Comparative investigation of the DNA inter-strand crosslinks induced by ACNU, BCNU, CCNU and FTMS using high-performance liquid chromatography- electrospray ionization tandem mass spectrometry

Article abstract of DOI:10.1016/j.ijms.2014.04.018

Chloroethylnitrosoureas (CENUs) are an important family of alkylating agents employed in the clinical treatments of cancer. They exert cytotoxicity by inducing DNA interstrand crosslinks (ICLs) between guanine and the complimentary cytosine, namely dG-dC crosslink. Many investigations have been performed on the DNA ICLs involved in the anticancer efficacy of CENUs, but no conclusive comparisons between these agents have been published. In this work, the levels of dG-dC crosslink in calf thymus DNA induced by four CENUs, including nimustine (ACNU), carmustine (BCNU), lomustine (CCNU) and fotemustine (FTMS), were quantitatively determined using HPLC-ESI-MS/MS. The obtained time-courses for the dG-dC crosslinking levels indicated that there is an induction period with very low crosslinking activity at the initial stage of the treatment by BCNU and CCNU. The induction period provides a convincing evidence for the presumed mechanism that the formation of dG-dC crosslinks was initiated by the monoalkylation of guanine followed by the second alkylation of the complimentary cytosine. The crosslinking activity of ACNU is remarkably higher than those of BCNU, CCNU and FTMS at all time points. The crosslinking activities of CENUs were found to be related to their stability in aqueous solution. ACNU has the shortest half-life among the four CENUs, but has highest crosslinking levels; on the contrary, CCNU has the lowest crosslinking activity with the longest half-life. Moreover, a correlation was found between the crosslinking activity and the anticancer efficiency. ACNU with the highest crosslinking activity showed the better survival gain for high-grade glioma than BCNU, CCNU and FTMS as reported in an epidemiological study. This suggests that dG-dC crosslink can possibly be employed as a potential biomarker for evaluating the anticancer efficiency of novel CENU drugs.

Full text of DOI:10.1016/j.ijms.2014.04.018

G Model
MASPEC151671–7
ARTICLE IN PRESS
International Journal of Mass Spectrometry xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Mass Spectrometry
journal homepage: www.elsevier.com/locate/ijms
1
2
3
Comparative investigation of the DNA inter-strand crosslinks induced
by ACNU, BCNU, CCNU and FTMS using high-performance liquid
chromatography–electrospray ionization tandem mass spectrometry
Lijiao Zhao∗, Lili Li, Jie Xu, Rugang Zhong
Q1
4
5
6
Beijing Key Laboratory of Environmental and Viral Oncology, College of Life Sciences and Bioengineering, Beijing University of Technology, Beijing 100124,
PR China
7
821
a r t i c l e i n f o
a b s t r a c t
9
10
11
12
13
14
Article history:
Chloroethylnitrosoureas (CENUs) are an important family of alkylating agents employed in the clinical
treatments of cancer. They exert cytotoxicity by inducing DNA interstrand crosslinks (ICLs) between
guanine and the complimentary cytosine, namely dG–dC crosslink. Many investigations have been per-
formed on the DNA ICLs involved in the anticancer efficacy of CENUs, but no conclusive comparisons
between these agents have been published. In this work, the levels of dG–dC crosslink in calf thymus DNA
induced by four CENUs, including nimustine (ACNU), carmustine (BCNU), lomustine (CCNU) and fotemus-
tine (FTMS), were quantitatively determined using HPLC–ESI-MS/MS. The obtained time-courses for the
dG–dC crosslinking levels indicated that there is an induction period with very low crosslinking activity at
the initial stage of the treatment by BCNU and CCNU. The induction period provides a convincing evidence
for the presumed mechanism that the formation of dG–dC crosslinks was initiated by the monoalkylation
of guanine followed by the second alkylation of the complimentary cytosine. The crosslinking activity
of ACNU is remarkably higher than those of BCNU, CCNU and FTMS at all time points. The crosslinking
activities of CENUs were found to be related to their stability in aqueous solution. ACNU has the shortest
half-life among the four CENUs, but has highest crosslinking levels; on the contrary, CCNU has the lowest
crosslinking activity with the longest half-life. Moreover, a correlation was found between the crosslink-
ing activity and the anticancer efficiency. ACNU with the highest crosslinking activity showed the better
survival gain for high-grade glioma than BCNU, CCNU and FTMS as reported in an epidemiological study.
This suggests that dG–dC crosslink can possibly be employed as a potential biomarker for evaluating the
anticancer efficiency of novel CENU drugs.
Received 28 January 2014
Received in revised form 24 April 2014
Accepted 24 April 2014
Available online xxx
15
16
17
18
19
20
Keywords:
Chloroethylnitrosoureas
DNA interstrand crosslinks
Quantitative analysis
HPLC–ESI-MS/MS
© 2014 Published by Elsevier B.V.
22
1. Introduction
achieve higher antineoplastic activity or lower toxicity. Fotemus-
tine (FTMS) is an active nitrosourea in metastatic melanoma, and it
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
23
24
25
26
27
28
29
30
31
32
33
Chloroethylnitrosoureas (CENUs) are an important family of
alkylating agents widely employed in the clinical treatment of
cancer, including Hodgkin’s disease, malignant melanoma, and var-
ious solid tumors. Especially, owing to the ability of crossing the
blood–brain barrier, CENUs are efficient chemotherapeutics for
brain tumor and other central nervous system neoplasms [1–4].
As listed in Table 1, nimustine (ACNU), carmustine (BCNU), lomus-
tine (CCNU) and semustine (MeCCNU) represent the typical CENUs
chemotherapeutics used in the clinical treatment of cancer [5–7].
In recent years, several novel CENUs chemotherapies were devel-
oped by modifying the structure of the moiety on the N3 atom to
was the first drug to show significant efficacy in brain metastases
[8,9]. The phosphoalanine group grafted on the N3 atom is highly
lipophilic and increase its ability of crossing the blood–brain barrier
[10]. 1-(2-Chloroethyl)-3-sarcosinamide-1-nitrosourea (SarCNU)
was selected for formulation by the National Cancer Institute
because of its therapeutic advantage in the treatment of malignant
glioma [11]. Ranimustine (MCNU), a derivative of CENU developed
and clinically used in Japan, showed excellent responses against
chronic myelogenous leukemia, polycythemia vera and thrombo-
cythemia [12,13].
CENUs are bifunctional alkylating agents, which exert cyto-
toxicity by inducing DNA interstrand crosslinks (ICLs) within the
complimentary guanine and cytosine base pair. This covalent
crosslink interferes with the normal DNA replication by preven-
ting the separation of the double strands, and finally leads to the
∗
Corresponding author. Tel.: +86 13910167017; fax: +86 01067391667.
E-mail address: zhaolijiao@bjut.edu.cn (L. Zhao).
Q2
http://dx.doi.org/10.1016/j.ijms.2014.04.018
1387-3806/© 2014 Published by Elsevier B.V.
Please cite this article in press as: L. Zhao, et al., Int. J. Mass Spectrom. (2014), http://dx.doi.org/10.1016/j.ijms.2014.04.018

G Model
MASPEC151671–7
ARTICLE IN PRESS
L. Zhao et al. / International Journal of Mass Spectrometry xxx (2014) xxx–xxx
2
Table 1
Chloroethylnitrosoureas used in clinical treatment or developed as preclinical therapeutics for cancer.
Proprietary name
Chemical structures
Indications
Nimustine (ACNU)
Carmustine (BCNU)
Brain tumors, small cell lung cancer and Hodgkin’s disease
Brain tumors and Hodgkin’s disease
Lomustine (CCNU)
Brain tumors
Brain tumors
Semustine (MeCCNU)
Fotemustine (FTMS)
SarCNU
Brain tumors and melanoma
Malignant glioma
Ranimustine (MCNU)
Tauromustine (TCNU)
Brain tumors, myeloma, malignant lymphoma, and leukemia
Myeloma, malignant glioma and lymphoma
Chlorozotocin (DCNU)
Pancreatic tumors
51
52
53
54
55
56
57
58
59
apoptosis of cancer cells [14–16]. The CENUs-induced DNA ICLs
were demonstrated to occur between the N1 site of guanine and the
N3 site of the complimentary cytosine. The supposed mechanism of
the formation of dG–dC crosslink is shown in Fig. 1. The chloroethyl
diazonium ion produced by the decomposition of CENUs alkylates
guanine on the O6 site to form O6-chloroethylguanine (O6-ClEt-
Gua) followed by further alkylation of the complimentary cytidine
on the N3 site via a cationic intermediate, N1,O6-ethanoguanine
[17–19]. The levels of dG–dC crosslink in cells are much lower than
the monoadducts induced by CENUs, such as N7-(2-hydroxyethyl)-
guanine and N7-(2-chloroethyl)-guanine [20–22]. However, there
is considerable correlation between the crosslinking activity and
the cytotoxicity of CENUs. Therefore, dG–dC crosslink is possibly
be used as a biomarker for evaluating the chemotherapeutic effect
of CENUs.
To understand the anticancer mechanism of CENUs and develop
more efficient chemotherapeutics, significant efforts have been
devoted to investigate the CENUs-induced DNA ICLs by in vitro or
60
61
62
63
64
65
66
67
68
Fig. 1. Supposed mechanism for the formation of dG–dC crosslinks induced by CENUs.
Please cite this article in press as: L. Zhao, et al., Int. J. Mass Spectrom. (2014), http://dx.doi.org/10.1016/j.ijms.2014.04.018

G Model
MASPEC151671–7
ARTICLE IN PRESS
L. Zhao et al. / International Journal of Mass Spectrometry xxx (2014) xxx–xxx
3
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
in vivo studies. Hayes et al. [23] determined the DNA ICLs induced
by FTMS and MeCCNU in linear pBR322 plasmid using agarose gel
electrophoresis, and observed that the coincubation of ellagic acid
reduced ICLs considerably. Penketh et al. [24] compared the levels
of ICL induced by CENUs and 1,2-bis(sulfonyl)hydrazine derivatives
(BSHs) in T7 DNA using fluorescent assay with Hochest 33258 as a
probe specifically recognizing the crosslinked double strand DNA.
Also using fluorescent assay, Ueda-Kawamitsu et al. [25] measured
the time course of DNA ICLs in L1210 cells treated with BCNU, and
observed that the percentage of crosslinks reached the maximum
after 6 h exposure and subsequentlydecreased presumablybecause
of DNA repair. Tong et al. [26] first isolated 1-[N3-deoxycytidyl],2-
[N1-deoxyguanosinyl]ethane (dG–dC crosslink) from DNA exposed
to BCNU. Bodell et al. [22] measured the levels of dG–dC crosslink
using HPLC, and observed a significant correlation between LD10 of
CENUs and the dG–dC crosslinking level. High performance liquid
chromatography–tandem mass spectrometry (HPLC–MS/MS) has
been frequently used for the quantitative analysis of DNA adducts
because of its high sensitivity, specificity and accuracy. Fischhaber
et al. [27] determined the levels of BCNU-induced dG–dC crosslink
using HPLC–MS/MS and provided the first direct evidence that
BCNU had no strong sequence preference for interstrand cross-
linking. In our previous work [28–30], dG–dC crosslinks induced
by MeCCNU in calf thymus DNA and in synthetic oligonucleotides
were determined by HPLC–MS/MS. The results indicated that the
dG–dC crosslink stayed at a relatively low level during the first 2 h of
the treatment and then underwent obvious increase. This provided
convincing evidence for the crosslinking mechanism proposed in
our theoretical studies [31,32], which suggested that the reaction
was initiated by the formation of the monoadduct followed by the
second alkylation on the complementary strand of DNA to form
crosslinks. Even though many previous studies were focused on the
DNA ICLs involved in the anticancer efficiency of CENUs, no con-
clusive comparisons between these agents have been published.
In this work, the levels of CENU-induced dG–dC crosslink were
determined using HPLC–ESI-MS/MS. Comparison was performed
between the levels of dG–dC crosslink in calf thymus DNA treated
by ACNU, BCNU, CCNU and FTMS, respectively.
in 4 mL of dry dioxane followed by incubation with 160 ␮L of 2-
fluoroethanol dioxane, 650 mg of triphenylphosphine and 400 ␮L
of diethylazodicarboxylate at room temperature for 1 h. Then 15 mL
of 5% NaHCO3 solution was added. The mixture was stirred at room
temperature for 15 min. The product was extracted with 30 mL of
methylene chloride for three times. The oily residue was dissolved
in 6 mL of methanol followed by addition of 30 mL of concentrated
ammonium hydroxide. The mixture was kept at 60 ◦C for 3 h to pro-
duce O6-(2-fluoroethyl)-2ꢀ-deoxyguanosine (O6-FEt-dGuo), which
was dissolved in methanol and purified by silica gel column chro-
matography (200–300 mesh) using ethyl acetate and petroleum
ether as the solvent. The obtained O6-FEt-dGuo was used as the
starting material for the following synthesis of dG–dC crosslink.
Twenty milligram of O6-FEt-dGuo were incubated with 5 mg of 2ꢀ-
deoxycytidine in 100 ␮L of DMSO at 55 ◦C for 20 days. The final
product was purified by HPLC with a 4.6 mm × 250 mm Zorbax
SB-C18 column (5 ␮m in particle size) to collect the fractions con-
taining dG–dC. The mobile phase consists of ammonium acetate
solution at the concentration of 10 mM (0.1% acetic acid, pH 6.8)
(A) and acetonitrile (B) with a flow rate of 1 mL/min. A gradient
of 5–10% buffer B in 20 min was employed with a linear gradient
to 30% buffer B over 10 min. Then an isocratic wash of 30% buffer
B was used for 3 min followed by a gradient down to 5% buffer B
over 2 min. The fraction between 20 and 22 min was collected. UV
detector was set at 258 nm. The synthesis of the internal standard,
15N3-dG-dC, was carried out using the same procedure as the unla-
beled dG–dC except 15N3-2ꢀ-deoxycytidine was used in the final
step. The final product was characterized by NMR, MS, IR and UV
spectroscopy. The data were consistent with the results obtained
previously [29,33].
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
2.3. Treatment of calf thymus DNA with CENUs
159
Calf thymus DNA was dissolved in phosphate-buffered saline
(10 mmol/L Tris–HCl, 50 mmol/L NaCl, 50 mmol/L NaH2PO4,
pH = 7.4) to a concentration of 0.5 mg/mL. ACNU was dissolved in
deionized water immediately prior to use and directly added to
8 mL of DNA solution to achieve the final concentrations of 1, 2, 4
and 8 mM, respectively. The reaction mixtures were incubated at
37 ◦C for 12 h in the dark. Aliquots of 400 ␮L solution were removed
from the reaction mixture at 1, 2, 3, 4, 6, 8, 10 and 12 h, respectively.
For each sample, DNA was precipitated by adding 800 ␮L of ice-cold
ethanol followed by centrifuge at 12,000 rpm for 5 min. The DNA
pellets were washed with 70% ethanol and then 100% ethanol. All
the DNA samples were dried with a stream of nitrogen and stored
at −20 ◦C until enzymatic hydrolysis. The treatments of calf thy-
mus DNA with BCNU, CCNU and FTMS, respectively, were carried
out using the same protocol as ACNU except that the drugs were
dissolved in ethanol, but not deionized water. For each time point,
the control DNA samples were incubated at the same conditions as
the treated DNA samples.
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
107
108
2. Experimental
2.1. Chemicals and materials
109
110
111
112
113
114
115
116
117
118
119
ACNU, BCNU, CCNU, FTMS, acetonitrile (HPLC grade), 2ꢀ-
deoxyguanosine, 2ꢀ-deoxycytidine, calf thymus DNA and snake
venom phosphodiesterase I were purchased from Sigma–Aldrich
(St. Louis, MO, USA). Nuclease S1, alkaline phosphatase (CIAP)
and deoxyribonuclease I were obtained from TaKaRa Biotech-
nology (Tokyo, Japan). 15N3-2ꢀ-deoxycytidine was purchased
from Cambridge Isotope Laboratories (Andover, MA, USA). All
other chemicals, reagents and solvents were purchased from
Sigma–Aldrich. Microcon YM-10 centrifugal columns were pur-
chased from Millipore (Billerica, MA, USA). Deionized water was
purified by a PALL deionizer.
2.4. Enzymatic hydrolysis of DNA
178
The concentration of DNA samples was determined before the
enzymatic hydrolysis. The obtained DNA pellets were redissolved in
200 ␮L of Tris–HCl buffer (10 mM, pH 7.0). The concentration of the
DNA solution was determined by the UV absorbance at 260 nm. One
OD260 corresponds to approximately 50 ␮g/mL for double-strand
DNA.
Each DNA sample was spiked with 15 ␮L of internal standard
15N3-dG-dC (1 ␮M) followed by enzymatic hydrolysis according
to the previously reported protocols [29,30]. DNA samples were
digested by four enzymes, including DNase I, nuclease S1, alka-
line phosphatase and snake venom phosphodiesterase I. Briefly,
the DNA solutions were first heated at 98 ◦C for 5 min and promptly
179
180
181
182
183
184
185
186
187
188
189
190
120
2.2. Synthesis of the standards
121
122
123
124
125
126
127
128
The synthesis of [l-(3-deoxycytidyl),2-(l-deoxyguanosinyl)]
ethane (dG–dC) and the internal standard 15N3-dG-dC were carried
out according to a previously published method [29,33], with some
modifications. Briefly, 325 mg of 2ꢀ-deoxyguanosine was incubated
with 1.8 mL of acetic anhydride in a mixture containing 18 mg
of 4-dimethylamiopyridine, 2.0 mL of triethylamine and 50 mL of
dry pyridine at 50 ◦C for 20 h. After crystallization and filtration,
the obtained N2,3ꢀ,5ꢀ-triacetyl-2ꢀ-deoxyguanosine was dissolved
Please cite this article in press as: L. Zhao, et al., Int. J. Mass Spectrom. (2014), http://dx.doi.org/10.1016/j.ijms.2014.04.018

G Model
MASPEC151671–7
ARTICLE IN PRESS
L. Zhao et al. / International Journal of Mass Spectrometry xxx (2014) xxx–xxx
4
191
192
193
194
195
196
197
198
199
200
201
202
chilled in an ice-bath for 10 min. Each DNA solution (200 ␮L) was
hydrolyzed by adding 180 units of DNase I (60 ␮L, buffered in
CH3COONa 20 mM, NaCl 150 mM, pH 5.0) and 300 units of nucle-
ase S1 (60 ␮L, buffered in CH3COONa 10 mM, NaCl 150 mM, ZnSO4
0.05 mM, pH 4.6). After incubation at 37 ◦C for 6 h, the mixture was
further incubated overnight at 37 ◦C with the addition of 30 units
(60 ␮L) of alkaline phosphatase and 10 mill units of phosphodieste-
rase I (5 ␮L) buffered in Tris–HCl 500 mM, MgCl2 10 mM (pH 9.0).
Finally, the DNA samples were filtered with molecular weight cen-
trifugal filters (Microcon YM-10) for HPLC–ESI-MS/MS analysis. A
buffer control without DNA was prepared for each set of samples
and processed as negative controls following the same procedure.
(solvent A) and acetonitrile (solvent B). The gradient was started
from an isocratic elution with 2% B for the first 5 min followed
by a linear increase to 80% B in 25 min. After being held at 80%
B for 3 min, the solvent composition was brought back to the ini-
tial composition of 2% B in the next 2 min and equilibrated at those
conditions for 30 min. The instrumental parameters of the mass
spectrometer were set as follows: spray voltage 4000 V; sheath
gas (nitrogen) pressure 50 psi; aux gas (nitrogen) pressure 15 psi;
capillary temperature 300 ◦C; and tube lens offset 89 V. Collision
energy was set to 20 eV using argon at 1.0 mTorr. The source CID
was set to 8 V. The fragmentation pattern of dG–dC and 15N3-
dG-dC was shown in Fig. S1 in the Supporting Information. The
amounts of dG–dC crosslink were quantified by selecting reaction
monitoring (SRM) with the transition of m/z 521→289 for dG–dC
and 524→292 for 15N3-dG-dC. Under the proposed HPLC condi-
tions, dG–dC standard and 15N3-dG-dC internal standard coeluted
at 22 min (see Fig. S2 in the Supporting Information).
According to previous work [29,30], the level of dG–dC crosslink
in DNA from cells were reported by the number of crosslinked dG
and dC in every 107 base pairs calculated by Eq. (1). In Eq. (1),
C refers to the determined concentrations of dG–dC crosslink; V
refers to the volume of the enzymatic digestion solution, which is
400 ␮L; C0 is the concentration of the calf thymus DNA measure by
UV; V0 is the volume of the DNA sample before enzymatic digestion,
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
203
2.5. Quantitation of dG–dC crosslink by HPLC–ESI-MS/MS
204
205
206
207
208
209
210
211
212
HPLC–ESI-MS/MS analysis was carried out on a Thermo TSQ
QUANTUM Discovery MAX triple quadrupole tandem mass spec-
trometer interfaced with Thermo Finnigan HPLC system (Thermo
Finnigan, San Jose, CA). The electrospray ionization (ESI) was per-
formed in the positive mode. The fractions of dG–dC crosslink were
separated with a 2.1 mm × 150 mm (5 ␮m in particle size) Zorbax
SB-C18 column (Agilent Technologies, Palo Alto, CA) and eluted
at a flow rate of 0.1 mL/min. The injection volume was 25 ␮L. The
mobile phase consisted of deionized water with 0.01% acetic acid
Fig. 2. SRM chromatograms of dG–dC crosslinks in the DNA digestion mixtures from (A) control samples, (B) ACNU-treated DNA, (C) BCNU-treated DNA, (D) CCNU-treated
DNA and (E) FTMS-treated DNA.
Please cite this article in press as: L. Zhao, et al., Int. J. Mass Spectrom. (2014), http://dx.doi.org/10.1016/j.ijms.2014.04.018

G Model
MASPEC151671–7
ARTICLE IN PRESS
L. Zhao et al. / International Journal of Mass Spectrometry xxx (2014) xxx–xxx
5
237
238
which is 200 ␮L; and M0 represents the average molecular weight
of the four deoxynucleotides (325 g/mol).
dG − dC crosslinks
C × V × 107
239
240
=
(1)
107 base pairs
C0 × V0/2M0
3. Results and discussion
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
Fig. 2 shows the SRM chromatograms of dG–dC and 15N3-dG-
dC in the DNA enzymatic hydrolysates from the control sample
(Fig. 2A), and DNA treated with ACNU (Fig. 2B), BCNU (Fig. 2C),
CCNU (Fig. 2D) and FTMS (Fig. 2E). In Fig. 2B–D, the retention times
for dG–dC in the digestion mixtures are about 22 min, and their
corresponding isotope labeled standard, 15N3-dG-dC, has the same
retention time. Fig. 2A indicates that there is no signal detected with
the SRM transition for dG–dC in the DNA hydrolysates from the
control samples. This indicates that there is no significant matrix
interference or contamination in the analyte channels from internal
standards or hydrolytic enzymes, so the specificity of the method
was acceptable.
Fig. 3 shows the time-course of the levels of dG–dC crosslinks
induced by the four CENUs. The corresponding values for the deter-
mined levels of dG–dC crosslinks at each time point are listed in
Table S1 in the Supporting Information. For the four CENUs, the
levels of dG–dC crosslink display dose-dependence with the con-
centration increased from 1 to 8 mM. The crosslinking levels show
a common increasing trend during 0–12 h, and reached a maxi-
mum at 12 h. For ACNU, the crosslinking levels at 12 h were 527,
1203, 1358 and 1445 dG–dC crosslinks/107 base pairs for 1, 2, 4
and 8 mM concentration, respectively. For BCNU, CCNU and FTMS,
the maximum crosslinking levels were 505, 263 and 585 dG–dC
crosslinks/107 base pairs, respectively, observed at 12 h for the
highest concentration of 8 mM. However, there is a significant dif-
ference between the patterns of the time-course of the four CENUs.
In Fig. 3B and C, there is an “induction period” in the initial stage of
the treatment by BCNU and CCNU, in which the dG–dC crosslinks
stay at relatively low levels. Approximately, the induction period is
2 h for BCNU, and 4 h for CCNU. Especially for CCNU, the induction
period is very apparent for all the four concentrations of the drug
with the crosslinking levels lower than 20 dG–dC crosslinks/107
base pairs. In our previous work [29], a similar induction period was
also observed in MeCCNU treated oligonucleotide duplexes. The
induction period provides a convincing evidence for the supposed
mechanism of the formation of dG–dC crosslinks induced by CENUs,
which postulates that the reaction was initiated by the formation
of the guanine monoadduct induced by chloroethyl cations arising
from the decomposition of CENUs, and then the dG–dC crosslink
is formed via the second alkylation of the complimentary cytosine
base.
For the time-courses of the crosslinking levels induced by ACNU
and FTMS (Fig. 3A and D), no obvious induction period was observed
in the initial stage of the treatments. This suggests that the for-
mations of dG–dC crosslinks induced by ACNU and FTMS are
faster than those by BCNU and CCNU. CENUs can undergo spon-
taneous decomposition to yield chloroethyldizonium ions, which
alkylate guanine followed by the formation of dG–dC crosslinks.
Our theoretical study revealed that the decomposition of CENUs
was the rate-limiting step in the whole process of the formation of
crosslinks [31,32]. Previous studies reported that the half-lives of
ACNU, BCNU, CCNU and FTMS were 34, 49, 53 and 43 min, respec-
tively [34,35]. From these theoretical and experimental evidences,
it can be inferred that ACNU and FTMS with shorter half-lives
decompose to chloroethyl diazonium ions more quickly and con-
sequently induce the formation of dG–dC crosslinks more quickly,
which finally results in the disappearance of the induction period.
On the contrary, BCNU and CCNU, which have longer half-lives and
Fig. 3. Plots of dG–dC crosslinking levels (crosslinks/107 base pairs) vs time (h)
in DNA treated by (A) ACNU, (B) BCNU, (C) CCNU and (D) FTMS with the drug
concentrations at 1, 2, 4 and 8 mM (n = 3).
decompose slower, slow down the formation of dG–dC crosslinks
by yielding less chloroethyl diazonium ions in the initial stage of the
treatment. Therefore, there is an induction period in the formation
of crosslinks induced by BCNU and CCNU.
299
300
301
302
Please cite this article in press as: L. Zhao, et al., Int. J. Mass Spectrom. (2014), http://dx.doi.org/10.1016/j.ijms.2014.04.018

G Model
MASPEC151671–7
ARTICLE IN PRESS
L. Zhao et al. / International Journal of Mass Spectrometry xxx (2014) xxx–xxx
6
shortest half-life exhibit highest crosslinking activity; while CCNU
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
with the longest half-life lead to the lowest crosslinking level.
From the above results, we presumed that the crosslinking activity
of CENUs might be related to their stability in aqueous solution.
Because the cytotoxicity of CENUs is related to the formation of
DNA ICLs, it was presumed that the anticancer activity of the
drugs was correlated to the level of dG–dC crosslinks. Wolff et al.
[36] compared the anticancer efficiency of ACNU, BCNU, CCNU,
FTMS and other CENUs by performing the survival gain analysis
of 364 studies describing 24,193 patients with high-grade glioma
treated in 504 cohorts. They demonstrated that the highest sur-
vival gain was achieved by ACNU (8.9 months) followed by CCNU
(5.3 months) and FTMS (2.0 months), while BCNU provided no sur-
vival gain. Except for CCNU, the dG–dC crosslinking levels obtained
in this work are correlated to the survival gain reported in the
epidemiological study, which is ACNU > FTMS > BCNU. CCNU has
relatively high survival gain, but lowest crosslinking level, which
may be related to its relatively high lipophilicity in the treatment
of glioma. Bodell et al. [37,38] treated human glial-derived cells
and 9 L rat gliosarcoma cells by ACNU and BCNU, and determined
the DNA ICLs using alkaline elution assay. They observed that
the crosslinking levels induced by ACNU were higher than those
induced by BCNU in all cell lines, which is also consistent with our
results.
4. Conclusion
331
In summary, the levels of dG–dC crosslinks induced by ACNU,
BCNU, CCNU and FTMS were quantitatively determined using
HPLC–ESI-MS/MS. The obtained time-courses of the crosslink-
ing levels for the four CENUs showed a time-dependent trend
at all concentrations of the drugs. At the initial stage of the
treatment (approximately the first 2–4 h), an obvious induction
period with very low crosslinking levels was observed in the
time-courses of BCNU and CCNU. The existence of the induc-
tion period corresponded to the dG–dC crosslinks being rarely
detectable at the beginning of the treatment. This provides con-
vincing evidence for the presumed mechanism that the first step
for the formation of dG–dC crosslinks was the monoalkylation
of guanine followed by the second alkylation of the complimen-
tary cytosine. The comparison of the crosslinking levels for the
four drugs indicated that ACNU induced remarkably higher lev-
els of dG–dC crosslinks than the other three drugs at all time
points. The crosslinking activities of CENUs were found to be
related to their stability in aqueous solution, i.e., ACNU with the
lowest stability has the highest crosslinking levels; on the con-
trary, CCNU with the highest stability has the lowest crosslinking
levels. Moreover, the low stability may also contribute to the
disappear of the induction period in the formation of crosslinks
induced by ACNU and FTMS. A correlation was found between
the determined crosslinking level and the previously reported
anticancer efficiency. ACNU, which has the highest crosslinking
activity, showed better survival gain for high-grade glioma than
BCNU, CCNU and FTMS in an epidemiological study. This sug-
gests that dC–dG crosslink can possibly be employed as a potential
biomarker for evaluating the anticancer efficiency of novel CENU
drugs. This work contributes to a further understanding of the anti-
cancer mechanism of CENUs, and will assist in the development
of novel bifunctional anticancer agents with high specificity and
efficiency.
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
Fig. 4. Levels of dG–dC crosslinks in calf thymus DNA treated by the four CENUs
at the concentration of (A) 1 mM, (B) 2 mM, (C) 4 mM and (D) 8 mM. Symbol des-
ignations are
(n = 3).
for ACNU, ꢀ for BCNU,
for CCNU and ꢁ for FTMS treatment
303
304
305
306
Fig. 4 shows the comparison of the crosslinking levels induced
by the four CENUs at various concentrations. It is predomi-
nant that ACNU has the highest crosslinking activity followed by
FTMS, BCNU and CCNU. The results indicates that ACNU with the
Acknowledgements
365
This work was supported by grants from the National Natu- Q3 366
ral Science Foundation of China (No. 21277001), the Beijing Nova
367
Please cite this article in press as: L. Zhao, et al., Int. J. Mass Spectrom. (2014), http://dx.doi.org/10.1016/j.ijms.2014.04.018

G Model
MASPEC151671–7
ARTICLE IN PRESS
L. Zhao et al. / International Journal of Mass Spectrometry xxx (2014) xxx–xxx
7
368
369
Program (No. 2009B08), and the Beijing Municipal Education Com-
mission Science and Technology Project (No. KZ201110005003).
mismatch repair-defective cells in vitro and in xenografts, Int. J. Cancer 85
(2000) 590–596.
[18] R.A. Finch, K. Shyam, P.G. Penketh, A.C. Sartorelli, 1,2-Bis(methylsulfonyl)-1-(2-
chloroethyl)-2-(methylamino)carbonylhydrazine (101 M): a novel sulfonylhy-
drazine prodrug with broad-spectrum antineoplastic activity, Cancer Res. 61
(2001) 3033–3038.
[19] L.-J. Zhao, R.-G. Zhong, Y. Zhen, An ONIOM study on the crosslinked base pairs
in DNA reacted with chloroethylnitrosoureas, J. Theor. Comput. Chem. 6 (2007)
631–639.
[20] W.J. Bodell, Repair of DNA alkylation products formed in 9L cell lines treated
with 1-(2-chloroethyl)-1-nitrosourea, Mutat. Res. 522 (2003) 85–92.
[21] F.-X. Chen, W.J. Bodell, G.-N. Liang, B. Gold, Reaction of N-(2-chloroethyl)-N-
nitrosoureas with DNA: effect of buffers on DNA adduction, cross-linking, and
cytotoxicity, Chem. Res. Toxicol. 9 (1996) 209–214.
[22] W.J. Bodell, DNA alkylation products formed by 1-(2-chloroethyl)-1-
nitrosourea as molecular dosimeters of therapeutic response, J. Neurooncol.
91 (2009) 257–264.
[23] M.T. Hayes, J. Bartley, P.G. Parsons, G.K. Eaglesham, A.S. Prakash, Mechanism
of action of fotemustine, a new chloroethylnitrosourea anticancer agent: evi-
dence for the formation of two DNA-reactive intermediates contributing to
cytotoxicity, Biochemistry 36 (1997) 10646–10654.
[24] P.G. Penketha, K. Shyama, A.C. Sartorelli, Comparison of DNA lesions produced
by tumor-inhibitory 1,2-bis(sulfonyl)hydrazines and chloroethylnitrosoureas,
Biochem. Pharmacol. 59 (2000) 283–291.
[25] H. Ueda-Kawamitsu, T.A. Lawson, P.R. Gwilt, In vitro pharmacokinetics and
pharmacodynamics of 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU), Biochem.
Pharmacol. 63 (2002) 1209–1218.
[26] W.P. Tong, M.C. Kirk, D.B. Ludlum, Formation of the cross-link 1-[N3-
deoxycytidyl], 2-[N1-deoxyguanosinyl]-ethane in DNA treated with N,Nꢀ-bis
(2-chloroethyl)-N-nitrosourea, Cancer Res. 42 (1982) 3102–3105.
[27] P.L. Fischhaber, A.S. Gall, J.A. Duncan, P.B. Hopkins, Direct demonstration in
synthetic oligonucleotides that N,Nꢀ-bis(2-chloroethyl)-nitrosourea crosslinks
N1 of deoxyguanosine to N3 of deoxycytidine on opposite strands of duplex
DNA, Cancer Res. 59 (1999) 4364–4368.
[28] L.-J. Zhao, T. Ren, B.-Q. Bai, R. Zhang, R.-G. Zhong, Comparative investigations of
deoxyribonucleic acid interstrand crosslinks induced by semustine using flu-
orescence and high performance liquid chromatography-mass spectrometry,
Chinese J. Anal. Chem. 39 (2011) 476–480.
[29] B.-Q. Bai, L.-J. Zhao, R.-G. Zhong, Quantification of meCCNU-induced dG–dC
crosslinks in oligonucleotide duplexes by liquid chromatography/electrospray
ionization tandem mass spectrometry, Rapid Commun. Mass Spectrom. 25
(2011) 2027–2034.
[30] B.-Q. Bai, L.-J. Zhao, R.-G. Zhong, Analysis of deoxyribonucleic acid interstrand
crosslinks induced by nitrosourea with high performance liquid chromatogra-
phy–electrospray ionization tandem mass spectrometry, Chinese J. Anal. Chem.
38 (2010) 532–536.
[31] L.-J. Zhao, X.-Y. Ma, R.-G. Zhong, A density functional theory investigation on the
formation mechanisms of DNA interstrand crosslinks induced by chloroethyl-
nitrosoureas, Int. J. Quantum Chem. 113 (2012) 1299–1306.
[32] L.-J. Zhao, X.-Y. Ma, R.-G. Zhong, Comparative theoretical investigation of the
formation of DNA interstrand crosslinks induced by two kinds of N-nitroso
compounds: nitrosoureas and nitrosamines, J. Phys. Org. Chem. 25 (2012)
1153–1167.
[33] W.J. Bodell, K. Pongracz, Chemical synthesis and detection of the cross-link
1-[N3-(2ꢀ-deoxycytidyl)]-2-[N1-(2ꢀ-deoxyguanosinyl)]ethane in DNA reacted
with l-(2-Chloroethyl)-l-nitrosoure, Chem. Res. Toxicol. 6 (1993) 434–438.
[34] N. Morikawa, T. Mori, T. Abe, H. Kawashima, M. Takeyama, S. Hori, Pharma-
cokinetics of cytosine arabinoside, methotrexate, nimustine and valproic acid
in cerebrospinal fluid during cerebrospinal fluid perfusion chemotherapy, Biol.
Pharm. Bull. 23 (2000) 784–787.
[35] B.H. Gordon, R.P. Richards, M.P. Hiley, A.J. Gray, R.M.J. Ings, D.B. Campbell, A
new method for the measurement of nitrosoureas in plasma: an HPLC pro-
cedure for the measurement of fotemustine kinetics, Xenobiotica 19 (1989)
329–339.
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
370
Appendix A. Supplementary data
371
372
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ijms.2014.04.018.
373
References
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
[1] T.C.A. Johannessen, R. Bjerkvig, B.B. Tysnes, DNA repair and cancer stem-like
cells: potential partners in glioma drug resistance, Cancer Treat. Rev. 34 (2008)
558–567.
[2] P. Merle, D. Morvan, D. Caillaud, A. Demidem, Chemotherapy-induced
bystander effect in response to several chloroethylnitrosoureas: an origin inde-
pendent of DNA damage, Anticancer Res. 28 (2008) 21–27.
[3] M.D. Bacolod, S.P. Johnson, A.E. Pegg, M.E. Dolan, R.C. Moschel, N.S. Bullock,
Q.M. Fang, O.M. Colvin, P. Modrich, D.D. Bigner, H.S. Friedman, Brain tumor
cell lines resistant to O6-benzylguanine/1,3-bis(2-chloroethyl)-1-nitrosourea
chemotherapy have O6-alkylguanine-DNA alkyltransferase mutations, Mol.
Cancer Ther. 3 (2004) 1127–1135.
[4] T.-T. Liu, L.-J. Zhao, R.-G. Zhong, DFT investigations of phosphotriesters hydrol-
ysis in aqueous solution: a model for DNA single strand scission induced by
N-nitrosoureas, J. Mol. Model. 19 (2013) 647–659.
[5] T. Walbert, M.R. Gilbert, M.D. Groves, V.K. Puduvalli, W.K.A. Yung, C.A. Conrad,
G.C. Bobustuc, H. Colman, S.H. Hsu, B.N. Bekele, W. Qiao, V.A. Levin, Combination
of 6-thioguanine, capecitabine, and celecoxib with temozolomide or lomustine
for recurrent high-grade glioma, J. Neurooncol. 102 (2011) 273–280.
[6] S. Watanabe, S. Sato, S. Nagase, S. Ohkuma, Chemotherapeutic choice of ran-
imustine or nimustine on the basis of regional polyamine levels in rat brain,
Methods Find. Exp. Clin. Pharmacol. 30 (2008) 115–120.
[7] A.D. Bolzan, M.S. Bianchi, Genotoxicity of streptozotocin, Mutat. Res. 512 (2002)
121–134.
[8] C. Jacquillat, D. Khayat, P. Banzet, M. Weil, P. Fumoleau, M.F. Avril, M. Namer,
J. Bonneterre, P. Kerbrat, J.J. Bonerandi, R. Bugat, P. Montcuquet, D. Cupissol, R.
Lauvin, C. Vilmer, C. Prache, J.P. Bizzari, Final report of the French multicenter
phase II study of the nitrosourea fotemustine in 153 evaluable patients with dis-
seminated malignant melanoma including patients with cerebral metastases,
Cancer 66 (1990) 1873–1878.
[9] D. Khayat, B. Giroux, J. Berille, V. Cour, B. Gerard, M. Sarkany, P. Bertrand, J.P.
Bizzari, Fotemustine in the treatment of brain primary tumors and metastases,
Cancer Investig. 12 (1994) 414–420.
[10] A. Daponte, S. Signoriello, L. Maiorino, B. Massidda, E. Simeone, A.M. Grimaldi,
C. Caraco, G. Palmieri, A. Cossu, G. Botti, A. Petrillo, S. Lastoria, E. Cavalcanti, P.
Aprea, N. Mozzillo, C. Gallo, G. Comella, P.A. Ascierto, Phase III randomized study
of fotemustine and dacarbazine versus dacarbazine with or without interferon-
␣ in advanced malignant melanoma, J. Transl. Med. 11 (2003) 38.
[11] N. Ni, T. Sanghvi, S.H. Yalkowsky, Stabilization and preformulation of anticancer
drug-SarCNU, Int. J. Pharm. 1/2 (2002) 257–264.
[12] S. Sekido, K. Ninomiya, M. Iwasaki, Biologic activity of MCNU: a new antitumor
agent, Cancer Treat. Rep. 63 (1979) 961–970.
[13] Y. Takaue, T. Watanabe, Y. Hoshi, T. Abe, K. Matsunaga, S.I. Saito, A. Hirao,
Y. Kawano, T. Ninomiya, Y. Kuroda, T. Koyama, T. Suzue, T. Shimokawa, H.
Uchiyama, A. Watanabe, T. Matsushita, A. Kikuta, A. Yokobayashi, R. Murakami,
A. Manabe, R. Hosoya, M. Ohira, T. Fujimoto, Effectiveness of high-dose MCNU
therapy and hematopoietic stem cell autografts treatment of childhood acute
leukemia/lymphoma with high-risk features, Cancer 67 (1991) 1830–1837.
[14] L.F.Z. Batista, W.P. Roos, M. Christmann, C.F.M. Menck, B. Kaina, Differential
sensitivity of malignant glioma cells to methylating and chloroethylating anti-
cancer drugs: p53 determines the switch by regulating xpc, ddb2, and DNA
double-strand breaks, Cancer Res. 67 (2007) 11886–11895.
[15] T. Yamaguchi, H. Kanemitsu, S. Yamamoto, M. Komatsu, H. Uemura, K. Tamura,
T. Shirai, N,Nꢀ-bis(2-chloroethyl)-N-nitrosourea (BCNU)-induced apoptosis of
neural progenitor cells in the developing fetal rat brain, J. Toxicol. Pathol. 23
(2010) 25–30.
[16] P.G. Penketh, R.P. Baumann, K. Ishiguro, K. Shyam, H.A. Seow, A.C. Sartorelli,
Lethality to leukemia cell lines of DNA interstrand cross-links generated by
cloretazine derived alkylating species, Leuk. Res. 32 (2008) 1546–1553.
[17] S. Fiumicino, S. Martinelli, C. Colussi, G. Aquilina, C. Leonetti, M. Crescenzi,
M. Bignami, Sensitivity to DNA cross-linking chemotherapeutic agents in
[36] J.E.A. Wolff, S. Berrak, S.E.K. Webb, M. Zhang, Nitrosourea efficacy in high-grade
glioma: a survival gain analysis summarizing 504 cohorts with 24193 patients,
J. Neurooncol. 88 (2008) 57–63.
[37] T. Aida, W.J. Bodell, Cellular-resistance to chloroethylnitrosoureas, nitrogen
mustard, and cis-diamminedichloroplatinum(II) in human glial-derived cell-
lines, Cancer Res. 47 (1987) 1361–1366.
[38] W.J. Bodell, M. Cerosa, T. Aida, M.S. Berger, M.L. Rosenblu, Investigation of resis-
tance to DNA cross-linking agents in 9L cell lines with different sensitivities to
chloroethylnitrosoureas, Cancer Res. 45 (1985) 3460–3464.
Please cite this article in press as: L. Zhao, et al., Int. J. Mass Spectrom. (2014), http://dx.doi.org/10.1016/j.ijms.2014.04.018