Modification of the antitumor effect of doxorubicin by phosphorylated retinoids conjugated to α-fetoprotein

Full text of DOI:10.1023/A:1015388429085

Pharmaceutical Chemistry Journal
Vol. 35, No. 12, 2001
MODIFICATION OF THE ANTITUMOR EFFECT OF DOXORUBICIN
BY PHOSPHORYLATED RETINOIDS CONJUGATED TO a-FETOPROTEIN
1
1
1
1
1
D. V. Arsenov, S. V. Babitskaya, I. I. Vashkevich, I. D. Dad’kov, M. A. Kisel’,
1
1
B. B. Kuz’mitskii, and O. A. Strel’chenok
Translated from Khimiko-Farmatsevticheskii Zhurnal, Vol. 35, No. 12, pp. 14 – 17, December, 2001.
Original article submitted June 28, 2001.
Retinoids are known to possess a broad spectrum of bio-
logical activity, which is related to the ability of these com-
pounds to interact with RAR and RXR – the intrisic recep-
tors containing transcription factors belonging to the ste-
roid – thyroid superfamily of nuclear hormonal receptors [1].
Exhibiting strong antiproliferative and cytodifferentiating ef-
fects, retinoids are of interest for the chemotherapy of
oncologic disorders [2, 3].
on the phosphorylation stage of and, hence, is applicable
only to the synthesis of unsaturated acid derivatives.
R"
O
O
Me
Me
Me
–
CNHCHCHZOPO
Me
Me
–
R'
O
R¢
R²
Z
N-(4-hydroxyphenyl)-all-trans-retinoylamide (fenretini-
de) represents a group of retinoids synthesized in attempts at
obtaining a compound of this type possessing minimum tox-
icity [4]. Fenretinide reduced the incidence of cancer of the
mammary glands, skin, lung, bladder, and prostate in experi-
mental animals treated with carcinogens [5]. It was also
found that fenretinide inhibited the growth of human ovarian
carcinoma transplanted in mice and potentiated the antitumor
activity of cisplatin [6, 7]. Treatment of the neuroblastoma
cells with fenretinide prior to introducing cisplatin, etoposide,
or carboplatin into the culture medium led to apoptosis with
evidence of a pronounced synergism in the drug action [8].
Below we report on the synthesis of new retinoids repre-
senting all-trans-retinoic acid (RA) amides with O-phos-
phates of ethanolamine, L-serine, L-threonine, and L-tyro-
sine. We have studied the influence of these retinoids conju-
gated to a-fetoprotein (a-FP) on the antitumor action of the
antibiotic doxorubicin with respect to inoculated Ehrlich
ascitic carcinoma. The synthesized compounds are structur-
NR-Ser-PA
COOH
COOH
COOH
H
CH
H
—
—
NR-Thr-PA
NR-Tyr-PA
3
NR-EA-PA
H
H
—
The problem of acylating and phosphorylating serine,
threonine, and tyrosine presents certain difficulty because
amino acid molecules contain three reactive functional
groups. Additional difficulty is related to the fact that free
serine, threonine, and tyrosine are soluble only in water: this
circumstance hinders using acylating agents based on fatty
acids. We suggest using the initial substrates (serine,
threonine, and tyrosine) in the form of methyl ester hydro-
chlorides. The acylation is performed with a mixed anhy-
dride of RA and n-butylcarbonic acid. For tyrosine, the reac-
tion at room temperature leads to the formation of a consid-
erable amount of bis-O,N-acyl derivative. Conducting the
acylation of tyrosine at –18°C for 1 h yields a virtually pure
methyl ester (NR-Tyr-PA).
The obtained N-acylamino acid methylates were
phosphorylated using a b-cyanoethyl method, which was
originally proposed for the nucleotide synthesis [12]. How-
ever, the phosphorylation of N-acyltyrosine methylate under
the originally proposed conditions (with pyridine as a sol-
vent) proceeds with a low yield, which is probably related to
the formation of an ionic bond between a phenolic hydroxy
group and pyridine. A more satisfactory yield of the
phosphorylated diester is achieved when the reaction is con-
ally
analogous
to
N-palmitoyl-O-phospho-L-serine
N-palmitoyl-O-phospho-L-tyrosine
(
(
NP-Ser-PA) and
NP-Tyr-PA), which are antagonists of the receptors of
lysophosphatidic acid (an intercell messenger inducing cell
proliferation and cancer cell invasion) [9].
It should be noted that recently described synthesis of
NP-Ser-PA and NP-Tyr-PA [10, 11] involves hydrogenolysis
1
Institute of Bioorganic Chemistry, National Academy of Sciences of
Belarus, Minsk, Belarus.
657
0
091-150X/01/3512-0657$25.00 © 2001 Plenum Publishing Corporation

6
58
D. V. Arsenov et al.
Frel.
the possibility of complexation between a-FP and the syn-
thesized retinoids by measuring the quenching of fluores-
cence from tyrosine and tryptophan-162 residues as a result
of the induction-resonance energy transfer from an excited
state of the donor to the retinoyl acceptor. Owing to an over-
1
0
0
0
0
.0
.8
.6
.4
.2
lap of the fluorescence spectrum of the donor (lmax
=
=
3
28 nm) and the absorption spectrum of the acceptor (lmax
345 nm), an effective fluorescence quenching is observed
when the protein fluorophore is excited at a wavelength of
290 nm. It was found that the efficiency of NP-Ser-PA and
NP-Tyr-PA binding to the oncofetal protein differs but little
from their binding to free retinoic acid (Fig. 1). Adding a dif-
ferent ligand (docosahexaenoic acid incapable of quenching
fluorescence in this spectral interval [14]) to the a-FP solu-
tion did not lead to any significant change in the fluorescence
intensity. Therefore, the fluorescence quenching is related to
an induction-resonance energy transfer realized within a re-
gion bounded by the Forster radius (usually, on the order of
30 – 50 Å) that is within the boundaries of the ligand protein
complex studied.
The influence of the complexes between a-FP and
retinoids was studied on a model of Ehrlich ascitic carci-
noma in mice. The results presented in Table 2 shows that
doxorubicin intravenously injected in a dose of 3 ´ 2 mg/kg
produces no significant effect upon the dense ascites frac-
tion, but reliably inhibits (28%, p < 0.05) the cell growth in
the ascitic form of Ehrlich carcinoma. The compound
NR-Tyr-PA (as well as a-FP) introduced by the same scheme
in the corresponding dose produces no significant change in
the tumor development either. Nor does the antitumor action
(Ehrlich ascitic tumor growth inhibition) of the doxorubicin
complex with NR-Tyr-PA reach a level characteristic of free
anthracycline. Only adding a-FP (250 mg/kg) to the complex
increases the model tumor growth inhibition up to 38%
(p < 0.05), although the decrease in the ascites residue vol-
ume amounts only to 21% (which is statistically insignifi-
cant). On the whole, the antitumor action of this
retinoid – a-FP complex is not superior to the doxorubicin
effect. Nevertheless, the experimental results reliably indi-
cate that a-FP is a significant factor accounting the the posi-
tive effect of the composition studied. A certain specific pro-
tein content in this composition not only eliminates the nega-
3
10
350
370
l, nm
1
2
3
4
0
1
2
3
4
[
ligand]/[a-FP]
Fig. 1. The intrinsic fluorescence of a-FP quenched as a result of
adding ligands: (1 ) NR-Ser-PA; (2 ) NR-Tyr-PA; (3 )
all-trans-retinoic acid; (4 ) docosahexaenoic acid. The inset shows
the fluorescence spectrum of a-FP excited at a wavelength of
90 nm (solid curve) and a fragment of the absorption spectrum of
retinoids (dashed curve).
2
ducted in acetonitrile for 5 days at 20°C (Table 1). This mod-
ified method is convenient because both protective moieties
methyl ester and b-cyanoethyl group) can be simultaneously
(
removed with a virtually quantitative yield under mild alka-
line conditions. During the hydrolysis, relatively labile am-
ide and phosphate bonds are fully retained, and the RA
polyene chain is neither subject to isomerization or oxida-
tion, nor exhibits polymerization.
The targeted delivery of the synthesized retinoids to a tu-
mor tissue can be provided by a-FP. Selecting this molecule
as a vector is explained by the fact that receptors of this pro-
tein are expressed in considerable amounts in the tumor cells
[
13]. At the same time, RA is capable of specifically binding
to a-FP [14]. However, the synthesized compounds possess a
more complicated structure as compared to that of free RA,
which may hinder their binding to a-FP. We have estimated
1
TABLE 1. Yields and H NMR Spectra of Phosphorylated Retinoids
1
Compound
NR-Ser-PA
Yield, %
24
Empirical
3 3
H NMR spectrum (in CD SOCD ): d, ppm
C
23
C
24
C
29
C
22
H
H
H
H
34NO
36NO
38NO
34NO
7
7
7
5
P
P
P
P
1.0 (s, 6H); 1.4 – 2.4 (m, 15H); 3.9 – 4.0 (bs, 2H); 4.3 – 4.4 (m, 1H); 5.8 – 7.0 (m, 6H);
.2 – 8.4 (m, 3H)
1.0 (s, 6H); 1.2 – 1.4 (d, 3H); 1.4 – 2.4 (m, 15H); 4.1 – 4.3 (m, 2H); 5.8 – 7.0 (m, 6H);
.2 – 8.4 (m, 3H)
1.0 (s, 6H); 1.4 – 2.4 (m, 15H); 2.9 – 3.1 (m, 2H); 4.3 – 4.4 (m, 1H); 5.8 – 7.0 (m, 6H);
.0 – 7.2 (q, 4H); 8.2 – 8.4 (m, 3H)
1.0 (s, 6H); 1.4 – 2.4 (m, 15H); 3.4 – 3.5 (bs, 2H); 3.9 – 4.0 (bs, 2H); 5.8 – 7.0 (m, 6H);
.2 – 8.4 (m, 3H)
8
NR-Thr-PA
NR-Tyr-PA
NR-EA-PA
27
19
28
8
7
8

Modification of the Antitumor Effect of Doxorubicin
659
TABLE 2. Effect of Complexation with Retinoic Acid Derivatives and a-Fetoprotein on the Antitumor Activity of Doxorubicin with Respect
to the Ehrlich Ascitic Carcinoma in C3H Mice
Tumor process parameters
Preparation (dose)
Number of tumor cells, Tumor growth inhibition,
Ascites residue volume
6
p
´10
%
First series
Control
1.9 ± 0.62
1.7 ± 0.72
2.1 ± 1.36
1.8 ± 0.29
1.9 ± 0.62
1.5 ± 0.94
601.5 ± 42.87
433.2 ± 30.87
523.9 ± 147.39
470.5 ± 50.61
498.2 ± 87.67
373.2 ± 97.50
–
–
Doxorubicin (2 mg/kg)
NR-Tyr-PA (6.8 mg/kg)
a-FP (0.25 mg/kg)
28
13
22
17
38
< 0.05
> 0.05
> 0.05
> 0.05
< 0.05
Doxorubicin + NR-Tyr-PA (2.0 + 6.8 mg/kg)
a)
Doxorubicin + NR-Tyr-PA + a-FP
Second series
Control
1.3 ± 0.46
0.7 ± 0.37
1.1 ± 0.26
1.0 ± 0.58
0.6 ± 0.19
1.2 ± 0.44
499.4 ± 70.15
249.3 ± 41.38
407.9 ± 60.53
315.4 ± 47.66
212.1 ± 71.97
367.1 ± 36.08
–
–
b)
Doxorubicin (3 mg/kg)
50
< 0.02
< 0.05
< 0.05
< 0.02
< 0.05
c)
Doxorubicin + NR-Ser-PA + a-FP
Doxorubicin + NR-Thr-PA + a-FP
Doxorubicin + NR-Tyr-PA + a-FP
Doxorubicin + NR-EA-PA + a-FP
– 64
b)
b)
37
58
c)
– 47
Notes. a The three-component mixture in the first series of experiments was composed of doxorubicin (2 mg/kg), retinoid (6.8 mg/kg), and
b
c
a-FP (0.25 mg/kg); in the second series, doxorubicin (3 mg/kg), retinoid (6.8 mg/kg), and a-FP (0.25 mg.kg); relative to control; relative to
doxorubicin.
tive influence of NR-Tyr-PA upon the antitumor activity of
doxorubicin but even renders the inhibiting effect of the
complex more pronounced in comparison to that of the anti-
biotic component alone.
The above data provided a base for the second series of
experiments, aimed at a comparative study of the modifying
action of four RA derivatives in compositions of the doxoru-
bicin – ligand – a-FP type. In this series, only RA derivatives
were replaced in the complex. Taking into account the rela-
tively low activity of doxorubicin in the experiments de-
scribed above, the drug dose was increased from 2 to
ues are 64 and 47% lower as compared to the effect of free
doxorubicin). Based on these data, we may suggest that (i) an
optimum composition is offered by the doxorubicin complex
with NR-Tyr-PA and (ii) a-FP produces a positive influence
on the antitumor properties of this complex. We may also
suggest that the absence of a reliable effect for the mixture of
doxorubicin with NR-Tyr-PA is explained by the formation
of an ion complex between the components, related to the in-
teraction between the amino group of daunosamine and the
phosphate moiety of NR-Tyr-PA. The presence of a-FP
seems to destroy this ion complex, binds NR-Tyr-PA, and
delivers it to the target tumor cells.
3
mg/kg.
It was established that the increase in the doxorubicin
dose led to the antitumor effect being more pronounced,
reaching 50% with respect to the tumor cell growth inhibi-
tion; the ascites residue volume was found to decrease as
well (Table 2). The complex of doxorubicin with NR-Thr-PA
in the presence of a-FP exhibited a lower antitumor activity
as compared to that of free doxorubicin: the Ehrlich carci-
noma cell growth was inhibited by 37% (p < 0.05). The RA
complex containing only NR-Tyr-PA produced a more pro-
nounced cytotoxic action upon the tumor as compared to
both the aforementioned complex and free doxorubicin.
The two other complexes, involving doxorubicin conju-
gated to NR-Ser-PA and NR-EA-PA, produced no significant
cytotoxic action upon Ehrlich ascitic carcinoma in compari-
son to mice in the untreated control group. Moreover, the
presence of these retinoids reduced the cytotoxic activity of
doxorubicin (the corresponding tumor growth inhibition val-
Thus, adding a-FP and synthetic N-all-trans-retinoyl-L-
tyrosine-O-phosphate to doxorubicin reliable increases the
chemotherapeutic effect of antibiotic, while the other synthe-
sized retinoids were found ineffective with respect to the tu-
mor model studied.
EXPERIMENTAL CHEMICAL PART
The experiments were performed with all-trans-retinoic
acid, n-butylchloroformate, b-cyanoethylphosphate (Ba salt),
N,N¢-dicyclohexylcarbodiimide, and amino acids from
Sigma (USA) and doxorubicin from Pharmacia AB (Swe-
den). a-FP was isolated from retroplacental serum of
parturient women as described in [15]. The analysis for phos-
phorus was performed using the method described in [16].

6
60
D. V. Arsenov et al.
General method for the synthesis of retinoids. To a so-
lution of RA (152 mg, 0.5 mmole) and triethylamine (52 mg,
.51 mmole) in 3 ml of a THF – acetonitrile (1 : 2) mixture,
cooled to – 18°C, was added n-butylchloroformate (70 mg,
.51 mmole) and the reaction mixture was allowed to stand
at – 18°C for 20 min. Then the precipitate of triethylamine
hydrochloride was separated and the remaining mixed anhy-
dride solution was added to a solution of ethanolamine or
amino acid methylate (1 mmole) in 1 ml of methanol (also
cooled to – 18°C). The reaction mixture was kept for 20 min
at – 18°C and stirred for 2 h at room temperature (for tyro-
sine, the acylation was conducted for 1 h at – 18°C). Then
EXPERIMENTAL BIOLOGICAL PART
The Ehrlich ascitic carcinoma was modeled by a conven-
tional method [17] in inbred C3H female mice weighing
0
20 – 24 g. The animals were kept on a standard laboratory
0
diet with free access to water. The tumor was taken from the
given strain mice and intraperitoneally implanted into test
6
mice (2 ´ 10 carcinoma cells). On the next day, the test mice
were randomized to form experimental groups, each of seven
mice, and the untreated control group of 15 mice.
Doxorubicin (3 mg/kg) and its complexes with retinoids
were injected into test mice according to an interrupted
schedule, on the second, fourth, and sixths day after tumor
implantation. The substances were injected into a lateral tail
vein with 0.2 ml of an isotonic NaCl solution. The composi-
tions of doxorubicin complexes are indicated in Table 2.
The results were evaluated on the eighth day in terms of
the tumor growth inhibition, ascites residue volume, and
number of tumor cells. The tumor growth inhibition, ex-
pressed as percentage relative to control, was calculated by
the formula [(control – test)/control] ´ 100%. The results
were statistically processed; activities of the standard drug
and the complexes tested were compared in terms of the Stu-
dent t-criterion.
1
0 ml of 0.5 M HCl cooled to – 5°C was added and the mix-
ture was extracted with 20 ml of diethyl ether. The ether ex-
tract was washed with 10 ml of water, dried over anhydrous
Na SO , and evaporated. The residue was dissolved in chlo-
2
4
roform (2 ml) and applied onto a chromatographic column
filled with alumina (alkaline, activity degree II according to
Brockman). The column was eluted with 10 ml of a chloro-
form – methanol (9 : 1, v/v) mixture. Finally, the eluate was
evaporated to obtain N-retinoyl derivatives of ethanolamine
or amino acid methylate (yield, 95%). TLC on silica gel re-
veals a single yellow spot at R = 0.4 – 0.5 (benzene –
f
dioxane – acetic acid, 25 : 5 : 1, v/v).
REFERENCES
To this product were added
a
solution of
b-cyanoethylphosphate (pyridinium salt, 2 mmole) in 3 ml of
anhydrous pyridine and N,N¢-dicyclohexylcarbodiimide
1
2
. R. Evans, Science, 240(4854), 889 – 895 (1988).
. M. A. Smith, D. R. Parkinson, B. D. Cheson, et al., J. Clin.
Oncol., 10, 839 (1992).
. C. W. Boone, G. J. Kelloff, and W. E. Malone, Cancer Res.,
5
(
2
413 mg, 2 mmole) and the reaction mixture was stirred for
0 h at 20°C. Then the mixture was cooled to 0°C and 0.5 ml
of water was added, after which stirring was continued for
0 min at room temperature (for tyrosine derivative, the
3
0(1), 2 – 9 (1990).
4. R. C. Moon, H. J. Thompson, P. J. Becci, et al., Cancer Res.,
9(4), 1339 – 1346 (1979).
. R. C. Moon and R. G. Metha, Prev. Med., 18(5), 576 – 591
1989).
3
3
5
6
7
8
phosphorylation was performed in acetonitrile at 20°C for 5
days). The precipitate of N,N¢-dicyclohexylurea was sepa-
rated by filtration. The filtrate was evaporated to dryness and
the residue was purified by chromatography on a silica gel
column eluted with chloroform – methanol (70 – 60 : 30 –
(
. F. Formelli and L. Cleris, Cancer Res., 53(22), 5374 – 5376
(1993).
. F. Formelli and L. Cleris, Eur. J. Cancer, 36(18), 2411 – 2419
2000).
. P. E. Lovat, M. Ranalli, F. Bernassola, et al., Int. J. Cancer,
8(6), 977 – 985 (2000).
9. W. H. Moolenaar, Exp. Cell Res., 253, 230 – 238 (1999).
(
4
0, v/v) mixtures. Fractions containing the target product
8
were evaporated and the residue was dissolved in 1 ml of
THF. The solution was added dropwise with stirring to 1.5 M
NaOH (4 ml) at 0°C and allowed to stand for 25 min. Then
the reaction mixture was acidified to pH 1.5 with 1 M HCl
and extracted with a chloroform – methanol (2 : 1, v/v) mix-
ture. The extract was washed with a chloroform – water
10. T. Sugiura, A. Tokumura, L. Gregory, et al., Arch. Biochem. Bi-
ophys., 311(2), 358 – 368 (1994).
1
1. R. Bittman, B. Swords, K. Liliom, and G. Tigyi, J. Lipid Res.,
7(2), 391 – 398 (1996).
12. J. G. Moffatt, J. Am. Chem. Soc., 85, 1118 – 1123 (1963).
3
1
3. C. Esteban, M. Geuskens, and J. Uriel, Int. J. Cancer, 49,
25 – 430 (1991).
4. C. Aussel and R. Masseyeff, Biochem. Biophys. Res. Commun.,
19(3), 1122 – 1127 (1984).
4
(
10 : 9) mixture, evaporated, and purified by chromatogra-
1
phy on a silica gel column. The target phosphates were
eluted with chloroform – methanol (30 – 20 : 70 – 80, v/v)
mixtures. Fractions containing the target product were com-
1
15. L. I. Survilo, I. I. Vashkevich, M. N. Ermolenko, et al., Izv. Nat.
Akad. Navuk Belarusi, Ser. Khim. Navuk, No. 2, 81 – 85 (2000).
16. V. E. Vaskovsky, E. V. Kostetsky, and I. M. Vasendin, J. Chro-
mat., 114, 129 – 141 (1975).
bined and evaporated. TLC on silica gel: R = 0.05 – 0.10
f
(
chloroform – methanol – 28% ammonia, 9 : 7 : 2, v/v).
17. A. Goldin and S. Carter, in: Cancer Medicine, J. F. Holland and
E. Frei (eds.), Philadelphia (1973), pp. 605 – 628.
Yields and some analytical data are listed in Table 1.