EXAFS and IR structural study of platinum-based anticancer drugs' degradation by diethyl dithiocarbamate

Article abstract of DOI:10.1021/ic051904u

Platinum compounds constitute a discrete class of DNA-damaging anticancer drug agents, including cisplatin, carboplatin, and oxaliplatin. The toxicity of such drugs raises the problem of waste detoxification. Diethyl dithiocarbamate (DDTC) is recommended by the World Heath Organization (WHO) for the destruction of cisplatin, but the degradation product has not been structurally characterized. This paper deals with the extended X-ray absorption fine structure (EXAFS) and IR structural study of the reaction products of DDTC with cisplatin, carboplatin, and oxaliplatin. Cisplatin and carboplatin give the same reaction product: Pt(DDTC)₂. In the case of oxaliplatin, we observed the formation of [(diaminocyclohexane)(DDTC)Pt(II)]. In all cases, the replacement of labile ligands by strong ligands should lead to inactive compounds. Our results suggest that the WHO inactivation protocol might be extended to carboplatin and oxaliplatin. Nevertheless, this should be validated by toxicity tests of the degradation products.

Full text of DOI:10.1021/ic051904u

Inorg. Chem. 2006, 45, 3393−3398
EXAFS and IR Structural Study of Platinum-Based Anticancer Drugs’
Degradation by Diethyl Dithiocarbamate
Diane Bouvet,*,† Alain Michalowicz,† Sylvie Crauste-Manciet,‡,§ Denis Brossard,‡,§ and Karine Provost†
Laboratoire de Physique Structurale des Mole´cules et Mate´riaux, UniVersite´ Paris XII,
61 aVenue du Ge´ne´ral De Gaulle, 94010 Cre´teil Cedex, France, Laboratoire de Pharmacie
Gale´nique, UniVersite´ Paris V, 75006 Paris, France, and SerVice de Pharmacie,
CHI Poissy Saint Germain en Laye, 78105 Saint Germain en Laye, France
Received November 3, 2005
Platinum compounds constitute a discrete class of DNA-damaging anticancer drug agents, including cisplatin,
carboplatin, and oxaliplatin. The toxicity of such drugs raises the problem of waste detoxification. Diethyl
dithiocarbamate (DDTC) is recommended by the World Heath Organization (WHO) for the destruction of cisplatin,
but the degradation product has not been structurally characterized. This paper deals with the extended X-ray
absorption fine structure (EXAFS) and IR structural study of the reaction products of DDTC with cisplatin, carboplatin,
and oxaliplatin. Cisplatin and carboplatin give the same reaction product: Pt(DDTC)2. In the case of oxaliplatin, we
observed the formation of [(diaminocyclohexane)(DDTC)Pt(II)]. In all cases, the replacement of labile ligands by
strong ligands should lead to inactive compounds. Our results suggest that the WHO inactivation protocol might be
extended to carboplatin and oxaliplatin. Nevertheless, this should be validated by toxicity tests of the degradation
products.
Introduction
Platinum compounds constitute a discrete class of DNA-
damaging anticancer agents, and their clinical uses are now
widespread. The first platinum coordination complex intro-
duced in 1972 as a cytotoxic agent was cisplatin.1,2 Clinical
experience has shown that cisplatin is highly nephrotoxic
and presents intrinsic and/or acquired tumor resistance,
leading to the development of other platinum compounds.
Carboplatin [diamine(1,1-cyclobutanedicarboxylato(2-)-
O,O′)-platinum(II)] is a second generation platinum antitumor
drug used in the same type of cancers as cisplatin.3 In this
platinum(II) complex, the metallic ion is bonded in the cis
position to two amines and chelated by a cyclobutanedicar-
boxylate ligand, as shown in Figure 1 b.
Figure 1. Formulas of (a) cisplatin, (b) carboplatin, and (c) oxaliplatin.
antitumoral drug approved for metastatic carcinoma treat-
ments.4 Oxaliplatin is formed by a platinum(II) ion chelated
by two bidentate ligands: one diaminocyclohexane (DACH)
and one oxalate (ox), as shown in Figure 1c.
Carboplatin has been shown to be less toxic than cisplatin
and equally active.3,5 However, it presents a cross-resistance
with the first generation drug. In preclinical studies, oxali-
platin has demonstrated efficacy on a broad spectrum of
experimental tumors, including cisplatin- and carboplatin-
resistant cell lines.4,6
Oxaliplatin cis-[(1R,2R)-1,2-cyclohexanediamine-N,N′]-
[oxalato(2-)-O,O′]platinum(II) is a third-generation platinum
* To whom correspondence should be addressed. E-mail: bouvet@
univ-paris12.fr.
-
Sodium diethyl dithiocarbamate (DDTC) [(C2H5)2NCS2 -
† Laboratoire de Physique Structurale des Mole´cules et Mate´riaux,
Universite´ Paris XII.
Na+] is recommended by the World Health Organization for
‡ Laboratoire de Pharmacie Gale´nique, Universite´ Paris V.
§ Service de Pharmacie, CHI Poissy Saint Germain en Laye.
(1) Rosenberg, B.; Camp, L. V.; Krigas, T. Nature 1965, 205, 698.
(2) Jamieson, E. R.; Lippard, S. J. Chem. ReV. 1999, 99, 2467.
(3) Micetich, K. C.; Barnes, D.; Erickson, L. C. Cancer Res. 1985, 45,
4043.
(4) Graham, M. A.; Lockwood, G. F.; Greenslade, D.; Brienza, S.; Bayssas,
M.; Gamelin, E. Clin. Cancer Res. 2000, 6, 1205.
(5) Kelland, L. R.; Sharp, S. Y.; O’Neill, C.; Raynaud, F. I.; Beale, P. J.;
Judson, I. R. J. Inorg. Biochem. 1999, 77, 111.
(6) Desoize, B.; Madoulet, C. Crit. ReV. Oncol./Hematol. 2002, 42, 317.
10.1021/ic051904u CCC: $33.50
Published on Web 03/22/2006
© 2006 American Chemical Society
Inorganic Chemistry, Vol. 45, No. 8, 2006 3393

Bouvet et al.
the destruction of cisplatin in different wastes, such as
glassware, aqueous solutions, or spills,7 on the basis of the
absence of mutagenic activity of the residue. Furthermore,
pharmacological studies have shown that the salt NaDDTC
is effective in reducing several kinds of nephrotoxicity when
administered just after a cisplatin injection.8,9 Nevertheless
the structures of the degradation products are still unknown.
Moreover, to our knowledge, there is no similar recom-
mendation for other platinum compounds such as carboplatin
and oxaliplatin. We have thus undertaken the structural study
of these three drugs’ reaction products with DDTC.
Up to now, we have not obtained any crystal suitable for
an X-ray diffraction study. Thus, we applied X-ray absorption
spectroscopy (XAS), which has already proved to be a
reliable technique, for studying the platinum local structure
in similar structural problems.10-12 The extended X-ray
absorption fine structures (EXAFSs) of the three native drugs
and their degradation products with chloride ions have
already been studied,12 and their XAS features are fully
understood. Following this previous work, this paper deals
with the structural studies of the three drugs’ degradation
by DDTC.
of pure product and cellulose. The quantities of products were
calculated in order to obtain an edge jump ∆µx near 1, with
a total absorbance after the edge µx(E) < 2.
Precipitate spectra were recorded in transmission mode
at LURE (the French synchrotron radiation facility), with
the following experimental conditions: LIII Pt edge and
energy range 11 450-12 600 eV. Precipitates with carbo-
platin and oxaliplatin spectra were recorded on the EXAFS13
station, with a Si(111) channel-cut monochromator. Precipi-
tate with cisplatin spectrum was recorded on the EXAFS2
station, with a Si(111) double crystal monochromator. The
EXAFSspectrawereextractedusingthestandardprocedures13-15
available in the program “Exafs pour le Mac”,16 including
spline-smoothing atomic absorption determination, energy-
dependent normalization, and fast Fourier transform (FT) (E0
) 11 560 eV, FT range ) 2-14 Å-1). Each spectrum was
recorded three times and averaged.
Since the three native drugs have already been character-
ized, they were used as model compounds. Structural models
for unknown compounds are based on the methodology
established for the known model compounds.12
EXAFS modeling of the complete structures was per-
formed in three steps:
Experimental Section
(i) Fourier filtering of the main peak and fitting of the
EXAFS contribution of the neighbors directly bonded to
platinum. The aim of this preliminary fit is to determine the
type and number of first Pt(II) neighbors.
(ii) Construction of the molecular models with the
Chem3D program using characteristic crystallographic dis-
tances and the knowledge of the ligand structures. Preparation
of the resulting FEFF17 input files with the code CRYS-
TALFF.18
(iii) Ab initio EXAFS modeling with FEFF7. Sorting of
the most important single and multiple scattering paths.
Equivalent paths are grouped in order to fit with a single
amplitude-phase function. Several models of ligand binding
are tested and, when parameters are physically acceptable,
quantitatively compared using the statistical F-test.19,20
All the fits of steps (i) and (iii) are performed with Round
Midnight code16 by comparing the experimental spectra to
the EXAFS standard formula (eq 1).
Samples Preparation. Cisplatin, carboplatin, and oxali-
platin were obtained from pharmaceutical preparations,
respecively, from Qualimed, Merck, and Sanofi. Carboplatin
and oxaliplatin were diluted to 10 mg/mL and cisplatin was
diluted in microfiltrated water to 1 mg/mL, with respect to
their solubilities. Each compound is mixed in solution with
a large excess of diethyl dithiocarbamate (0.33 M). Thus,
the reaction is expected to be complete.
For 1 mL of each solution, a yellow precipitate im-
mediately appears under these conditions. The precipitates
were washed three times in water, separated by centrifugation
at 8000 rpm, and dried in an oven at 60 °C for 2 days before
XAS measurement. The XAFS spectra of the supernatant
liquid do not show any Pt(II)LIII edge jump within the error
bars. This proves that >99% of platinum has precipitated.
Infrared Measurements. About 1-2 mg of each degra-
dation product of cisplatin, carboplatin, and oxaliplatin was
prepared as KBr pellets. The mixed products were com-
pressed under vacuum in order to make the pellets transparent
to infrared radiation up to 400 cm-1. Each pellet spectrum
(including the KBr reference pellet) was recorded with an
IRTF Perkin-Elmer spectrometer BX II, in the range 6000-
450 cm-1, with a 2 cm-1 resolution.
Ni
2
2
2
∑
i
i
kø(k) ) -S0
|fi(k,Ri)| e-2σ k e-2R /λ(k) sin[2kRi +
Ri2
i
2δ1(k) + ψi(k)] (1)
The sum is done on the most important selected single and
2
multiple scattering paths. S0 is the amplitude reduction
XAS Measurements and Analysis. Precipitates were
prepared as compressed pellets of a homogeneous mixture
(13) Teo, B. K. Inorganic Chemistry concepts, EXAFS: Basic Principles
and Data analysis; Berlin, 1986.
(7) Castegnaro, M. OMS-IARC Sci. Publ. 1985, 73, 97.
(8) Reedjik, J. Chem. ReV. 1999, 99, 2499.
(14) Koenigsberger, D. C., Prins, R., Eds. John Wiley: New York, 1988.
(15) Lytle, F. W.; Sayers, D. E.; Stern, E. A. Physica B 1989, 158, 701.
(16) Michalowicz, A. J. Phys. IV 1997, 7, 235.
(9) Bodenner, D. L.; Dedon, P. C.; Keng, P. C.; Borch, R. F. Cancer
Res. 1986, 46, 2745.
(17) Rehr, J. J.; Zabinski, S. I.; Alberts, A. C. Phys. ReV. Lett. 1992, 69,
(10) Curis, E.; Provost, K.; Nicolis, I.; Bouvet, D.; Benazeth, S.; Crauste-
Manciet, S.; Brion, F.; Brossard, D. New J. Chem. 2000, 24, 1003.
(11) Curis, E.; Provost, K.; Bouvet, D.; Nicolis, I.; Crauste-Manciet, S.;
Brossard, D.; Benazeth, S. J. Synchrotron Radiat. 2001, 8, 716.
(12) Bouvet, D.; Provost, K.; Crauste-Manciet, S.; Curis, E.; Nicolis, I.;
Brossard, D.; Michalowicz, A. J. Synchrotron Radiat. 2005, in press.
3397.
(18) Provost, K.; Champloy, F.; Michalowicz, A. J. Synchrotron Radiat.
2001, 8, 1109.
(19) Joyner, R. W.; Martin, K. J.; Meehan, P. J. J. Phys. 1987, C20, 4005.
(20) Michalowicz, A.; Provost, K.; Laruelle, S.; Mimouni, A.; Vlaic, G. J.
Synchrotron Radiat. 1999, 6, 233.
3394 Inorganic Chemistry, Vol. 45, No. 8, 2006

Platinum-Based Anticancer Drugs’ Degradation by DDTC
factor, and Ni is the number of equivalent scattering paths
with an effective distance Ri. λ(k) is the mean-free path of
the photoelectron. σi is the Debye-Waller coefficient,
characteristic of the distance distribution width for path i.
δ1(k) is the central atom phase shift, and |fi(k)| and ψi(k) are
the amplitude and phase for the neighbor scattering factors,
calculated by FEFF7 code. Since theoretical phase shifts and
amplitudes were used, it is necessary to fit the energy
threshold E0. The photoelectron mean-free path is set to the
empirical formula (eq 2),
4
1
Γ
η
k
λ(k) ) k +
(2)
[
( )
]
Values for Γ and η are set to η ) 3.1 and Γ ) 0.7, as
established in our previous study.12 In the modeling of the
filtered first coordination sphere, we have fitted separated
Debye-Waller (DW) factors σi for each type of atom. For
the fit of the complete spectra, it was necessary to limit the
total number of fitted parameters in order to limit the
correlations and improve the fitting statistics. Thus, it was
decided to fit a global DW factor for all the shells. In all
cases, a global ∆E0 was refined.
Figure 2. Comparison of EXAFS spectra before and after reaction between
DDTC and cisplatin, carboplatin, or oxaliplatin.
The goodness of fit, or quality factor, is QF ) ∆ømin 2/ν,
2
where ∆ømin is the minimum value of the statistical ∆østat
(eq 3) and the degree of freedom ν ) Nind - Npar, where
N
ind is the number of independent points and Npar is the total
number of fitted parameters.
Nind [køth(i) - kø exp(i)]2
2
∆østat
)
(3)
∑
ꢀi2
Npt
i
In eq 3, Npt is the number of experimental points and ꢀi2 is
the experimental error. Statistical errors on the average
spectra and error bars of the fitted parameters are evaluated
as recommended by the IXS Standard and Criteria sub-
committee.21
The EXAFS spectra are fitted in the range 3-14 Å-1. The
R-space fitted range is 1-3 Å for the first shell fits and 1-5
Å for the complete spectra. The numbers of independent
points Nind ) 2∆k∆R/π are, respectively, 14 and 28. The
average experimental error is evaluated to be ꢀ ) 0.015.
Figure 3. Comparison of Fourier transform amplitudes before and after
reaction between DDTC and cisplatin, carboplatin, or oxaliplatin.
For cisplatin, the apparent modification is represented by
an important raise of the FT amplitude without any signifi-
cant change of the average Pt-ligand distance. On the
contrary, both carboplatin and oxaliplatin first shell radial
distribution functions are shifted to a greater distance. In
carboplatin, this distance shift is accompanied by an impor-
tant increase of the amplitude, while this amplitude is
apparently not significantly affected for oxaliplatin.
The FT spectra of the three modified compounds show a
common new feature, when compared to the native one: a
relatively intense peak appears around 4 Å. In the case of
oxaliplatin, this outbreak comes with the fading of a peak at
3.5 Å, which has been previously shown to be the signature
of the oxalate ligand.12
Results and Discussion
Experimental Spectra. The EXAFS experimental signals
of the three precipitates are presented in Figure 2. Signals
(a), (b), and (c) correspond to the products of reaction
between and cisplatin, carboplatin, and oxaliplatin, respec-
tively. As a control, signals of the three drugs before reaction
are represented in dotted lines. The corresponding Fourier
transforms (without phase-shift correction) are presented in
Figure 3. The evolution of the EXAFS spectra and their
Fourier transforms for the three drugs upon degradation by
DDTC reflects the huge changes induced in the Pt(II)
coordination sphere by this sulfur ligand.
At least, the EXAFS spectra and Fourier transforms of
DDTC-degraded cisplatin and carboplatin are superimposable
within the error bars: both compounds should be modeled
with the same short-range order. On the contrary, the
(21) IXS. Report on errors in EXAFS fitting. http://ixs.iit.edu/
subcommittee_reports/sc/ (accessed March 16, 2006).
Inorganic Chemistry, Vol. 45, No. 8, 2006 3395

Bouvet et al.
Table 1. Evolution of the Quality Factor (QF) Versus the Composition
of the PtN4-xSx Coordination Sphere (Best Corresponds to the Lowest
QF)
spectrum and the local structure of the oxaliplatin degradation
product are significantly different.
These qualitative observations can be rationalized in order
to guide the first guess of the quantitative models and
refinements developed in the modeling section. The native
coordination sphere of Pt(II) in cisplatin is composed of two
nitrogen and two chlorine atoms, while both native carbo-
platin and oxaliplatin have a cis-PtN2O2 coordination sphere.
Adjacent atoms in the periodic table, such as sulfur and
chlorine on one hand and nitrogen and oxygen on the other
hand, are known to have very close EXAFS signals.
Moreover, the Pt-Cl and Pt-S bond length found in the
literature are in the same range (2.2-2.4 Å), while the
common range for both Pt-N and Pt-O bond lengths is
1.95-2.05 Å. The replacement of a chlorine atom by a sulfur
atom should not change significantly the corresponding
EXAFS signal, while the replacement of a light atom (N/O)
should induce a huge change.
PtN4-xSx
cisplatin/carboplatin
QF
oxaliplatin
QF
x
0
1
2
3
4
30.5
14.0
3.80
3.27
1.94
7.52
1.48
1.41
2.05
2.79
Table 2. Fit Parameters of the First Coordination Shell for the
Precipitates
path
N
σ × 102 (Å)
Cisplatin/Carboplatin and DDTC
R (Å)
Rh (Å)
Pt-S
4
5.6 (2)
2.31 (1)
RXR ) 2.32
residual ) 3.23 × 10-2, Npar ) 2, QF ) 1.94, ∆E ) 7.5 eV
Oxaliplatin and DDTC
Pt-N/O
Pt-S
2
2
5.5 (4)
5.4 (3)
2.05 (1)
2.29 (1)
RXR ) 2.04
RXR ) 2.32
residual ) 5.36 × 10-2, Npar ) 4, QF ) 1.41, ∆E ) 7.5 eV
In the case of DDTC-degraded cisplatin, the doubling of
the main peak amplitude suggests that the two nitrogens, at
least, must have been substituted by sulfur atoms, leading
to a PtS4 or a PtS2Cl2 coordination sphere. Because Cl-
ligands are more labile than amines, the first solution is much
more likely. Furthermore, the carboplatin degradation product
has been obtained in the absence of any chloride ions. Thus,
the platinum environment can only be composed of nitrogen,
oxygen, and sulfur atoms. Because DDTC-degraded cisplatin
and carboplatin Fourier transforms are superimposable up
to 5 Å, we expect that, in both products, the Pt(II) ion has
a PtS4 first coordination sphere.
In the case of the oxaliplatin degradation product, no huge
increase of the main peak in the Fourier transform was
observed. Because both ligands in oxaliplatin are bidentate,
we expect that only one ligand may have been displaced,
leading to a Pt(N/O)2S2 coordination sphere. The absence
of the oxalate signature in the DDTC-degraded oxaliplatin
spectrum suggests that the oxalate is displaced while the
DACH remains bonded to the Pt(II) center.
EXAFS Modeling of the First Coordination Sphere. As
quoted before, the coordination spheres of cisplatin (or
carboplatin) and oxaliplatin are drastically modified upon
reaction with DDTC. In both cases, the type and the number
of atoms in the first shell were determined using the method
recommended by Penner-Hahn and co-workers22 and modi-
fied by Michalowicz et al.20 A PtN4-xSx model has been
tested, where x is a whole number varying from 0 to 4. For
each value of x, the Debye-Waller factor σ and bond lengths
R were fitted. The goodness of fit was estimated using the
quality factor QF. Table 1 presents the variation of quality
factor versus x in the fitting of the first coordination sphere
of cisplatin/carboplatin and oxaliplatin degradation products
with a PtN4-xSx model. The models and the parameters of
the best simulations are reported in Table 2.
For oxaliplatin DDTC precipitate, the best fit is obtained
with two light atoms (O or N) and two sulfur atoms. These
quantitative results confirm the qualitative assumptions
presented in the previous section. As expected, the platinum
environment after reaction presents a huge difference be-
tween cisplatin/carboplatin and oxaliplatin. All fitted dis-
tances are coherent with PtN or PtS crystallographic distances
found in the literature for similar complexes.23-25
Structural Characterization of the Pt-DDTC Bonding
Mode by IR Spectroscopy. The reaction of the DDTC
ligand with cisplatin and carboplatin leads to the formation
of four Pt-S bonds, while two Pt-S bonds are involved in
the degradation product of oxaliplatin. Structures of different
dithiocarbamate (DTC) metal complexes have proved that
DTC can bind the metal ion either in a monodentate mode
or in a bidentate one.24-28 Nevertheless, the EXAFS refine-
ment of the first coordination sphere is unable to distinguish
both modes of fixation.
To discriminate between monodentate and bidentate
DDTC binding on Pt, we have decided to examine the IR
spectra of the degradation product. We have restricted our
IR analysis to the C-S vibrational IR signal of these
compounds. Nakamoto29 and Manav et al.28 have shown that
the C-S vibration band around 1000 cm-1 in the IR spectra
of DTC metal complexes is a good way to distinguish both
binding modes. When the Pt-DTC binding is monodentate,
the two C-S vibration modes are different and the elongation
mode should occur as a doublet separated by 20 cm-1 28. On
the contrary, the bidentate Pt-DTC binding mode is sym-
(23) Hansson, C.; Kukushkin, V. Y.; Lo¨vqvist, K.; Yong, S.; Oskarsson,
A° . Acta Crystallogr. 2003, C59, m432.
(24) Heath, G. A.; Hockless, D. C. R.; Prenzler, P. D. Acta Crystallogr.
1996, C52, 537.
(25) Christidis, P. C.; Rentzeperis, P. J. Acta Crystallogr. 1979, B35, 2543.
(26) Faraglia, G.; Fedrigo, M. A.; Sitran, S. Transition Met. Chem. 2002,
27, 200.
(27) Marzano, C.; Bettio, F.; Baccichetti, F.; Trevisan, A.; Giovagnini, L.;
Dregona, D. Chem.-Biol. Interact. 2004, 148, 37.
In the case of cisplatin or carboplatin degradation products,
the first sphere signal is fitted with a four-sulfur-atom model.
(28) Manav, N.; Mishra, A. K.; Kaushik, N. K. Spectrochim. Acta, Part A
2004, 60, 3087.
(29) Nakamoto, K. Infrared and Raman spectra of inorganic and coordina-
tion compounds, 3th ed.; John Wiley & Sons: New York, 1978.
(22) Clark-Baldwin, K.; Tierney, D. L.; Govindaswamy, N.; Gruff, E. S.;
Kim, C.; Koch, S. A.; Penner-Hahn, J. E. J. Am. Chem. Soc. 1998,
120, 8401.
3396 Inorganic Chemistry, Vol. 45, No. 8, 2006

Platinum-Based Anticancer Drugs’ Degradation by DDTC
Figure 6. Most important scattering paths of the carboplatin degraded
EXAFS signal.
Figure 7. Comparison between experimental and theoretical signals for
carboplatin degraded by DDTC: (a) EXAFS spectrum and (b) corresponding
FT.
Figure 4. Infrared spectrum of (a) cisplatin (dotted lines)/carboplatin-
DDTC degradation products and (b) oxaliplatin-DDTC degradation product.
Figure 8. Model for precipitate obtained for oxaliplatin.
Figure 5. Model for precipitates obtained with carboplatin and cisplatin.
Table 3. Fit Parameters of the Degradation Product of Carboplatin or
Cisplatin in the Presence of DDTC
path
N
σ × 102 (Å)
R (Å)
Pt-S
Pt-C
4
2
2
4
4.9 (1)
4.9 (1)
4.9 (1)
4.9 (1)
2.31 (1)
2.84 (2)
4.12 (1)
4.59 (1)
Pt-C-N′
Figure 9. Most important scattering path for the oxaliplatin degraded
EXAFS signal
Pt-S-Pt-S
residual ) 1.84 × 10-2, Npar ) 6, QF ) 2.17, ∆E ) 7.7 (1) eV
geometry of the DDTC bidentate ligand.24,25 All other
parameters present physical acceptable values.
metric and should be detected as a single C-S vibration
band.
The comparison between experimental and refined FEFF-
simulated EXAFS spectra of cisplatin and DDTC degradation
product, and their respective FT, is presented in Figure 7.
The four-shell model detailed above fits to the complete
EXAFS spectrum within the error bars. Moreover, the FT
contribution around 4 Å is well-reproduced if we take into
account both multiple scattering groups: Pt-C-N′ and
Pt-S-Pt-S. The 4 Å multiple scattering signals of carbo-
platin and cisplatin degraded by DDTC are identical. This
result definitely rules out the possibility of a PtS2Cl2
structure: in such a hypothetical model, the 4 Å signal should
be less intense.
EXAFS Model for DDTC-Degraded Oxaliplatin. As
discussed above, the signature of oxalate ligand binding at
3.5 Å is missing after reaction. This suggests that this ligand
is displaced by DDTC. We have already shown in the IR
section that the Pt-DDTC bond is bidentate. According to
these results, we propose that, in the degradation complex,
the platinum ion is bonded to both bidentate DACH and
DDTC ligands, as shown in Figure 8.
Parts a and b of Figure 4 display, respectively, the
superimposed IR spectra of DDTC-degraded cisplatin and
carboplatin and the oxaliplatin degradation product IR
spectrum. As expected, the cisplatin and carboplatin IR
spectra are very similar, while some differences appear for
oxaliplatin. In both figures, the elongation C-S characteristic
vibration mode is a singlet. Thus, we can assume that the
Pt-DDTC ligand binding mode is bidentate for the three
compounds.
EXAFS Model for DDTC-Degraded Carboplatin/Cis-
platin. Figure 5 presents the only acceptable model with a
4S coordination sphere and a bidentate fixation of the ligand.
The complete EXAFS spectrum of this reaction product can
be modeled ab initio up to 5 Å using FEFF7 code. An
analysis of the most important scattering-path contributions
showed that it was possible to limit the refinement of the
platinum neighborhood to four groups of paths, as shown in
Table 3. The two first groups correspond to single scattering
paths Pt-S and Pt-C. The third group takes into account
the multiple scattering path through the alignment PtCN′.
In the later one, we used multiple scattering paths through
the platinum ion. These scattering paths are represented in
Figure 6. After XAFS refinement, the final model is coherent
with the already-fitted Pt-S distance and with the known
The oxaliplatin degradation product EXAFS spectrum can
be reproduced with a five-shell FEFF model. Figure 9
presents the scattering-paths scheme. The first three shells
correspond to single scattering paths Pt-N, Pt-S, and Pt-C
(one for DDTC and two for DACH). The two later ones
take into account multiple scattering paths, respectively,
Inorganic Chemistry, Vol. 45, No. 8, 2006 3397

Bouvet et al.
to a centro-symmetric square planar PtS4 structure. The
replacement of the two Pt-amine and the two labile Pt-
chlorine bonds in a cis geometry by four Pt-S bonds may
explain the absence of anticancer activity and toxicity. The
resulting complex might not be hydrolyzed and fixed
specifically on the DNA chain.
The local EXAFS structure of the carboplatin-DDTC
degradation product is the same as that for cisplatin-DDTC.
Furthermore, their IR spectra are identical. Thus, we can
assume that both drugs lead to the same final product. This
result suggests that the detoxification process already recom-
mended for cisplatin might be extended to carboplatin.
The structural feature of oxaliplatin-DDTC is quite
different: DDTC substitutes only the oxalate ligand. We have
shown elsewhere that the oxalate ligand is labile in Pt(II)
complexes.12 The present work confirms that it is more labile
than DACH.30 The binding of oxaliplatin to DNA is assumed
to be preceded by the hydrolysis of oxalate. Its substitution
by a strong ligand like DDTC should also lead to a nontoxic
and nonactive compound.
In summary, our results suggest that DDTC might be a
good candidate for the inactivation of carboplatin and
oxaliplatin. Nevertheless, this should be validated by activity
and toxicity tests.
Various nucleophilic sulfur ligands are proposed in dif-
ferent biological and clinical processes involving platinum-
based anticancer drugs. The EXAFS structural study of these
reactions, either precipitates or still in solution, is under
investigation and will be published later.
Figure 10. Comparison between experimental and theoretical signals for
oxaliplatin degraded by DDTC: (a) EXAFS spectrum and (b) corresponding
FT.
Table 4. Fit Parameters of the Degradation Product of Oxaliplatin in
the Presence of DDTC
path
N
σ × 102 (Å)
R (Å)
Pt-N
Pt-S
Pt-C
2
2
3
2
2
4.77 (4)
4.77 (4)
4.77 (4)
4.77 (4)
4.77 (4)
2.04 (1)
2.28 (1)
2.80 (2)
4.13 (3)
4.26 (3)
Pt-C-N′
Pt-N-Pt-S
residual ) 2.96 × 10-2, Npar ) 7,QF ) 1.45, ∆E ) 6.4 (3) eV
through the PtCN′ alignment and through the central atom.
Compared to these multiple scattering paths, the EXAFS
signals corresponding to the DACH ligand outer shells can
be neglected.
It appears that the common feature observed around 4 Å
in the three modified drug spectra is essentially due to the
Pt-C-N′ multiple scattering focusing effect. The best
simulation parameters are presented in Table 4. As for
cisplatin/carboplatin, the complete model distances are in
good agreement with the already fitted Pt-S and Pt-N
distances and with the known geometry of the DDTC
bidentate ligand. With such parameters, the EXAFS spectrum
and its Fourier transforms are well-reproduced (Figure 10).
Acknowledgment. We would like to thank Pr. Simone
Benazeth, Dr. Jacques Moscovici, Dr. Emmanuel Curis, and
Dr. Ioannis Nicolis for data collection at LURE, and Dr.
Nicolas Fray and Dr. Herve´ Cottin for IR data collection,
as well as all their helpful discussions. The financial
support by the French Ministry of Education is gratefully
acknowledged.
Conclusion
The detoxification of cisplatin by the DDTC process is
already validated. Our EXAFS and IR study completes the
knowledge of this process by the structural characterization
of the final nontoxic compound, despite the difficulty to
obtain monocrystals. The Pt(II) coordination sphere has been
completely substituted by two bidentate sulfur bonds, leading
IC051904U
(30) Wong, E.; Giandomenico, C. M. Chem. ReV. 1999, 99, 2451.
3398 Inorganic Chemistry, Vol. 45, No. 8, 2006