Thermoresponsive chlorambucil derivatives for tumour targeting

Article abstract of DOI:10.1002/anie.201101133

Putting the heat on: Drugs that are essentially inactive at 37°C and are active under mild hyperthermia can be synthesized by modification of chlorambucil with thermoresponsive groups (see picture; C gray, Cl green, F yellow, H white, N blue, O red). This

Full text of DOI:10.1002/anie.201101133

Communications
DOI: 10.1002/anie.201101133
Antitumor Agents
Thermoresponsive Chlorambucil Derivatives for Tumour Targeting
Catherine M. Clavel, Olivier Zava, Frꢀdꢀric Schmitt, Blanka Halamoda Kenzaoui,
Alexey A. Nazarov, Lucienne Juillerat-Jeanneret, and Paul J. Dyson*
Many of the most widely applied anticancer agents result in
severe side-effects, including, in extreme cases, secondary
cancers that are only detected a long time after administration
of the drug has ceased.[1] A promising approach to overcome
nonselectivity relies on drug enhancement by the application
of external techniques, such that the toxicity of the drug is low
until it is activated at the tumor site. One such combination
approach is to combine chemotherapy with hyperthermia at
the tumor site.[2] Indeed, the cytotoxicity of some anticancer
drugs is enhanced under mild hyperthermia (40–428C), even
though the drugs also work under normal conditions, and
were not intentionally designed for this application.[3] The
thermosensitivity of small-molecule drugs can be enhanced by
attaching them to thermosensitive macromolecules, for
example, liposomal drug carriers,[4] that are insoluble at
378C and become soluble under hyperthermia and cross the
cell membrane.[5] Other polymers that are soluble at 378C
may aggregate in a heated
molecular-weight thermosensitive drugs that exclude macro-
molecular carriers would be attractive, and herein we describe
rationally designed thermoactive molecular derivatives of
chlorambucil (CLB);[9] these derivatives are essentially inac-
tive at 378C, and are activated by mild hyperthermia (418C)
in vitro. This behavior, which is referred to as thermoactivity,
was achieved by covalently linking perfluorinated “pony
tails” to CLB, as perfluorinated compounds were shown to be
highly thermoresponsive in the field of catalysis.[10]
A series of perfluorinated CLB derivatives (1–3) and their
hydrocarbon equivalents (4–7) were synthesized. CLB was
coupled by an ester link with either a fluorinated alcohol or its
hydrocarbon equivalent using N,N’-dicyclohexylcarbodiimide
(DCC) and 4-dimethylaminopyridine (DMAP), as shown in
Scheme 1 (see the Supporting Information for full details of
synthesis and characterization). CLB derivatization with
hydrophobic chains makes compounds 1–3 and 7 less soluble
tumor.[6] For example,
ThermoDox is
a
drug
delivery system based on
a low temperature sensi-
tive liposome (LTSL) con-
taining doxorubicin. This
system is currently in pha-
se III clinical trials in com-
bination with radiofre-
quency ablation (RFA)
for the treatment of hepa-
tocellular
carcinoma
(HCC).[7] When heated to
428C, the LTSL releases
100% of the doxorubicin
drug in less than 20 sec-
Scheme 1. Synthesis of CLB derivatives 1–7.
onds.[8]
development
However,
of
the
low-
than CLB in water at 378C, with the solubility of 3 and 7
increasing rapidly with temperature. Every derivative has a
higher partition coefficient (logPoctanol/water) value (6.26–12.47)
than CLB (logPo/w = 3.92), with the perfluorinated analogues
being the most lipophilic (see the Supporting Information for
temperature–solubility profiles in water and logPo/w values).
Consequently, the anticancer activity of CLB and its
derivatives was initially evaluated in two human ovarian
carcinoma cell lines (A2780 and its cisplatin-resistant variant
A2780cisR) by using the MTT assay. IC50 values were
obtained on cells treated with the compounds incubated for
72 h at 378C, and also under mild hyperthermia, that is, the
cells were incubated for 4 h at 41.58C followed by 68 h at
378C (Table 1). At 378C, the hydrocarbon analogues 4 and 5
[*] C. M. Clavel, Dr. O. Zava, Dr. A. A. Nazarov, Prof. P. J. Dyson
Institut des Sciences et Ingꢀnierie Chimiques
Ecole Polytechnique Fꢀdꢀrale de Lausanne (EPFL)
1015 Lausanne (Switzerland)
E-mail: paul.dyson@epfl.ch
Dr. F. Schmitt, B. Halamoda Kenzaoui, Dr. L. Juillerat-Jeanneret
University Institute of Pathology
Centre Hospitalier Universitaire Vaudois (CHUV)
1011 Lausanne (Switzerland)
Supporting information for this article (detailed descriptions of the
synthesis and characterization of all compounds, procedures for the
cytotoxicity determination, the plasmid DNA interaction study, and
the comet assays) is available on the WWW under http://dx.doi.org/
10.1002/anie.201101133.
7124
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7124 –7127

Table 1: Cytotoxicity of CLB and 1–7 in A2780 and A2780cisR cells at
378C and 41.58C.
in the HT-29 cell line; CLB and 1 have comparable IC50 values
while 3 becomes more active than CLB at 378C. Although 6
has a lower IC50 value at 418C, it is still essentially inactive. In
the HT-29 cell line, only 7 becomes less active at the higher
temperature. The compounds behave similarly in the HCEC
cell line; CLB and 1 are slightly cytotoxic at 378C and both
have increased cytotoxicity at 418C. Reasonable thermores-
ponsive behavior is observed for 3, which is inactive at 378C
and at higher temperature becomes as active as CLB at 378C.
The IC50 values for 6 and 7 do not change significantly at
418C. In the HT-1080 cell line, CLB and 1 have comparable
cytotoxicity at both 37 and 418C. CLB and 1 are less cytotoxic
at 418C than at 378C, whereas the toxicity of 3, 6, and 7
increases from modest at 378C to moderate at 418C, with 3
becoming more active than CLB and as active as 1. In the
THP-1 cell line, apart from CLB that remains moderately
cytotoxic at both temperatures, all compounds show an
increase in their cytotoxicity with temperature, with 1
becoming as cytotoxic as CLB and 3 undergoing the largest
change in IC50 values (D = 32 mm) in this cell line. At 378C,
CLB is the most cytotoxic of the series in the U-937 cell line,
followed by 1, while 6 and 7 are moderately toxic, and 3 is
inactive. At 418C, CLB, 1, 6, and 7 are less cytotoxic and 1
becomes more cytotoxic than CLB. Compound 3 is the only
derivative that exhibits an increase in cytotoxicity with
temperature. In the TK-6 cell line at 378C, CLB and 1 are
reasonably cytotoxic compared to the other compounds, and
at 418C, CLB, 1, and 3 become more toxic; 1 is as toxic as
CLB at 378C. The IC50 values of 6 and 7 increase at higher
temperature; 7 remains essentially unchanged and 6 becomes
inactive.
From the data provided in Table 2, compound 3, which has
the highest logPo/w value (see the Supporting Information), is
the only derivative with universal thermoactive behavior in all
the tested cell lines. The extent of this behavior depends on
the cell line with the best thermoactive effect in adherent cell
lines (HT-29, HCEC, and HT-1080), in which 3 is as cytotoxic
as CLB at 418C. Compound 7 does not show thermores-
ponsive behavior in U-937 and TK-6 cell lines. In general, 6
shows the same thermoactive behavior as 3 but to a lesser
extent, that is, the difference in cytotoxicity at 37 and 418C is
smaller, and in U-937 and TK-6 cells no temperature
selectivity is observed, thus indicating the importance of
perfluorinated chains over aliphatic chains to induce the
thermoresponsive behavior. The short perfluorinated chain in
1 leads to a compound with an overall cytotoxicity that is
comparable to that of CLB with only a modest thermores-
ponsive effect observed in the HT-1080 cell line.
Compound
IC50 in A2780 [mm]
IC50 in A2780cisR [mm]
378C
41.58C
378C
41.58C
CLB
1
2
3
4
5
6
7
12Æ6
26Æ8
38Æ12
>200
6Æ2
39Æ14
38Æ9
23Æ7
37Æ5
14Æ7
45Æ7
51Æ10
51Æ4
43Æ5
65Æ5
25Æ6
>200
16Æ1
19Æ2
45Æ10
>200
52Æ1
67Æ2
27Æ6
40Æ4
17Æ7
35Æ6
66Æ2
41Æ5
9Æ1
40Æ4
>200
are more cytotoxic than CLB. In contrast, 6 is less active than
CLB in A2780, but has a comparable IC50 value to CLB in
A2780cisR cell line at 378C. Compound 2 is more active than
CLB at 378C, but only in A2780cisR cell line. Compound 2 is
less cytotoxic than CLB in A2780 cells, as are 1, 3, and 7, with
the latter two compounds being inactive (IC50 > 200 mm) at
378C. Under mild hyperthermia, CLB and its derivatives have
equivalent cytotoxicity or are less cytotoxic relative to
treatment at 378C. Compound 1 has the same cytotoxicity
as CLB at 41.58C, and 6 appears to be less active than CLB in
both cell lines.
Compounds 2 and 4 become even more active than CLB
at 41.58C. On increasing the temperature to 41.58C, 3 and 7
become active in both cell lines. Notably, in the A2780cisR
cell line at 41.58C, 3 and 7 are even more active than CLB at
either 378C (IC50 = 43 mm) or at 41.58C (IC50 = 52 mm;
Table 1). Compounds 1, 3, 6, 7, and CLB were further studied
in five human cancer cell lines (HT-29 colon cancer cells, HT-
1080 fibrosarcoma cells, THP-1 and U937 monocytic mye-
loma cells, and TK-6 lymphoblastoid cells) and HCEC human
endothelial cells as a model for angiogenic cells at 378C and
under mild hyperthermia that includes 2 h at 418C (Table 2).
In the HT-29 cell line, CLB and 1 are moderately toxic at
378C while 3, 6, and 7 are inactive with IC50 values above
100 mm. At 418C, CLB, 1, 3, and 6 have increasing cytotoxicity
Table 2: Cytotoxicity of CLB, 1, 3, 6, and 7 on adherent (A) and non-
adherent (B) cells at 378C and 418C.
A)
IC50 [mm]
IC50 [mm]
IC50 [mm]
HT-29
HT-1080
HCEC
378C
418C
378C
418C
378C
418C
CLB
1
3
6
7
54Æ3
48Æ2
189Æ4
>200
44Æ1
44Æ1
50Æ3
111Æ2
>200
21Æ1
23Æ2
81Æ2
64Æ3
67Æ2
35Æ1
32Æ1
32Æ2
40Æ2
52Æ4
64Æ4
89Æ4
125Æ1
99Æ1
114Æ3
54Æ1
49Æ2
64Æ2
101Æ1
109Æ4
Preliminary investigations into the mechanism of action of
3 with respect to the synergistic effect with hyperthermia were
undertaken. Since CBL is a DNA alkylating agent,[11] the
interaction of 3 with DNA was evaluated in vitro with
pBR322 plasmid DNA, by incubation with 3 at various
concentrations (50, 100 and 200 mm) for 24 h at 378C, or 1 h at
418C followed by 23 h at 378C. The effect on DNA was
visualized by gel electrophoresis (Figure 1), which showed
that 3 interacts slightly with DNA at 378C, as the amount of
supercoiled DNA (SC) decreases with the formation of
alkylated DNA. The exact nature of the damaged DNA is not
120Æ3
B)
IC50 [mm]
THP-1
IC50 [mm]
U-937
IC50 [mm]
TK-6
378C
418C
378C
418C
378C
418C
CLB
1
3
6
7
25Æ5
35Æ1
84Æ2
105Æ1
94Æ4
28Æ1
26Æ2
52Æ2
77Æ2
82Æ5
4Æ1
6Æ1
25Æ1
18Æ2
62Æ5
81Æ5
59Æ3
22Æ4
31Æ2
83Æ4
86Æ1
51Æ2
13Æ1
22Æ5
60Æ2
174Æ5
58Æ2
>200
60Æ1
48Æ3
Angew. Chem. Int. Ed. 2011, 50, 7124 –7127
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7125

Communications
no difference in DNA damage at 37 and 418C, and 1, which
has a short perfluorinated chain, gives rise to a slight
thermoresponsive effect (see the Supporting Information).
These data are in excellent agreement with the universal
thermoactive behavior of 3, that is, hyperthermia alone does
not have any effect on TK-6 cells and is slightly toxic for HT-
29 cells, and in the presence of 3 little change is observed at
378C, whereas under mild hyperthermia the comet assay
shows a significant increase in DNA damage.
Figure 1. Agarose gels of plasmid DNA incubated at 378C for 24 h
(left) or at 418C for 1 h and 23 h at 378C (right) with 3 at various
concentrations. The first line (ref) shows the control DNA.
In conclusion, we have shown that an organic anticancer
drug modified with a perfluorinated chain can exert thermor-
esponsive behavior to target tumor tissue. The concept uses
hyperthermia to trigger drug cytotoxicity inside heated tumor
cells. The fluorinated CLB derivative 3 has shown thermoac-
tive behavior in the eight tested cell lines, and, under mild
hyperthermia, is as cytotoxic as the parent drug, but does not
show significant toxicity at 378C. This discovery is a first step
toward the rational design of other thermoactive anticancer
drugs.
clear, but could correspond to alkylation that results in cross-
linking,[12] although there is also evidence to suggest that CLB
can induce strand breaks and produce open circular DNA.[13]
Under hyperthermia, this transformation is increased and the
appearance of a new band between the two reference bands is
also observed. Therefore, hyperthermia helps to trigger the
activity of 3 with DNA in vitro.
To evaluate the damage caused by 3 on cellular DNA, a
comet assay[14] was performed on HT-29 and TK-6 cells. Cells
were incubated with 3 (or CLB or 1 as controls; see the
Supporting Information) for 24 h at 378C, or 2 h at 418C
followed by 22 h at 378C. Damage was visually evaluated by
using fluorescence microscopy (4’,6-diamidino-2-phenylin-
dole (DAPI) staining) by classifying cells into five categories
according to the comet tails (Figure 2 and Figure 3).[15] At
378C, 3 causes modest damage to cellular DNA. Under mild
hyperthermia, however, the amount of cellular DNA damage
is proportional to the concentration of 3, whereas CLB shows
Received: February 15, 2011
Revised: May 3, 2011
Published online: June 17, 2011
Keywords: antitumor agents · drug design · fluorine ·
.
hyperthermia · thermoactivity
[1] T. Henne, D. Schmhl, Cancer Treat. Rev. 1985, 12, 77.
[2] a) B. Hildebrandt, P. Wust, O. Ahlers, A. Dieing, G. Sreenivasa,
T. Kerner, R. Felix, H. Riess, Crit. Rev. Oncol. Hematol. 2002, 43,
33; b) R. D. Issels, Eur. J. Cancer 2008, 44, 2546.
[3] a) M. Urano, Int. J. Hyperthermia 1999, 15, 79; b) E. Gabano, D.
Colangelo, A. R. Ghezzi, D. Osella, J. Inorg. Biochem. 2008, 102,
629.
[4] a) D. Needham, G. Anyarambhatla, G. Kong, M. W. Dewhirst,
Cancer Res. 2000, 60, 1197; b) L. H. Lindner, Clin. Cancer Res.
2004, 10, 2168; c) L. Li, T. L. M. ten Hagen, D. Schiper, T. M.
Wijnberg, G. C. van Rhoon, A. M. M. Eggermont, L. H. Lind-
ner, G. A. Koning, J. Controlled Release 2010, 143, 274; d) G.
Kong, G. Anyarambhatla, W. P. Petros, R. D. Braun, O. M.
Colvin, D. Needham, M. W. Dewhirst, Cancer Res. 2000, 60,
6950; e) A. M. Ponce, Z. Vujaskovic, F. Yuan, D. Needham,
M. W. Dewhirst, Int. J. Hyperthermia 2006, 22, 205; f) S.
Unezaki, K. Maruyama, N. Takahashi, M. Koyama, T. Yuda,
A. Suginaka, M. Iwatsuru, Pharm. Res. 1994, 11, 1180.
[5] a) J. A. Mackay, A. Chilkoti, Int. J. Hyperthermia 2008, 24, 483;
b) D. E. Meyer, B. C. Shin, G. A. Kong, M. W. Dewhirst, A.
Chilkoti, J. Controlled Release 2001, 74, 213; c) M. R. Dreher, W.
Liu, C. R. Michelich, M. W. Dewhirst, A. Chilkoti, Cancer Res.
2007, 67, 4418.
Figure 2. DAPI-stained nuclei after the comet assay at different
damage stages: class 0 (left, intact nucleus, untreated cell), class 2
(center, medium-sized tail, cell treated with 50 mm of 3 at 378C),
class 4 (right, damaged nucleus, 100 mm of 3 under hyperthermia).
[6] D. Raucher, A. Chilkoti, Cancer Res. 2001, 61, 7163.
[7] http://clinicaltrials.gov/ct2/show/NCT00617981; www.celsion.-
com.
[8] a) J. K. Mills, D. Needham in Methods in Enzymology, Vol. 387
(Ed.: D. Nejat), Academic Press, New York, 2004, pp. 82 – 113;
b) M. L. Hauck, S. M. LaRue, W. P. Petros, J. M. Poulson, D. Yu,
I. Spasojevic, A. F. Pruitt, A. Klein, B. Case, D. E. Thrall, D.
Needham, M. W. Dewhirst, Clin. Cancer Res. 2006, 12, 4004 –
4010.
[9] a) J. E. Ultmann, G. A. Hyman, A. Gellhorn, JAMA J. Am. Med.
Assoc. 1956, 162, 178 – 183; b) S. Sachanas, G. A. Pangalis, T. P.
Vassilakopoulos, P. Korkolopoulou, F. N. Kontopidou, M. Atha-
nasoulia, X. Yiakoumis, C. Kalpadakis, G. Georgiou, S. Masour-
Figure 3. Percentage of damaged cellular DNA incubated with various
concentrations of 3 at 378C for 24 h (dark gray) or at 418C for 2 h
followed by 22 h at 378C (light gray).
7126
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7124 –7127

idis, M. Moschogiannis, P. Tsirkinidis, V. Pappis, S. I. Kokoris,
M. P. Siakantaris, P. Panayiotidis, M. K. Angelopoulou, Leuk.
Lymphoma 2011, 52, 383 – 393.
Kohn, Biochim. Biophys. Acta Gene Struct. Expression 1986,
868, 71 – 76.
[13] a) I. Boldogh, G. Roy, M.-S. Lee, A. Bacsi, T. K. Hazra, K. K.
Bhakat, G. C. Das, S. Mitra, Toxicology 2003, 193, 137 – 152;
b) A. Minnock, L.-S. Lin, J. Morgan, S. D. G. Crow, M. J. Waring,
L. Sheh, Bioconjugate Chem. 2001, 12, 870 – 882.
[14] a) A. Collins, Mol. Biotechnol. 2004, 26, 249; b) R. R. Tice, E.
Agurell, D. Anderson, B. Burlinson, A. Hartmann, H. Kobaya-
shi, Y. Miyamae, E. Rojas, J. C. Ryu, Y. F. Sasaki, Environ. Mol.
Mutagen. 2000, 35, 206.
[10] a) M. Wende, R. Meier, J. A. Gladysz, J. Am. Chem. Soc. 2001,
123, 11490; b) M. Wende, J. A. Gladysz, J. Am. Chem. Soc. 2003,
125, 5861; c) F. O. Seidel, J. A. Gladysz, Adv. Synth. Catal. 2008,
350, 2443.
[11] a) A. Gilman, F. S. Philips, Science 1946, 103, 409 – 436; b) P. B.
Farmer, Pharmacol. Ther. 1987, 35, 301 – 358; c) L. F. Povirk,
D. E. Shuker, Mutat. Res. Rev. Genet. Toxicol. 1994, 318, 205 –
226.
[15] T. Kumaravel, B. Vilhar, S. Faux, A. Jha, Cell Biol. Toxicol. 2009,
25, 53.
[12] a) K. W. Kohn, J. A. Hartley, W. B. Mattes, Nucleic Acids Res.
1987, 15, 10531 – 10549; b) W. B. Mattes, J. A. Hartley, K. W.
Angew. Chem. Int. Ed. 2011, 50, 7124 –7127
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7127