Selective targeting of 2′-deoxy-5-fluorouridine to urokinase positive malignant cells in vitro

Article abstract of DOI:10.1016/j.bmcl.2010.03.029

A urokinase targeting conjugate of 2′-deoxy-5-fluorouridine (5-FUdr) was synthesized and tested for tumor-cell selective cytotoxicity in vitro. The 5-FUdr prodrug 2′-deoxy-5-fluoro-3′-O-(3-carboxypropanoyl)uridine (5-FUdrsuccOH) containing an ester-labile

Full text of DOI:10.1016/j.bmcl.2010.03.029

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2908–2911
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Selective targeting of 20-deoxy-5-fluorouridine to urokinase positive
malignant cells in vitro
b
b
b
a
Kara L. Vine a, , Julie M. Locke , John B. Bremner , Stephen G. Pyne , Marie Ranson
*
a School of Biological Sciences and Centre for Medical Bioscience, University of Wollongong, NSW 2522, Australia
b School of Chemistry and Centre for Medicinal Chemistry and Pharmacology, University of Wollongong, NSW 2522, Australia
a r t i c l e i n f o
a b s t r a c t
A urokinase targeting conjugate of 20-deoxy-5-fluorouridine (5-FUdr) was synthesized and tested for
tumor-cell selective cytotoxicity in vitro. The 5-FUdr prodrug 20-deoxy-5-fluoro-30-O-(3-carboxypropa-
noyl)uridine (5-FUdrsuccOH) containing an ester-labile succinate linker was attached to the specific uro-
kinase inhibitor plasminogen activator inhibitor type II (PAI-2) and was found to preferentially kill
urokinase-over expressing cancer cells. Up to 7 molecules of 5-FUdr were incorporated per PAI-2 mole-
cule without affecting protein activity. This is the first time a small organic cytotoxin has been conjugated
to PAI-2.
Article history:
Received 29 January 2010
Revised 4 March 2010
Accepted 5 March 2010
Available online 9 March 2010
Keywords:
20-Deoxy-5-fluorouridine
PAI-2
Ó 2010 Elsevier Ltd. All rights reserved.
Anticancer prodrug
Targeted therapy
Selected delivery
The fluoropyrimidine, 5-fluorouracil (5-FU), and its cytotoxic
metabolite 20-deoxy-5-fluorouridine (5-FUdr; 1) are used in combi-
nation with folinic acid as standard treatments in the adjuvant
setting for a variety of carcinomas, including those of the stomach,
colon and breast.1–3 However like many cytotoxic chemo-therapeu-
tics, these drugs suffer from a number of disadvantages due to their
lack of specificity and, as a result, dose limiting toxicity is often prob-
lematical. To overcome this obstacle, efforts have been made to
develop a variety of non-toxic prodrugs of 5-FU and 5-FUdr that be-
come activated only at the site of interest. Recently, particular atten-
tion has been given to the ligand-directed targeting approach,
whereby the targeting of 5-FUdr to tumor cells has been described
through the use of rabbit serum albumin,4 epidermal growth factor,5
small cyclic peptides and monoclonal antibodies directed against
tumor associated antigens.6–10 This last strategy has demonstrated
some therapeutic potential although a number of limitations still
persist. These include the inefficient uptake of the antibody conju-
gates into solid tumors due to poor blood perfusion11 and large anti-
body size, loss of antibody activity and specificity after chemical
modification,12 and rapid elimination of the conjugate after initia-
tion of an immune response due to the use of antibodies of non-hu-
man origin.13
has the ability to preferentially target urokinase-positive meta-
static breast cancer cells in vitro14 and tumor xenografts and
micrometastates in vivo.15–17 With these characteristics, together
with its relatively small size (47 kDa) and immunoneutrality,18 it
was anticipated that by linking a derivatized version of 5-FUdr to
PAI-2, the cytotoxin could be specifically delivered to, and concen-
trated exclusively within, urokinase expressing tumor cells, there-
by reducing collateral damage. Herein we describe the synthesis
and characterization of a 20-deoxy-5-fluorouridine derivative and
its conjugation to the targeting protein PAI-2 via a succinate linker.
In vitro tumor-cell selective cytotoxicity is also reported.
A series of reactions were performed to prepare 20-deoxy-5-flu-
orouridine (5-FUdr, 1) (Scheme 1) for chemical attachment to PAI-
A promising alternative targeting strategy is that based on the
urokinase plasminogen activator (uPA) system and its natural
inhibitor, plasminogen activator inhibitor type II (PAI-2). PAI-2
Scheme 1. Synthesis of 20-deoxy-5-fluoro-30-O-(3-carboxypropanoyl)-uridine (5-
FUdrsuccOH, 4). Reagents and conditions: (a) 4,40-dimethoxytriphenylmethyl
chloride, DMAP, Et3N, pyridine, rt, 16 h (61%); (b) succinic anhydride, DMAP,
CH2Cl2, rt, 24 h (86%); (c) 8:2 AcOH/H2O, 2 h rt (68%).
* Corresponding author. Tel.: +61 2 42214256; fax: +61 2 42214135.
E-mail address: kara@uow.edu.au (K.L. Vine).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.03.029

K. L. Vine et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2908–2911
2909
aSy), at least 5 free lysine groups are accessible for chemical conjuga-
tion at the surface of the protein. However, this was only determined
for the amino nitrogens that were >50% accessible and did not in-
clude the N-terminal amino acid, suggesting that incorporation of
a larger number of cytotoxin molecules is possible. ESI-MS data also
demonstrated that the 5-FUdrsucc-PAI-2 preparations contained a
mixture of several 5-FUdrsucc-PAI-2 moieties with an average of
3–4 cytotoxin units incorporated per protein molecule. The relative
amount of free, unmodified PAI-2 was <10%.
The ability of 5-FUdrsucc-PAI-2 to form stable complexes with
uPA was next investigated in order to confirm that the uPA inhib-
itory activity of PAI-2 was retained after modification. 5-FUdrsucc-
PAI-2 was found to form stable complexes of ꢀ75 to 100 kDa with
both the B chain (25 kDa) and the A + B chain (50 kDa) of uPA,
respectively. This was similar to that observed for unmodified
PAI-2, confirming that the modification of lysine functional groups
leading to the incorporation of up to 7 molecules of 5-FUdrsucc
does not compromise the inhibitory ability of the protein. Further-
more, when a sample of the conjugate was retested after 1 year
(after sterile storage at 4 °C), it was found to retain its inhibitory
activity as evidenced by the formation of high molecular weight
uPA/PAI-2 complexes (Fig. 2).
To assess the targeting ability of the conjugate in vitro, the
newly synthesized 5-FUdrsucc-PAI-2 and corresponding free cyto-
toxins were evaluated for inhibitory activity in two breast adeno-
carcinoma cell lines (Table 1) varying in their expression levels of
uPA. The cytotoxicity of 5-FUdr with the succinate linker attached
(5-FUdrsuccOH, 4) against the metastatic MDA-MB-231 cell line
(expressing high levels of uPA and uPAR), was 13-fold lower than
Scheme 2. Activation of 5-FUdrsuccOH (4) to form the active ester
5 and
conjugation to PAI-2 to form the uPA directed prodrug 5-FUdrsucc-PAI-2 via amide
bond formation with free surface lysine residues. Reagents and conditions: (a) N-
hydroxysuccinimide, dicyclohexylcarbodiimide rt, 3 h; (b) PAI-2 (PBS pH 7.4), rt, 3 h
(>85%).
2. This involved the introduction of a free carboxyl group19 by for-
mation of an ester between the secondary alcohol of a selectively
protected 5-FUdr derivative (2)6,20 and succinic acid which pro-
ceeded in >65% yield. The carboxyl group of 5-FUdrsuccOH (4)6,21
was then activated by forming the ester 5 with 1.1 equiv of NHS
and 6 equiv of DCC in anhydrous DMF (Scheme 2). The molar
equivalents employed were based on those described previously
for the activation of 4.6 However, it was noted that the active ester
formed rapidly at RT as evidenced by TLC analysis (Rf 0.55, silica
DCM) and longer reaction times, such as those described by Goer-
lach and co-workers6 were unnecessary. The active ester 5 was
then conjugated without being isolated or further purified to afford
5-FUdrsucc-PAI-2 (Scheme 2).
To optimize coupling, different molar excesses of the activated
ester 5 were added to PAI-2 in order to obtain conjugates with the
highest possible substitution. The greatest 5-FUdrsucc: PAI-2 ratio
was obtained using a 50-fold molar excess of 5, yielding up to 7 mol-
ecules of 5-FUdr for every protein molecule (Fig. 1). This was deter-
mined by ESI-MS analysis and corresponds well to previous reports
regarding cDTPA incorporation into PAI-2.15,22 According to the 3D
structure of PAI-2 (prepared using the SwissModelÒ program, ExP-
that of 5-FUdr (IC50 0.21
lM and IC50 2.85 lM, respectively, Ta-
ble 1), suggesting that succinylation of the 30-group of 5-FUdr re-
sults in the loss of some activity. Similarly, Goerlach et al.
reported a 10-fold decrease in activity for 5-FUdrsuccOH against
a murine thymoma cell line in vitro and hypothesized that it
may have been due to the inefficient uptake of the modified com-
pound into the cell by the nucleoside transport system.6 The low-
ered IC50 value of 4 compared to 1 may also be indicative of the
relative stability of the succinate linker in culture media.10 The
IC50 of the 5-FUdrsucc-PAI-2 conjugate, with an average of 3 mol-
Figure 1. ESI-MS of 5-FUdrsucc-PAI-2. Conjugates were prepared for electrospray ionization mass spectrometry (ESI-MS) in Milli Q water and were made up fresh to a final
concentration of 1–10 M. Numbers displayed on the peaks represent the number of 5-FUdrsucc molecules incorporated into the PAI-2 targeting ligand. The difference in
mass for each species represents the exact weight of 5-FUdr, plus the succinate linker (327 g molꢁ1), giving a correlation coefficient (R2) = 1.
l

2910
MW (KDa)
K. L. Vine et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2908–2911
1
2
3
4
5
A
PAI-2 polymers
120
250
150
100
75
PAI-2 dimers
100
80
60
40
20
0
PAI-2:uPA
complexes
***
25
20
15
10
5
***
5-FUdrsucc-PAI-2
* Cleaved PAI-2
50
37
MCF-7
MDA
0
+ uPA
- uPA
25
20
0.0001
0.01
1
[Conjugate] in terms of 5FUdr (µM)
Figure 2. Samples fractioned by SDS PAGE showing the ability of unmodified and
modified PAI-2 to form stable complexes with HMW uPA. Briefly, PAI-2 and 5-
FUdrsucc-PAI-2 were incubated with uPA (2:1) at 37 °C for 30 min. Samples were
then fractionated under non-reducing conditions using 10% acrylamide SDS PAGE
and visualized by staining with Coomassie blue. Lane 1: uPA, Lane 2: PAI-2, Lane 3:
PAI-2: uPA, Lane 4: 5-FUdrsucc-PAI-2, Lane 5: 5-Fudrsuc-PAI-2: uPA. *Cleaved PAI-2
occurs when a portion of wild-type PAI-2 is cleaved by uPA in the CD-loop.23
140
120
100
80
B
60
40
ecules of 5-FUdr incorporated per protein molecule was calculated
to be 0.77 lM against the MDA-MB-231 cell line, which was almost
5-FUdrsucc-PAI-2
5-FUdrsucc-BSA
20
four times lower than 5-FUdrsuccOH. This is a significant finding as
potent cytotoxic activity is maintained after chemical modification,
a phenomenon which by no means should be assumed. Of impor-
tance was that more than four fold24 greater activity was observed
in the high uPA expressing cell line (MDA-MB-231) than the low
uPA expressing cell line (MCF-7) after treatment with the cytotoxin
conjugate (Table 1, Fig. 3A). This suggests that PAI-2 aids in the cel-
lular uptake of the modified cytotoxin through receptor mediated
endocytosis which results in partial hydrolysis of the succinate lin-
ker and subsequent release of 5-FUdr. Similarly, Zhang et al. report
partial release of 5-fluoro-20-deoxyuridine from the cyclic peptide
CNGRC using the same ester-labile linking system.10 The specific
targeting ability of PAI-2 was further exemplified in the colorectal
cell line HCT-116 where a non-specific negative control conjugate
(5-FUdrsucc-BSA) was found to be inactive at the highest concen-
tration tested (Fig. 3B), while the PAI-2 targeted conjugate killed
35% of cells at the highest concentration (31 nM equiv protein con-
cn). Furthermore, saturation of the urokinase plasminogen activa-
tor receptor (uPAR) on MCF-7 cells by pre-incubation with
0
0.001
0.01
0.1
1
10
[Conjugate] nM
Figure 3. The in vitro cytotoxicity of 5-FUdrsucc-PAI-2 against the high and low
uPA and uPAR expressing breast adenocarcinoma cell lines, MDA-MB-231 and MCF-
7 and the colorectal cell line HCT-116. (A) Breast cancer cells (1.0 ꢃ 104) were
cultured for 48 h in the presence of increasing concentrations of the protein
cytotoxin conjugate 5-FUdrsucc-PAI-2 and percent viability of the cells determined
in reference to PBS vehicle controls. Inset: 5-FUdrsucc-PAI-2 at a single concentra-
tion of 3.8
lM was incubated with MCF-7 cells pre-treated with (+) or without (ꢁ)
exogenous uPA (18 nM) for 48 h. Percent viability was determined in reference to
PBS controls. (B) HCT-116 cells (1.0 ꢃ 104) were cultured for 24 h in the presence of
increasing concentrations of the protein cytotoxin conjugate 5-FUdrsucc-PAI-2 or
the negative control conjugate 5-FUdrsucc-BSA and percent viability of the cells
determined in reference to PBS vehicle controls (see Supplementary data (Fig. S1)
for 72 h results). Data points are the means of triplicate experiments SE. *** P
<0.001 extremely significant.
ated via a uPA-dependent mechanism. This has previously been re-
ported for PAI-2 conjugated to the
a
-emitting radioisotope 213Bi
whereby PAI-2-213Bi was found to be selectively toxic to uPA
expressing breast cancer cells in vitro, but not to freshly isolated,
normal human leukocytes on which cell-surface localized active
uPA was not detectable.22 Although the full mechanism of action
of the 5-FUdrsucc-PAI-2 remains to be elucidated, it is anticipated
that the conjugate would be digested in the lysosomes after inter-
nalization via receptor mediated endocytosis. Non-specific ester-
ases present in the lysosomes could hydrolyze the ester bond
between 5-FUdr and the succinyl linker to liberate free 5-FUdr,
which is likely to be metabolized to produce the active compo-
exogenous uPA, followed by treatment with 3.8
lM 5-FUdrsucc-
PAI-2 for 48 h showed significant increase in cytotoxicity
a
(as measured by the decrease in cell viability (P = 0.0002)) com-
pared to MCF-7 cells with endogenous receptor bound uPA only
(Fig. 3A, inset). MDA-MB-231 cells with high endogenous uPAR
bound uPA were even more sensitive (viability = 6.0% 0.29, data
not shown), showing a reduction in cell viability of between 3
and 4 times that of MCF-7 cells treated with (+) or without (ꢁ)
uPA. Considering cell surface uPA is essential for PAI-2:uPA:uPAR
endocytosis,26 this indicates that cytotoxicity is most likely medi-
Table 1
The effect of 5-FUdrsucc-PAI-2 and unconjugated cytotoxins 5-FUdr (1) and 5-FUdrsuccOH (4) on two mammary adenocarcinoma cell lines varying in their expression levels of
uPA
uPAb
IC50
(lM)
a
Cell line
5-FUdr (1)
5-FUdrsuccOH (4)
PAI-2
5-FUdrsucc-PAI-2
MCF-7
MDA-MB-231
Low
High
NRc [ꢂ20.3]
NR [>144]
2.85 ( 0.331)
NAd
NA
NR [>3.13]
0.77 ( 0.167)
0.21 ( 0.003)
a
IC50 values were calculated based on moles of cytotoxin from sigmoidal dose response curves (variable slope), generated using GraphPad Prism V. 5 (GraphPad Software
Inc.). Values are the mean of triplicate experiments SEM.
b
Relative expression levels of uPA.22,25
NR: Not reached, even at the highest concentration tested, in brackets.
NA: Not active, even at the highest concentration tested (i.e., 1 lM).
c
d

K. L. Vine et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2908–2911
2911
6. Goerlach, A.; Krauer, K. G.; McKenzie, I. F.; Pietersz, G. A. Bioconjugate Chem.
1991, 2, 96.
7. Henn, T. F.; Garnett, M. C.; Chhabra, S. R.; Bycroft, B. W.; Baldwin, R. W. J. Med.
Chem. 1993, 36, 1570.
8. Krauer, K. G.; McKenzie, I. F.; Pietersz, G. A. Cancer Res. 1992, 52, 132.
9. Pietersz, G. A.; Krauer, K. J. Drug Target 1994, 2, 183.
10. Zhang, Z.; Hatta, H.; Ito, T.; Nishimoto, S. Org. Biomol. Chem. 2005, 3, 592.
11. Jain, R. K. J. Natl. Cancer Inst. 1989, 81, 570.
12. Wu, A. M.; Senter, P. D. Nat. Biotechnol. 2005, 23, 1137.
13. Dillman, R. O. Antibody Immunocon. Radiopharm. 1990, 3, 1.
14. Ranson, M.; Tian, Z.; Andronicos, N. M.; Rizvi, S.; Allen, B. J. Breast Cancer Res.
Treat. 2002, 71, 149.
15. Allen, B. J.; Tian, Z.; Rizvi, S. M.; Li, Y.; Ranson, M. Br. J. Cancer 2003, 88, 944.
16. Li, Y.; Rizvi, S. M.; Ranson, M.; Allen, B. J. Br. J. Cancer 2002, 86, 1197.
17. Stutchbury, T. K.; Al-Ejeh, F.; Stillfried, G. E.; Croucher, D. R.; Andrews, J.; Irving,
D.; Links, M.; Ranson, M. Mol. Cancer Ther. 2007, 6, 203.
nents required for inhibition of thymidylate synthase and hence
DNA synthesis.
In conclusion, the use of the cytotoxic metabolite 5-FUdr in the
clinic is currently limited due to its narrow therapeutic window.
PAI-2, an irreversible, specific inhibitor of the metastatic marker
uPA has the ability to overcome this problem by specifically deliv-
ering the potent molecule to urokinase expressing tumor cells, ulti-
mately reducing toxic side effects and overcoming many of the
limitations associated with MAb use in targeted therapy. This is
the first time a small organic cytotoxin has been conjugated to
PAI-2. Further studies assessing the in vivo efficacy of this prodrug
are under investigation.
18. Croucher, D. R.; Saunders, D. N.; Lobov, S.; Ranson, M. Nat. Rev. Cancer 2008, 8, 535.
1
19. Compound 3: H NMR (CDCl3) d 2.34 (ddd, J = 6, 8, 14 Hz, 1H, H20), 2.58 (dd,
Acknowledgments
J = 6, 14 Hz, 1H, H20), 2.69 (m, 4H, COCH2CH2CO), 3.39 (dd, J = 2, 11 Hz, 1H, H50),
3.52 (dd, J = 3, 11 Hz, 1H, H50), 3.78 (s, 6H, 2 ꢃ OCH3), 4.19 (d, J = 2 Hz, 1H, H40),
5.45 (d, J = 6 Hz, 1H, H30), 6.27 (br t, J = 6 Hz, 1H, H10), 6.84 (dd, J = 3, 9 Hz, 4H,
ArH), 7.22 (t, J = 7 Hz, 1H, phenyl ArH), 7.28–7.31 (overlapping m, 6H, ArH),
7.38 (d, J = 7 Hz, 2H, ArH), 7.83 (d, JHF = 6 Hz, 1H, H6). 13C NMR (CD3OD) d 25.1,
24.8, 34.2, 51.2, 59.4, 71.3, 80.1, 81.3, 83.4, 109.3, 109.4, 123.2, 123.9, 124.1,
126.0, 131.0 (d, JCF = 29 Hz, CFCH), 136.9 (d, JCF = 239 Hz, CF), 140.0, 145.2,
153.4 (d, JCF = 26 Hz, CFCO), 154.7 (2C), 167.6, 172.4. LRES-MS m/z 646.9 ([M]ꢁ).
This study was supported by the University of Wollongong
through the Centre for Medical Bioscience (CMB), the Centre for
Medicinal Chemistry and Pharmacology (CMCP), a URC Small
Grant, a University Cancer Research Grant and a UPA student
scholarship for K. L. Vine. We are also grateful to the Illawarra Can-
cer Carers Inc., Kiama, Minnamurra and Gerringong Sunrise Rotary,
The Robert East Memorial Fund, Professor P. Clingan and other pri-
vate donors for funding assistance and to Dr. K. Benkendorff for
support. Assoc. Professor M. Ranson was a recipient of a Cancer
Institute NSW Fellowship award.
1
20. Compound 2: H NMR (CDCl3) d 2.25 (ddd, J = 7, 7, 14 Hz, 1H, H20), 2.51 (ddd,
J = 3, 7, 14 Hz, 1H, H20), 3.38 (dd, J = 3, 11 Hz, 1H, H50), 3.42 (dd, J = 3, 11 Hz, 1H,
H50), 3.76 (s, 6H, 2 ꢃ OCH3), 4.10 (d, J = 3 Hz, 1H), 4.56 (t, J = 3 Hz, 1H), 6.31 (t,
J = 6 Hz, 1H, H10), 6.83 (dd, J = 2, 9 Hz, 4H, ArH), 7.20 (t, J = 7 Hz, 1H, ArH), 7.29
(overlapping m, 6H, ArH), 7.40 (d, J = 7 Hz, 2H, ArH), 7.83 (d, JHF = 6 Hz, 1H, H6).
13C NMR (CDCl3) mixture of rotamers d 41.0, 55.0, 55.1, 55.2, 55.4, 63.3, 71.96,
72.03, 85.5, 85.6, 86.5, 87.1, 113.3, 127.0, 128.0, 129.9, 130.0, 135.2 (d,
JCF = 23 Hz, CFCH), 140.6 (d, JCF = 238 Hz, CF), 144.2, 149.1, 157.1 (d, JCF = 26 Hz,
CFCO), 158.6. LREI-MS m/z 548 ([M]+).
Supplementary data
1
21. Compound 4: H NMR (CD3OD) d 2.29 (ddd, J = 6, 6, 14 Hz, 1H, H20), 2.44–2.50
(overlapping m, 3H, CH2 and H20), 2.60 (t, J = 7 Hz, 2H, CH2), 3.81 (d, J = 3 Hz,
2H, H50), 4.14 (d, J = 2 Hz, 1H, H40), 5.30 (d, J = 6 Hz, 1H, H30), 6.25 (t, J = 7 Hz,
1H, H10), 8.19 (d, JHF = 6 Hz, 1H, H6). 13C NMR (CD3OD) 31.6, 32.9, 38.4, 62.7,
76.3, 86.5, 86.8, 126.1 (d, JCF = 35 Hz, CFCH), 142.0 (d, JCF = 234 Hz, CF), 151.4,
160.3 (d, JCF = 25 Hz, CFCO), 175.3, 181.1. LRES-MS m/z 344.9 ([M]ꢁ).
22. Ranson, M.; Tian, Z.; Andronicos, N. M.; Rizvi, S.; Allen, B. J. Breast Cancer Res.
Treat. 2002, 71, 149.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.03.029.
References and notes
23. Lobov, S.; Wilczynska, M.; Bergstrom, F.; Johansson, L. B.; Ny, T. J. Mol. Biol.
2004, 344, 1359.
24. This value is likely to be considerably higher. An exact IC50 value was not
obtained due to limited amounts of material.
25. Andronicos, N. M.; Ranson, M. Br. J. Cancer 2001, 85, 909.
26. Al-Ejeh, F.; Croucher, D.; Ranson, M. Exp. Cell Res. 2004, 297, 259.
1. Meyerhardt, J. A.; Mayer, R. J. N. Eng. J. Med. 2005, 352, 476.
2. Kelder, W.; Hospers, G. A.; Plukker, J. T. Expert Rev. Anticancer Ther. 2006, 6, 785.
3. van Laar, J. A.; Rustum, Y. M.; Ackland, S. P.; van Groeningen, C. J.; Peters, G. J.
Eur. J. Cancer 1998, 34, 296.
4. Barbanti-Brodano, G.; Fiume, L. Experientia 1974, 30, 1180.
5. Poznansky, M. J.; Juliano, R. L. Pharmacol. Rev. 1984, 36, 277.