchilin V.
P., Jain R. K., Cancer Res., 55, 3752—3756 (1995).
8) Maeda H., Matsumura Y., Crit. Rev. Ther. Drug Carrier Syst., 6, 193—
210 (1989).
9) Kopecek J., Rejmanova P., Strohalm J., Ulbrich K., Rihova B., Chytry
V., Lloyd J. B., Duncan R., U. S. Patent 5037883 (1991).
10) Kopecek J., Duncan R., J. Controlled Release, 6, 315—327 (1987).
11) Duncan R., Hume I. C., Kopeckova P., Ulbrich K., Strohalm J.,
Kopecek J., J. Controlled Release, 10, 51—63 (1989).
We found that the cellular uptake of F-PVA varied be-
tween cell lines, with mouse macrophage-like tumor J774.1
and mouse leukemia P388 showing the highest levels of up-
take (Fig. 6). Fluorescence microscopic examination revealed
that F-PVA was distributed mainly in the lysosomal/endoso-
mal fraction. Furthermore, the level of [125I]-PVA accumu-
lated in the J774.1 cells was significant higher than that in
the P388 cells (Fig. 7). Half-lives of uptake and elimination
in the J774.1 cells were 6 h and 14 h, respectively (data are
not shown). Mean residence time (MRT) of [125I]-PVA in the
J774.1 cells was estimated to be 20 h. These data indicated
that the tumor cells remained to incorporate PVA enough to
release the DOX.
12) Duncan R., Kopeckova P., Strohalm J., Hume I. C., Lloyd J. B.,
Kopecek J., Br. J. Cancer, 57, 147—156 (1988).
Intracellular distribution of DOX was also observed by flu- 13) Duncan R., Kopeckova-Rejmanova P., Strohalm J., Hume I., Cable H.
C., Pohl J., Lloyd J. B., Kopecek J., Br. J. Cancer, 55, 165—174
(1987).
14) Mallapragada S. K., McCarthy-Schroeder S., “Handbook of Pharma-
orescence microscopy (Fig. 8). The fluorescence of DOX in
the cells treated with PVA–cis-ADOX was detected mainly in
the cytoplasm and nuclear region, whereas the fluorescence
ceutical Controlled Release Technology,” Marcel Dekker, New York,
treated with DOX was widely distributed in the cells. The in-
tracellular release of DOX from PVA–ADOX conjugates was
estimated by the HPLC method of Seymour.24) The level of
DOX released from PVA–cis-ADOX was significant higher
than that from PVA–trans-ADOX (Fig. 9). Furthermore, the
intracellular concentration of DOX was significantly higher
than that in the medium in both cell lines. Therefore, it was
confirmed that the tumor cells took the macromolecular con-
jugates which released the free DOX in the lysosomal/endo-
somal acidic conditions.
2000, pp. 31—46.
15) Kaneo Y., Hashihama S., Kakinoki A., Tanaka T., Nakano T., Ikeda Y.,
Drug Metab. Pharmacokinet., 20, 435—442 (2005).
16) Kratz F., Beyer U., Schutte M. T., Crit. Rev. Ther. Drug Carrier Syst.,
16, 245—288 (1999).
17) Shen W. C., Ryser H. J., Biochem. Biophys. Res. Commun., 102,
1048—1054 (1981).
18) Yamaoka T., Tabata Y., Ikada Y., J. Pharm. Pharmacol., 47, 479—486
(1995).
19) Hearn M. T., Methods Enzymol., 135, 102—117 (1987).
20) Beauchamp C. O., Gonias S. L., Menapace D. P., Pizzo S. V., Anal.
Biochem., 131, 25—33 (1983).
The cytotoxicities of PVA–cis-ADOX and PVA–trans-
21) Habeeb A. F., Anal. Biochem., 14, 328—336 (1966).

110
Vol. 31, No. 1
22) Fields R., Biochem. J., 124, 581—590 (1971).
23) Fields R., Methods Enzymol., 25, 464—468 (1972).
1193 (1988).
29) Bland N., Thorpe J., J. Chem. Soc., 101, 1490—1498 (1912).
24) Seymour L. W., Ulbrich K., Strohalm J., Kopecek J., Duncan R., 30) Malachowski R., Giedroyc M., Jerzmanowska Z., Chem. Ber., 61,
Biochem. Pharmacol., 39, 1125—1131 (1990).
25) deBelder A. N., Granath K., Carbohydr. Res., 30, 375—378 (1973).
26) Lowry O. H., Rosebrough N. J., Farr A. L., Randall R. J., J. Biol.
Chem., 193, 265—275 (1951).
27) Diener E., Diner U. E., Sinha A., Xie S., Vergidis R., Science, 231,
148—150 (1986).
2525—2538 (1928).
31) Pratt L., Smith B., Trans. Faraday Soc., 63, 2858—2867 (1967).
32) Remenyi J., Balazs B., Toth S., Falus A., Toth G., Hudecz F., Biochem.
Biophys. Res. Commun., 303, 556—561 (2003).
33) Di Marco A., Gaetani M., Scarpinato B., Cancer Chemother. Rep., 53,
33—37 (1969).
28) Yang H. M., Reisfeld R. A., Proc. Natl. Acad. Sci. U.S.A., 85, 1189—