Improved synthesis of N-benzylaminoferrocene-based prodrugs and evaluation of their toxicity and antileukemic activity

Article abstract of DOI:10.1021/jm5019548

We report on an improved method of synthesis of N-benzylaminoferrocene-based prodrugs and demonstrate its applicability by preparing nine new aminoferrocenes. Their effect on the viability of selected cancer cells having different p53 status was studied. The obtained data are in agreement with the hypothesis that the toxicity of aminoferrocenes is not dependent upon p53 status. Subsequently the toxicity of a selected prodrug (4) was investigated ex vivo using rat precision cut liver slices and in vivo on hybrid male mice BDF1. In both experiments no toxicity was observed: ex vivo, up to 10 μM; in vivo, up to 6 mg/kg. Finally, prodrug 4 was shown to extend the survival of BDF1 mice carrying L1210 leukemia from 13.7 ± 0.6 days to 17.5 ± 0.7 days when injected daily 6 times at a dose of 26 μg/kg starting from the second day after injection of L1210 cells.

Full text of DOI:10.1021/jm5019548

Subscriber access provided by SELCUK UNIV
Article
Improved Synthesis of N-Benzylaminoferrocene-Based Prodrugs
and Evaluation of Their Toxicity and Antileukemic Activity
Steffen Daum, Vasyl F. Chekhun, Igor N Todor, Natalia Yu Lukianova, Yulia V Shvets,
Leopold Sellner, Kerstin Putzker, Joe Lewis, Thorsten Zenz, Inge de Graaf, Geny
M.M. Groothuis, Angela Casini, Olexii Zozulia, Frank Hampel, and Andriy Mokhir
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm5019548 • Publication Date (Web): 29 Jan 2015
Downloaded from http://pubs.acs.org on February 3, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 41
1
2
3
4
5
6
7
8
9
Improved Synthesis of NꢀBenzylaminoferroceneꢀ
Based Prodrugs and Evaluation of Their Toxicity
and Antileukemic Activity
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Steffen Daum,† Vasiliy F. Chekhun,‡ Igor N. Todor,‡ Natalia Yu. Lukianova, ‡ Yulia V. Shvets,‡
Leopold Sellner, § Kerstin Putzker,│ Joe Lewis,│ Thorsten Zenz, § Inge A. M. de Graaf, # Geny
M.M. Groothuis, # Angela Casini, # Olexii Zozulia, † Frank Hampel, † Andriy Mokhir†*
† Department of Chemistry and Pharmacy, FriedrichꢀAlexanderꢀUniversity of Erlangenꢀ
Nürnberg, Organic Chemistry II, Henkestr. 42, 91054 Erlangen, Germany
‡ R. E. Kavetsky Institute of Experimental Pathology, Oncology, and Radiobiology of the
National Academy of Sciences of Ukraine, Vasilkivska 45, 03022 Kyiv, Ukraine
§ Department of Translational Oncology, National Center for Tumor Diseases (NCT) and
German Cancer Research Center (DKFZ) Heidelberg, Im Neuenheimer Feld 460, 69120
Heidelberg, Germany and Department of Medicine V, University Hospital Heidelberg, Im
Neuenheimer Feld 410, 69120 Heidelberg, Germany
# Department of Pharmacokinetics, Toxicology and Targeting, Groningen Research Institute of
Pharmacy, University of Groningen, A. Deusinglaan 1, 9713 AV Groningen, The Netherlands
ACS Paragon Plus Environment
1

Page 3 of 41
Journal of Medicinal Chemistry
1
2
│ Chemical Biology Core Facility, European Molecular Biology Laboratory Heidelberg,
Meyerhofstraße 1, 69117 Heidelberg, Germany
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
KEYWORDS Aminoferrocene, Prodrug, Leukemia, TP53, Reactive Oxygen Species.
ABSTRACT. We report on an improved method of synthesis of Nꢀbenzylaminoferroceneꢀbased
prodrugs and demonstrate its applicability by preparing nine new aminoferrocenes. Their effect
on the viability of selected cancer cells having different p53 status was studied. The obtained
data are in agreement with the hypothesis that the toxicity of aminoferrocenes is not dependent
upon p53 status. Subsequently the toxicity of a selected prodrug (4) was investigated ex vivo
using rat precision cut liver slices and in vivo on hybrid male mice BDF1. In both experiments no
toxicity was observed: ex vivo ꢀ up to 10 µM; in vivo ꢀ up to 6 mg/kg. Finally, prodrug 4 was
shown to extend the survival of BDF1 mice carrying L1210 leukemia from 13.7 + 0.6 days to
17.5 + 0.7 days when injected daily for 6 times at a dose of 26 µg/kg starting from the second
day after injection of L1210 cells.
ACS Paragon Plus Environment
2

Journal of Medicinal Chemistry
Page 4 of 41
1
2
3
4
INTRODUCTION
5
6
7
8
9
The intracellular milieu of cancer cells differs substantially from that of normal ones. This
difference can be utilized to design cancer specific, ROSꢀsensitive prodrugs.1 In particular, the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ
concentration of reactive oxygen species (ROS= 1O2, O2 , H2O2, HO•) is higher in the majority of
cancer cells.2 It has been suggested that this biochemical alteration is required to support the
cancerous phenotype. Therefore, designing prodrugs that need activation by ROS can lead to
antiꢀcancer drugs that are active against a broad range of cancers.3 Known prodrugs activated at
elevated ROS concentrations include the naturally occurring cyclic glycopeptide bleomycin,4
hydroxyferrocifen and its analogues,5 organochalcogeneꢀbased proꢀcatalysts,6 H2O2ꢀresponsive
proꢀalkylation agents,7 SNꢀ38 (a camptothecin derivative)8 and aminoferroceneꢀbased prodrugs.9
In particular, bleomycin coordinates intracellular iron ions (present in excess in cancer cells)
resulting in the formation of a potent catalyst for cleavage of genomic DNA.4
Hydroxyferrocifens and organochalcogenes are oxidized to scavengers of intracellular
nucleophiles (e.g. glutathione (GSH) or thioredoxin reductases) and ROSꢀgenerating catalysts,
respectively; thus, inducing cell death via oxidative stress.5,6 Furthermore, proꢀalkylating agents,
such as SNꢀ38 and aminoferrocenes are activated by the cleavage of a CꢀB bond in the presence
of cancerꢀspecific H2O2 concentrations.8,9 In the case of aminoferrocenes two toxic products are
obtained, one of them being a ROSꢀgenerating catalyst and the other a GSH or selenoenzymes
scavenger. These products act synergistically on each other leading to a strong elevation of ROS
concentration in cancer cells causing their death. The dual mode of action of aminoferrocenes is
unique.9 The only other drug reported up to date, acting by a related mechanism, is [(η6ꢀ
arene)Ru(azpy)I]+.10 In particular, this compound first oxidizes GSH to GSSG. Then, the reduced
complex is oxidized by O2 to give the initial complex and H2O2. However, in contrast to
ACS Paragon Plus Environment
3

Page 5 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
aminoferrocenes this compound is not a prodrug. Therefore, it can potentially affect also normal
cells. Moreover, the oxidized product (GSSG) can be recovered by intracellular reductases
thereby diminishing the effect of the drug. In contrast, products released from aminoferrocenes
irreversibly modify glutathione by alkylating it.9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In previous work we have demonstrated that selected aminoferrocenes are active against
primary chronic lymphoid leukemia (CLL) cells,9b however, in this preliminary study the genetic
status of CLL cells was not determined. Nevertheless, it is well documented that CLL is a
genetically heterogeneous disease and some CLLꢀtypes exhibit poor prognosis.11 For example,
TP53ꢀmutated CLL is poorly responsive to standard therapy. The TP53 gene encodes tumor
suppressor p53, a protein participating in cell cycle regulation. Therefore, search for new drugs
for the treatment of TP53ꢀmutated CLL and other hematological cancers is warranted. Since the
mode of action of aminoferrocenes relies on ROSꢀinduced chemical activation that is not directly
related to TP53,9 it is probable that these prodrugs will be active against both wild type and TP53
mutated/negative hematological cancers. Herein, we discuss a new series of data proving this
hypothesis using three Bꢀcell derived cancer cell lines with different TP53ꢀstatus and primary
CLL cells with defined TP53 status. Moreover, we report on studies of toxicity and antitumor
activity of a selected aminoferrocene (compound 4, Scheme 1) using rat precision cut liver slices
(ex vivo), and BDF1 (DBA/2,♀ x C57Bl/6,♂) mice carrying TP53ꢀmutated L1210 leukemia (in
vivo). For these biological experiments a large amount of prodrug 4 (up to 2 g) was required,
which was obtained using a substantially improved method of synthesis of Nꢀbenzylated
aminoferrocenes. We also demonstrated the general applicability of this method by preparing
nine new Nꢀsubstituted aminoferrocenes.
ACS Paragon Plus Environment
4

Journal of Medicinal Chemistry
Page 6 of 41
1
2
3
4
RESULTS AND DISCUSSION
5
6
Synthesis of N-alkylated aminoferrocene-based prodrugs
7
8
9
The previously reported procedure for preparation of compound 4 yielded under optimized
conditions only ~5 % of the product.9 It consisted of reductive amination of commercially
available aminoferrocene 1 in the presence of benzaldehyde and Na[B(CN)H3] with formation of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nꢀbenzylaminoferrocene 2,
followed by coupling of compound
2
to 4ꢀ
(hydroxymethyl)phenylboronic acid pinacol ester in toluene in the presence of triphosgene at 120
°C for 24 h (steps a, b in Scheme 1). Since the intermediate product 2 is unstable in air, these two
Scheme 1. Outline of tested pathways for optimization of synthesis of Nꢀalkylated
aminoferroceneꢀbased prodrugs
R'
R''
Possible modification
sites of produgs
N
H
2
b
R''
R'= H; R''=
F
R'
O
Fe
O
a
N
Fe
NH2
Fe
c
O
d
OCH3
6
O
B
O
4
5
4ꢀ10
1
O
O
N
O
B
N
B
Fe
Fe
H
O
3
O
N
e
N
O
8
9
OAc
O
O
O
O
N
B
O
Fe
Fe
7
f
O
O
N
B
O
O
Fe
12
O
R'=R''=
10
11
O
N
B
O
Cl
O
a: R’R’’C=O, Na[B(CN)H3]; b: 4ꢀhydroxymethylphenylboronic acid pinacol ester,
Cl3COC(O)OCCl3, 120 °C;9 c: same as b; d: benzyl chloride, Cs2CO3, tertꢀbutylamonium iodide
(TBAI); e: 1,2ꢀdiꢀ(chloromethyl)benzol, Cs2CO3, TBAI; f: compound 3, Cs2CO3, TBAI.
ACS Paragon Plus Environment
5

Page 7 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
steps were conducted in a one pot reaction. We observed that changes of conditions including (1)
optimization of the reductive amination combined with isolation of compound 2, (2) replacement
of triphosgene either for N,N’ꢀcarbonyldiimidazole (CDI) or CDI/CH3I mixture or 4ꢀ
nitrophenyloxycarbonylchloride and (3) increasing nucleophilicity of the hydroxyl group of 4ꢀ
hydroxymethylphenylboronic acid pinacol ester by its deprotonation on the presence of NaH led
to further reduction of the yield of the desired product. Moreover, application of substituted
benzaldehydes in this reaction sequence with an attempt to prepare analogues of compound 4
was hampered by even lower yields. The latter effect was especially pronounced for
pyridinecarboxaldehydes and 2ꢀ and 4ꢀsubstituted benzaldehydes carrying either electron
donating substituents like OMe, NMe2 or CH3 or thermally and hydrolytically unstable
substituents like OAc or NHAc. Therefore, an alternative approach for synthesis of compound 4
was developed. In particular, we used compound 3 as a starting material. In contrast to 2,
compound 3 is highly stable and can be obtained from commercially available starting materials
in large quantities with 74 % yield.9a We attempted alkylation of compound 3 (1) under
optimized Mitsunobu conditions for carbamates,12 using benzyl alkohol, Bu3P and 1,1’ꢀ
(azidodicarbonyl)dipiperidine (ADDP); (2) in the presence of Cs2CO3 using BnCl/(NBu4)I
mixture as an electrophile under conditions of Salvatore et al13; and (3) in the presence of NaH
using either benzyl chloride or bromide as electrophiles as described elsewhere for other starting
materials.14 While the first approach delivered only traces of the product, the other two reaction
procedures led to clean conversion of compound 3 into 4. Isolated yields of analytically pure
compound 4 varied in the range between 71 and 92 %, which is substantially improved with
respect to the previously reported synthetic protocol (5 %). We confirmed that this method is
suitable for scaling up the synthesis of compound 4 up to 2 g. We also used this protocol to
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
6

Journal of Medicinal Chemistry
Page 8 of 41
1
2
3
4
5
6
7
8
9
prepare nine new analogues of 4, including compounds, containing differently substituted
phenylmethyl and pyridylmethyl moieties. In particular, in the series of compounds 4, 5, 6
electron deficiency of the Ph residue is modulated by substituents in the meta position: H, F,
OMe. In compound 7 the paraꢀhydroxy group is protected as an acetic acid ester. In cells it is
expected to be deprotected via hydrolysis forming the highly electron rich 4ꢀ
hydroxyphenylmethylꢀcontaining aminoferrocene, which will be partially deprotonated at pH 7
with formation of the phenolate ion (pKa ~ 9ꢀ10). In contrast, 2ꢀ and 3ꢀpyridylꢀresidues in
compounds 8 and 9 are strongly electron deficient. At pH 7 they are expected to remain neutral
(pKa~ 5ꢀ6). Compound 10 is derived from 4 by substitution of a benzyl for a diphenylmethyl
residue. It is obtained by alkylation of intermediate 3 with diphenylmethylchloride with the
relatively low yield of 6 %. The latter result reflects the low reactivity of the corresponding
electrophile due to its steric bulkiness. Compound 10 is the first example of the aminoferroceneꢀ
based prodrug where the carbon atom connected to the nitrogen atom of the aminoferrocene core
carries two substituents (phenyl residues in this case). Synthesis of such derivatives using the
initially applied protocol was not possible since the reductive amination product formed from
aminoferrocene 1 and ketones in step a is not reactive in the following step b (Scheme 1). Since
all steps in the new protocol were conducted at 22 °C, thermally unstable or reactive substrates
could be applied. For example, compound 11, which carries 2ꢀchlormethyl residue could be
prepared from equimolar mixture of compound 3 and 1,2ꢀdi(chlormethyl)benzene. The product
was isolated with the yield of 34%. Xꢀray single crystal diffraction analysis allowed to determine
the structure of compound 11 (Figure 1), which is the first crystallographically characterized
aminoferroceneꢀbased prodrug.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
7

Page 9 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Molecular structure of compound 11 as determined by Xꢀray crystallography.
1
These data confirm the atom connectivity suggested based on the synthesis, H NMR
spectroscopy and mass spectrometry of 11. The geometry of a ferrocene unit in compound 11 is
similar to that in other known alkyloxycarbonylaminoferrocenes.15 Atoms of a carbamate group
(C2, N1, C3, O1, O2) are located in a plane, which is inclined by 24° to the plane of the
cyclopentadienyl ligand containing atoms C42, C41 and C45 (Figure 1). These data are
indicative of only weak conjugation between the ferrocene unit and the carbamate fragment. The
fully planar conformation is not realized due to the expected clash between the bulky substituents
at N1 atom with the hydrogen atoms attached to C42 and C45 atoms. It is evident that a fragment
–H2C1ꢀCl1 is not hindered by other atoms in structure 11 (Figure 1). Therefore, we expected that
it is accessible to nucleophilic reagents. In agreement with this expectation, we observed that 11
reacts with excess 3 with formation of doubly substituted product 12. We are currently exploring
11 for conjugation of the aminoferrocene core to other nucleophilic modifiers, e.g. those, which
can either target these prodrugs to specific compartments in cells (e.g. nuclei or mitochondria) or
direct them to cancer cells. Analogously to monoꢀaminoferroceneꢀbased compound 3, 1,1’ꢀbisꢀ
ACS Paragon Plus Environment
8

Journal of Medicinal Chemistry
Page 10 of 41
1
2
3
4
5
6
7
8
9
aminoferroceneꢀbased compound 13 can be efficiently alkylated at the optimized conditions
producing doubly Nꢀbenzylated product 14 with the yield of 55 % (Scheme 2). In this case the
yield is slightly lower than that observed in the alkylation of compound 3 – 71ꢀ92 %.
Thus, the synthetic method described above expands our possibilities in the modification of the
Nꢀsubstituent in aminoferrocenes at the sites indicated with red arrows in Scheme 1, whereas the
previously used protocol allowed only modifications at the sites indicated with blow arrows and
was restricted to thermally stable substrates. Moreover, the protocol developed in this work
allows scalingꢀup the synthesis of Nꢀsubstituted aminoferrocenes that enables their thorough
structural, spectroscopic and biological studies.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthesis of doublyꢀmodified aminoferrocenes using the developed improved
synthetic protocol and structures of controls.
a: BnCl, Cs2CO3, TBAI.
Effects of prepared prodrugs on the viability of B-cell-derived cancer cell lines having
different TP53-status
To explore the role of p53 in the mode of action of aminoferroceneꢀbased prodrugs we studied
their effects on the viability of three selected Bꢀcell derived cancer cell lines with different TP53
ACS Paragon Plus Environment
9

Page 11 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
status. The human mantle cell lymphoma (MCL) was represented by the JVMꢀ2 cell line. MCL
accounts for ~7 % of all lymphoid malignancies and has relatively poor prognosis due to almost
invariable relapse occurring after initial chemotherapy.16 The JVMꢀ2 cell line contains wild type
(wt) p53 and an additional t(11;14) translocation. The human Burkitt lymphoma was represented
in our panel by RAJI and BLꢀ2 cell lines. Notably, mutated p53 is expressed in RAJI, whereas wt
p53 in BLꢀ2 cells. Moreover, in RAJI a translocation t(8,14) as well as EpsteinꢀBarrꢀVirus
(EBV+) are detected, whereas in BLꢀ2 a translocation t(8,22) is present. The latter cell line is
EBV negative.17 The promyelocytic human leukemia cell line HLꢀ60, which expresses no
detectable p53 protein and p53 mRNA and whose TP53 gene underwent major deletions,18 was
also included for comparison (Table 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
First, we determined the relative amounts of intracellular reactive oxygen species in
JVMꢀ2, RAJI and BLꢀ2 by using 2,7ꢀdichlorodihydrofluorescein diacetate (DCFHꢀDA)ꢀbased
assay as described earlier.9 In particular, we observed that the relative mean fluorescence (F) of
JVMꢀ2 (F= 0.71 + 0.01) or BLꢀ2 cells (F= 0.77 + 0.03) measured after their incubation with
DCFHꢀDA for 30 min was comparable to that of HLꢀ60 cells (F=1.0). In contrast, in these
conditions the relative mean fluorescence of RAJI cells was ~3 fold higher than that of HLꢀ60
cells. These data indicate that ROS amounts in JVMꢀ2 and BLꢀ2 cells are similar to that in HLꢀ
60 cells, whereas in RAJI cells it is higher than that in HLꢀ60 cells. Since the prodrugs are
efficiently activated in HLꢀ60 cells due to their reaction with ROS,9 these data indicate that an
analogous activation should also be possible in JVMꢀ2, BLꢀ2 and RAJI cells provided that the
prodrugs are taken up into the cells.
The physiological concentration of boronꢀcontaining compounds in cells is negligible.
Since each molecule of the prodrug contains one boron atom, we could investigate the membrane
ACS Paragon Plus Environment
10

Journal of Medicinal Chemistry
Page 12 of 41
1
2
3
4
Table 1. Effects of the aminoferroceneꢀbased prodrugs on the viability of the selected cell lines
5
6
(IC50 values) with different p53ꢀstatus after 48 h incubation.i
7
8
9
Drugii
IC50 (µM)
BLꢀ2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HLꢀ60
no p53
>50
RAJI
mut p53
>50
JVMꢀ2
wt p53
~50
wt p53
>50
3
11 + 1
9 + 7
17 + 4
16 + 1
ND
35 + 6
36 + 5
36 + 2
37 + 3
>50
25 + 1
23 + 5
23 + 7
26 + 0
19 + 3
21 + 6
16 + 1
3.8 + 0.4
>50
29 + 6
30 + 4
30 + 4
37 + 5
41 + 7
45 + 5
24 + 6
8 + 0
>50
4
5
6
7
8
ND
>50
9
15 + 2
6 + 0
~50
21 + 4
7 + 1
>50
10
11
12
13
14
15
16
8 + 8
10 + 0
>50
ND
ND
ND
>50
29 + 19
>50
>50
>50
>50
>50
>50
>50
>50
3 + 1
3 + 1
3 + 1
3 + 1
17
i
no p53 – wt p53 is not expressed; mut p53 – mutated p53; wt p53 – active (wild type) p53 is
expressed. Data are given as mean of at least three independent experiments + SD. ND – not
determined.
ii Structure of prodrugs and their labeling scheme are given in Schemes 1 and 2.
permeability of these compounds by monitoring the boron content in the cells.9 Intracellular
boron was determined for all four studied cell lines by the curcumin assay in combination with
ACS Paragon Plus Environment
11

Page 13 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
UVꢀvisible spectrophotometry for the representative compound 4. In this assay the uptake
correlates with the baseline corrected absorbance at 550 nm (A(550 nm) − A(780 nm))
characteristic for the curcuminꢀboron complex. In particular, untreated cells give rise to A(550
nm) − A(780 nm) in the range of 0.34ꢀ0.38, whereas cells incubated with compound 4 give rise
to A(550 nm) − A(780 nm) between 1.30 and 2.36. When the uptake of compound 4 in HLꢀ60
cells is taken as a reference (1.0), the permeability in RAJI cell line can be determined to be
equal to 1.8 + 0.5, that in JVMꢀ2 cell line to be 1.5 + 0.1 and that in BLꢀ2 cell line 1.4 + 0.1.
These data indicate that, within the experimental error, compound 4 permeates equally well
through the membranes of all studied cell lines.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Furthermore, we studied the viability of three selected Bꢀcell derived cancer cell lines and
of HLꢀ60 cell line in the presence of known (3, 4, 13) and newly prepared aminoferroceneꢀbased
prodrugs by using the 3ꢀ(4,5ꢀdimethylthiazolꢀ2ꢀyl)ꢀ2,5ꢀdiphenyltetrazolium bromide (MTT)
assay (Table 1). Among the new aminoferrocenes reported in this paper, only compound 11,
which contains a 2ꢀCH2Cl residue in the Nꢀbenzyl substituent, turned out to be substantially
more toxic towards all cell lines tested than the parent compound 4 (p< 0.01). We can speculate
that in this case the stronger cytotoxic effect could be caused by alkylation of endogenous
nucleophilic moieties, e.g. thiol groups of GSH or proteins, by the 2ꢀCH2Cl residue. At this point
other possible mechanisms, e.g. enhanced uptake of this compound, cannot be excluded. Since
the experiments with the CLL cells (as described below) were performed with medium
containing human serum (HS), whereas the cell lines were incubated in medium with fetal calf
serum (FCS), we investigated whether the replacement of FCS by HS in experiments with HLꢀ60
cells affects the toxicity of compound 11 as well as that of reference compound 4. We observed
that while the effect of compound 4 was not dependent on the type of the serum used, the
ACS Paragon Plus Environment
12

Journal of Medicinal Chemistry
Page 14 of 41
1
2
3
4
5
6
7
8
9
toxicity of 11 was fully inhibited in the HSꢀcontaining medium even at the highest concentration
of the drug used (50 µM). Furthermore, in experiments with CLL cells compound 11 exhibited
no toxicity at concentrations of up to 10 µM (data not shown). These data indicate that HS is able
to scavenge in some way compound 11. Since the only structural difference between compounds
4 and 11 is the presence of a strongly electrophilic CH2Cl group in the latter case, we can suggest
that the scavenging occurs due to alkylation of nucleophilic centers of HS by compound 11.
Independent of the mechanism of the latter process, it may preclude the possible anticancer
activity of compound 11 in vivo. Therefore, 11 was not selected for further biological tests.
It is worth noting that neither the substitution of 3ꢀ and 4ꢀHꢀatoms in the phenyl residue
for either electron donating or accepting moieties (compounds 5, 6, 7) nor the replacement of the
phenyl residue with 2ꢀ and 3ꢀpyridyl moieties (compounds 8, 9) affected the activity of
aminoferrocenes substantially. Furthermore, substitution of the Nꢀbenzyl residue for Nꢀ
diphenylmethyl (compound 10) only slightly enhanced the activity towards the cancer cell lines
RAJI (IC50= 21 + 4 µM for 10 versus 35 + 6 µM for 4, p<0.05) and BLꢀ2 (IC50= 16 + 1 µM for
10 versus 25 + 1 µM for 4, p<0.01), whereas its effect on two other cell lines (HLꢀ60 and JVMꢀ
2) was found to be negligible (Table 1). Unfortunately, the Nꢀbenzylation of the carbamate group
of a 1,1’ꢀbisꢀaminoferroceneꢀderived prodrug did not exhibit the analogous favorable effect as
the same derivatization of the monoꢀaminoferrocenes (compare 13, 14 and 3, 4, Table 1).
We observed that the resistance of wt p53ꢀexpressing BLꢀ2 and JVMꢀ2 cell lines as well
as mut p53ꢀexpressing RAJI cell line towards aminoferrocenes is similar for the majority of the
prodrugs studied here, except that in a few cases (compounds 7, 8 and 9) the BLꢀ2 cell line was
found to be slightly more sensitive than RAJI and JVMꢀ2 cell lines (Table 1). Overall, these data
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
13

Page 15 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
support the hypothesis that the anticancer activity of aminoferroceneꢀbased prodrugs does not
require p53 activity.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Effect of aminoferrocenes on the viability of wild type and TP53-mutated/deleted chronic
lymphocytic leukemia cells
Subsequently, we investigated the effects of two representative aminoferroceneꢀbased prodrugs
3 and 4 on the viability of CLL cells isolated from 25 patients. In this group 8 patients had p53ꢀ
mutated CLL cells and 17 patients had wild type p53 CLL cells. The effect of compounds 3 and
4, as well as of fludarabine on the viability of CLL cells was determined by using ATPlite™
Luminescence Assay System (PerkinElmer, Waltham, USA). Fludarabine is a purine analogue,
which is currently used as a standard treatment of the CLL disease. The mode of action of this
drug depends upon p53,19 making it more active against wt p53 cancer types. Therefore,
treatment of CLL patients with fludarabine can cause a selective survival of p53ꢀmutant CLL
cells, which are more resistant and, therefore, less responsive to the therapy. We observed that
after treatment of p53ꢀmutated CLL cells (8 patients) with fludarabine (10 µM) for 48 h 48.7 +
16.1 % viable cells remained in the suspension, whereas the same treatment of wild type p53
CLL reduced the number of viable cells to 34.7 + 18.1 % (17 patients, Figure 2).
ACS Paragon Plus Environment
14

Journal of Medicinal Chemistry
Page 16 of 41
1
2
3
4
5
6
7
8
9
fludarabine
60.0
30.0
0.0
prodrug 4
prodrug 3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
wt TP53
mut TP53
Status of TP53 gene
Figure 2. Mutated p53 CLL cells (mut TP53, 8 patients) or wild type CLL cells (wt TP53, 17
patients) were treated with prodrugs 3, 4 or fludarabine (each 10 µM) for 48 h and the % of
viable cells in suspension was determined as described in the experimental section and plotted as
the relative number of viable cells (the total number of control untreated cells was taken as 100
%).
The response of the cells of the selected, small group of patients to fludarabine was strongly
heterogeneous, which is reflected in large standard deviations of the number of viable cells in
populations of treated CLL cells – 16.1 % for p53ꢀmutated CLL and 18.1 % for wild type p53
CLL. Therefore, the difference in activities of this drug towards mut p53 CLL and wt CLL cells
observed is not significant within the observed experimental error (Student’s tꢀtest).
Aminoferrocenes 3 and 4 exhibited in general a higher toxicity towards CLL cells than
fludarabine (p<0.001): 9.1ꢀ12.5 % viable cells left after their treatment with these prodrugs (10
µM) for 48 h (compare with 39.2 % obtained after the treatment of CLL cells with fludarabine).
Their toxicity turned out to be practically independent of the status of p53 in CLL cells (Figure
2). These data are in agreement with our results obtained for Bꢀcellꢀderived cell lines (see above)
ACS Paragon Plus Environment
15

Page 17 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and support the hypothesis that the mode of action of aminoferrocenes is not dependent upon
p53. This advantageous property can make these prodrugs potentially suitable for the treatment
of p53ꢀmutated CLL and other hematological cancers. Interestingly, some ferrocifens have been
found to induce senescence in cancer cells, which does not correlate to p53 status.5c
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ex vivo and in vivo toxicity of aminoferrocene 4
Based on the preliminary studies obtained so far, we selected compound 4 for further
evaluation of its toxicity ex vivo and in vivo. We have previously shown that this prodrug is not
toxic towards normal (nonꢀcancer) cells including fibroblasts (< 100 µM) and primary
mononuclear cells (MNC’s, < 10 µM).9 In order to validate these in vitro data in a more complex
experimental model, we first investigated the toxicity of compound 4 ex vivo using rat precisionꢀ
cut liver slices (PCLS) according to established procedures.20 PCLS are viable explants of tissue
with a reproducible and wellꢀdefined thickness, containing cells in their natural environment.
Notably, this technique is an FDAꢀapproved model for drug toxicity and metabolism studies, and
it has been applied already for the assessment of toxicity of cisplatin21 and experimental goldꢀ
based compounds.22 In this assay liver cells remain in their natural environment and therefore,
the obtained toxicity results are expected to be more relevant than those obtained earlier in
experiments with cell cultures.
Thus, rat liver slices were cut and preꢀincubated either for 1 h or 24 h before exposure to the
prodrug as detailed in the experimental section, to allow the cells to recover from the cold
ischemia and the slicing procedure. After preꢀincubation, slices were incubated in fresh WEGG
medium containing 2.5 µM, 5 µM or 10 µM of compound 4 for a period of 24 h. Afterwards they
were collected and the viability of the tissues was determined measuring the ATP content. The
ACS Paragon Plus Environment
16

Journal of Medicinal Chemistry
Page 18 of 41
1
2
3
4
5
6
7
8
9
choice of the preꢀincubation time appeared to be crucial: a marked decrease in the viability of 1
h preꢀincubated slices was already found at a concentration of 5 µM prodrug (Figure 3). In
contrast, slices preꢀincubated for 24 h remained practically fully viable (83 %) even at the
highest tested concentration (10 µM, Figure 3).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
120
60
0
1h pre
24h pre
0,1 µM 0,5 µM 1 µM
control 2,5 µM 5 µM 10 µM
[17]
[4]
Figure 3. Viability of precision cut liver slices (PCLS) after 24 h of incubation with the
aminoferroceneꢀbased prodrug 4 and positive control 17 following 1 h (1 h pre) or 24 h of preꢀ
incubation (24 h pre).
Since aminoferrocenes are activated and as such expected to be more toxic under oxidative
conditions9, we may speculate that after 1 hour preꢀincubation the slices are subject to oxidative
stress due to the organ excision and slicing procedure increasing their sensitivity to compound 4.
Indeed, others have reported that shortly after slice preparation the GSH concentration in the
slices is low.23 In contrast, after 24 h preꢀincubation slices reach a more physiological
equilibrium, which is reflected in normal intracellular GSH concentration.23 Preliminary
experiments exploring the toxicity of cisplatin in liver slices have shown that no effect of preꢀ
incubation exists for this compound (data not shown). This indicates that the effect of preꢀ
ACS Paragon Plus Environment
17

Page 19 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
incubation is probably related to the specific toxic mechanism of aminoferrocenes. Taken
together these data confirm that in normal (nonꢀmalignant) cells, which are characterized by low
ROS concentration, aminoferrocenes remain inactive.9 Furthermore, experiments were carried
out with ferrocene 15 as a negative control and 17 as a positive control (Scheme 2). Both
compounds were added to slices preꢀincubated for 24 h, and incubation was continued for
additional 24 h. As expected, ferrocene 15 showed no reduction in viability even at 40 µM,
whereas compound 17 showed a high toxicity with an IC50 of ca. 1 µM.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Based on the promising in vitro9 and ex vivo data on the toxicity of prodrug 4, we decided to
further study its effects on healthy hybrid male mice BDF1 (DBA/2,♀ x C57Bl/6, ♂) at variable
doses between 10 and 6000 µg/kg as described in the experimental section. Each dose was
introduced via an intraperitoneal injection. One group contained 6 animals, which were
monitored for 21 days after the drug injection. Every third day animals were weighted. All
animals survived this treatment. Their weight increased by 1.8ꢀ2.5 g during the time of the
experiment. The same weight increase was observed in the control group, which received no
treatment. These data suggest that prodrug 4 exhibits no fulminant toxicity in vivo at
concentrations of up to 6 mg/kg, although more subtle effects cannot be excluded.
Antitumor activity of prodrug 4 in vivo
Finally, antitumor activity of 4 was evaluated on hybrid male mice BDF1 (DBA/2,♀ x
C57Bl/6, ♂), which carry L1210 leukemia, expressing mutant p53.17 In an exploratory
experiment the treatment of mice (15 animals per group) was started on the second day after
injection of L1210 cells. It included six daily intraperitoneal injections of compound 4 at a dose
of 26 µg/kg in PBS solution (0.5 mL) containing 0.2 % (v/v) DMSO (total dose: 0.156 mg/kg).
ACS Paragon Plus Environment
18

Journal of Medicinal Chemistry
Page 20 of 41
1
2
3
4
5
6
7
8
9
The control groups received the same volume of PBS solution. We observed that survival of the
mice was extended from 13.7 + 0.6 days in the control group to 17.5 + 0.7 days in the treated
group (p<0.05, Figure 4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Control
4 (26 mkg/kg)
100%
80%
60%
40%
20%
0%
0
5
10
15
20
25
Time, days
Figure 4. A KaplanꢀMaier survival plot illustrating extension of the life span of BDF1 mice
(DBA/2,♀ x C57Bl/6, ♂), which carry L1210 leukemia, in the result of their treatment with six
daily intraperitoneal injections of compound 4 at a dose of 26 µg/kg.
This corresponds to 28 % lifetime extension that according to Sofina et al24 indicates a
significant antitumor effect thereby justifying further investigations of prodrug 4. It has been
previously reported that 7 days after intraperitoneal injection of L1210 in BDF1 mice (DBA/2,♀
x C57Bl/6, ♂) ascites fluid of the animals contains over 93 % of L1210 cells with respect to the
total cell number.25 In order to evaluate the drugꢀinduced oxidative stress, we took probes of
ascites fluid of treated and control groups and analyzed mitochondria integrity (JCꢀ1 assay) and
the intracellular ROS amount (CMꢀH2DCF DA assay) in the L1210 cells that were present in
ACS Paragon Plus Environment
19

Page 21 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
this sample. In particular, the fluorescence of L1210 cells isolated from the drugꢀtreated animals
and incubated with the JCꢀ1 probe was found to be 97.0 + 4.3 arbitrary units (a.u.), whereas the
control probe (cells isolated from untreated animals) exhibited significantly (p<0.05) lower
fluorescence of 48.2 + 3.1 a.u. These data indicate the disturbance of the function of
mitochondria in the cancer cells in the treated animals. Furthermore, we observed that L1210
cells from the treated animals, which were incubated with CMꢀH2DCF DA dye (230.0 + 18.5
a.u.), were more fluorescent than those from the control group (163.3 + 20 a.u., p< 0.05). These
data indicate that the treatment with compound 4 enhances the oxidative stress in cancer cells in
vivo.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSIONS
We developed a substantially improved method for synthesis of Nꢀsubstituted aminoferroceneꢀ
based prodrugs, which leads to yields of 72ꢀ91 %. In contrast, by using the previous method
these compounds could be obtained with yields of under 5 %.9 In the key step, easily accessible
starting material 4ꢀ(ferrocenylaminocarbonyloxymethyl)ꢀphenylboronic acid pinacol ester was
alkylated by variable benzyl chlorides or bromides at 22 °C, whereas in the previously used
protocol prolonged heating at 120 °C was required.9 Therefore, heat sensitive building blocks
could be introduced by the new approach, including for example an Nꢀ(4ꢀ
acetyloxyphenyl)methyl substituent. Overall, nine new Nꢀsubstituted prodrugs were prepared by
this protocol, which were obtained in high purity and amount and required limited purification.
Thus, we were able to characterize the biological effects of the new series of compounds in vitro,
ex vivo and in vivo. First, we observed that the prodrugs exhibit similar toxicity towards Bꢀcell
derived hematological cancer cell lines (JVMꢀ2, RAJI and BLꢀ2) having different p53 status.
ACS Paragon Plus Environment
20

Journal of Medicinal Chemistry
Page 22 of 41
1
2
3
4
5
6
7
8
9
Subsequently, we confirmed that the toxicity of aminoferrocenes is practically independent from
the p53 status in primary CLL cells. In preparation to the in vivo studies of the antitumor activity
of the prodrugs, we confirmed that compound 4 exhibits practically no toxicity up to 10 µM in
precision cut liver slices preꢀincubated for 24 h up (ex vivo), and up to 6 mg/kg in hybrid male
mice BDF1 (DBA/2,♀ x C57Bl/6, ♂) (in vivo). Finally, the antitumor activity of the most active
prodrug 4 was studied in hybrid male mice BDF1 (DBA/2,♀ x C57Bl/6, ♂), which carry L1210
leukemia. We observed that compound 4 administered in 6 daily doses of 26 µg/kg extends the
survival of the mice from 13.7 + 0.6 days to 17.5 + 0.7 days. The prolonged survival was
accompanied by increased oxidative stress of and membrane damage in L1210 cells isolated
from the ascites fluid of treated animals. It is worth mentioning that the dose used in the in vivo
antitumor experiment is much lower than the maximum tested dose of the in vivo toxicity
studies, where no weight loss of the animals has been observed. Although further studies are
necessary to validate the mechanisms of action of this new class of compounds, as well as more
advanced toxicity studies, our findings indicate that these aminoferrocene prodrugs can
potentially be applied for the treatment of cancers, which express mutated or no p53 and are
often poorly responsive to the conventional chemotherapy.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
General Information
Commercially available chemicals of the best quality from Aldrich/Sigma/Fluka (Germany) were
obtained and used without purification. Starting materials 1ꢀ3 (Scheme 1), 13 (Scheme 2) and
control compounds 16 and 17 (Scheme 2) were prepared as described previously.9 NMR spectra
were acquired on a Bruker Avance 400 or Bruker Avance III 600 spectrometer. ESI mass spectra
ACS Paragon Plus Environment
21

Page 23 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
were recorded on a Bruker ESI MicroTOF II. C, H, and N analysis was performed in the
microanalytical laboratory of the chemical institute of the FiedrichꢀAlexanderꢀUniversity of
7
8
9
ErlangenꢀNürnberg. UVꢀvis spectra were measured on
a
Lambda Bio+ UV/Vis
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
spectrophotometer (Perkin Elmer) by using quartz glass cuvettes (Hellma GmbH, Germany) with
a sample volume of 1 mL or microꢀcuvettes with a sample volume of 100 µL (BRAND GmbH,
Germany). The fluorescence of live cells was quantified using a Guava easyCyteTM 6ꢀ2L Flow
cytometer from Merck Millipore. The data were processed using the inCyteTM software package
from Merck Millipore and the ModFIT LTTM software from Verity Software House. The purity
of the prodrugs used in the biological tests was determined by C, H, and N analysis. According
to these data, the purity of the prodrugs and controls was greater than 95%. Crystal structure data
for prodrug 11 (Figure 1) were submitted to the Cambridge Structural Database (CSD). This
structure was assigned the deposition number CCDC 1011896.
Synthesis
Compound 4, protocol 1. Compound 3 (0.5 g, 1.08 mmol) was dissolved in anhydrous DMF (20
mL) under nitrogen atmosphere. Then Cs2CO3 (1.06 g, 3.25 mmol) and tetraꢀbutylammonium
iodide (TBAI, 1.20 g, 3.25 mmol) were added to the solution. After stirring for 30 minutes at 22
°C, benzyl chloride (412 mg, 374 µL, 3.25 mmol) was added. The mixture was stirred for 18 h at
22 °C. Then, the solvent was removed in vacuum (0.01 mbar) and the product was purified by
column chromatography on silica gel using petroleum ether/acetone (10/1, v/v) as eluent yielding
an orange oil. Triturating of this material with hexane gave the product as an orange solid. Yield:
1
427 mg, 0.77 mmol (71%). TLC (SiO2, eluent petroleum ether/acetone, 4:1, v/v) Rf = 0.5; Hꢀ
ACS Paragon Plus Environment
22

Journal of Medicinal Chemistry
Page 24 of 41
1
2
NMR (acetonꢀd6,300 MHz)δ (ppm) 7.73 (d, 3J = 7.1 Hz, 2 H), 7.39ꢀ7.27 (m, 7 H), 5.25 (s, 2 H),
5.02 (s, 2 H), 4.46 (s,2 H), 4.12 (s, 5 H), 3.97 (s, 2 H), 1.34 (s, 12 H). The spectra are in
agreement with those reported for this compound.9
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 4, protocol 2. Compound 3 (50.0 mg, 108 µmol) was dissolved in anhydrous DMF (2
mL) under a nitrogen atmosphere. The solution was treated with NaH (5.20 mg, 130 µmol, 60 %
mineral oil suspended) and stirred 30 min at 22 °C. Benzyl bromide (27.8 mg, 19.3 µL 163
µmol) was added and the reaction was stirred for 2 h at 22 °C. Then, the solvent was removed in
vacuum (0.01 mbar) and the product was purified by column chromatography on silica gel using
petroleum ether/acetone (4/1, v/v) as eluent. Yield: 55 mg, 100 µmol (92%).
Compound 5 was prepared analogously to compound 4 using protocol 1. The product was
obtained as orange solid. Yield: 253 mg, 444 µmol (68%). TLC (SiO2, eluent petroleum
1
3
ether/acetone, 4/1, v/v) Rf = 0.5; HꢀNMR (300 MHz, acetonꢀd6): δ (ppm) 7.75 (d, J = 6.6 Hz,
2H), 7.41 (m, 3H), 7.19ꢀ7.03 (m, 3H), 5.27 (s, 2H), 5.05 (s, 2H), 4.47 (s, 2H), 4.15 (s, 5H), 4.00
(s, 2H), 1.35 (s, 12H); 13CꢀNMR (101 MHz, acetonꢀd6): δ (ppm) 164.65, 162.23, 142.57, 140.30,
135.11, 130.78, 127.47, 122.65, 114.02, 113.81, 113.65, 113.43, 84.08, 69.28, 67.49, 64.68,
+
62.81, 62.75, 24.71; HRꢀMS (ESI +), m/z: calculated for C31H33BFFeNO4 [Mꢀeꢀ] 569.1836,
found 569.1826; C, H, N analysis: calculated for C31H33BFFeNO4 (%) – C 65.41, H 5.84, N
2.46; found – C 65.39, H 5.96, N 2.56.
Compound 6 was prepared analogously to compound 4 using protocol 1. The product was
obtained as orange oil. Yield: 274 mg, 471 µmol (73%). TLC (SiO2, eluent petroleum
ether/acetone, 4/1, v/v) Rf = 0.51; 1HꢀNMR (400 MHz, acetonꢀd6): δ (ppm) 7.75 (d, 3J = 6.7 Hz,
2H), 7.4ꢀ7.27 (m, 3H), 6.92ꢀ6.84 (m, 3H), 5.27 (s, 2H), 5.00 (s, 2H), 4.47 (s, 2H), 4.14 (s, 5H),
ACS Paragon Plus Environment
23

Page 25 of 41
Journal of Medicinal Chemistry
1
2
13
3
3.99 (s, 2H), 1.35 (s, 12H); CꢀNMR (101 MHz, acetonꢀd6): δ (ppm) 160.49, 155.01, 141.03,
140.41, 135.14, 129.92, 127.50, 118.84, 112.55, 112.46, 101.79, 84.09, 69.28, 68.86, 67.37,
64.59, 62.82, 54.98, 53.71, 24.76; HRꢀMS (ESI+), m/z: calculated for C32H36BFeNO5 [Mꢀeꢀ]+
581.2036, found 581.2030; C, H, N analysis: calculated for C32H36BFeNO5 (%) – C 66.12, H
6.24, N 2.41; found – C 65.69, H 6.25, N 2.43.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 7 was prepared analogously to compound 4 using protocol 1. The product was
obtained as orange oil. Yield: 36 mg, 59 µmol (9%). TLC (SiO2, eluent petroleum ether/acetone,
1
3
6/1, v/v) Rf = 0.29; HꢀNMR (300 MHz, acetonꢀd6): δ (ppm) 7.77 (d, J = 7.2 Hz, 2 H), 7.37
3
(m, 4 H), 7.12 (d, J = 8.5 Hz, 2 H), 5.27 (s, 2 H), 5.02 (s, 2 H), 4.48 (s, 2 H), 4.14 (s, 5 H),
3.99 (s, 2 H), 2.27 (s, 3 H), 1.35 (s, 5 H); 13CꢀNMR (75 MHz, acetonꢀd6): δ (ppm) 168.30,
156.22, 149.51, 139.47, 135.85, 134.27, 126.99, 126.76, 121.47, 121.19, 83.20, 69.31, 68.71,
68.44, 68.35, 68.06, 63.73, 23.84, 19.62; HRꢀMS (ESI+), m/z: calculated for C33H36BFeNO6 [Mꢀ
eꢀ]+ 609.1986, found 609.1984; C, H, N analysis: calculated for C33H36BFeNO6 (%) – C 65.05, H
5.96, N 2.30;found – C 65.31, H 6.04, N 2.26.
Compound 8 was prepared analogously to compound 4 using protocol 1. TLC (SiO2, eluent
petroleum ether/acetone, 4/1, v/v) Rf = 0.36; 1HꢀNMR (300 MHz, acetonꢀd6): δ (ppm) 8.60 (d, 4J
3
4
= 2.1 Hz, 1H), 8.48 (dd, J = 4.8 Hz, J = 1.5 Hz, 1H) 7.75ꢀ7.68 (m, 3 H), 7.38ꢀ7.32 (m, 3 H),
5.25 (s, 2 H), 5.05 (s, 2 H), 4.46 (s, 2 H), 4.12 (s, 5 H), 4.00 (s, 2 H), 1.33 (s, 12 H); HRꢀMS
(ESI+), m/z: calculated for C30H33BFeN2O4 [Mꢀeꢀ] 552.1883, found 552.1891.
Compound 9 was prepared analogously to compound 4 using protocol 1. The product was
obtained as orange oil. Yield: 84 mg, 152 µmol (23%). TLC (SiO2, eluent petroleum
ether/acetone, 4/1, v/v) Rf = 0.35; 1HꢀNMR (300 MHz, acetonꢀd6): δ (ppm) 8.57 (d, 3J = 5.0 Hz,
ACS Paragon Plus Environment
24

Journal of Medicinal Chemistry
Page 26 of 41
1
2
3
4
5
6
7
8
9
1 H), 7.78ꢀ7.70 (m, 3 H), 7.3ꢀ7.27 (m, 4 H), 5.22 (s, 2 H), 5.07 (s, 2 H), 4.48 (s, 2 H), 4.11 (s, 5
H), 3.98 (s, 2 H), 1.33 (s, 12 H); 13CꢀNMR (75 MHz, acetonꢀd6): δ (ppm) 155.28, 149.36,
149.26, 140.72, 135.61, 135.23, 134.85, 128.00 (two overlapping peaks), 124.20, 102.15, 84.57,
69.77, 68.03, 65.23, 63.25, 52.07, 25.20; HRꢀMS (ESI+), m/z: calculated for C30H33BFeN2O4
[Mꢀeꢀ]+ 552.1883, found 552.1878; C, H, N analysis: calculated for C30H33BFeN2O4 (%) – C
65.25, H 6.02, N 5.07; found – C 65.28, H 5.97, N 5.08.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 10. Compound 3 (1.0 g, 2.17 mmol) was dissolved in anhydrous DMF (12 mL) under
a nitrogen atmosphere. Then Cs2CO3 (4.24 g, 13.01 mmol) and TBAI (4.79, 13.01 mmol) were
added to the solution. After stirring for 30 minutes at ambient temperature, diphenylmethyl
chloride (2.31 mL, 2.64 g, 13.01mmol) was added. The mixture was stirred for 7 days at 22 °C.
Then, the solvent was removed in vacuum (0.01 mbar) and the product was purified by column
chromatography on silica gel using DCM ꢀ DCM/ethylacetate (4/1, v/v) followed by a second
column using petroleum ether/acetone (6/1, v/v). The product was isolated as orange solid.
Yield: 78 mg, 124 µmol (6%). TLC (SiO2, eluent petroleum ether/acetone, 4/1, v/v) Rf = 0.5; 1Hꢀ
NMR (400 MHz, acetonꢀd6): δ (ppm) 7.70 (d, 3J = 7.1 Hz, 2 H), 7.39ꢀ 7.30 (m, 10 H), 7.23 (d, 3J
= 6.8 Hz, 2 H), 6.76 (s, 1 H), 5.11 (s, 2 H), 4.33 (s, 2 H), 4.10 (s, 5 H), 3.99 (s, 2 H), 1.35 (s, 12
13
H); CꢀNMR (101 MHz, acetonꢀd6): δ (ppm) 154.87, 140.64, 140.24, 135.03, 128.95, 128.46,
127.48, 127.44, 126.87, 102.65, 84.08, 69.46, 69.35, 67.22, 64.73, 64.56, 24.77; HRꢀMS (ESI+),
m/z: calculated for C37H38BFeNO4 [Mꢀeꢀ]+ 627.22448, found 627.22496; C, H, N analysis:
calculated for C37H38BFeNO4 (%) – C 70.84, H 6.11, N 2.23;found – C 70.41, H 6.04, N 2.07.
Compound 11. Compound 3 (500 mg, 1.08 mmol) was dissolved in anhydrous DMF (25 mL)
under a nitrogen atmosphere. Then Cs2CO3 (4.24 g, 13.01 mmol) and TBAI (4.79, 13.01 mmol)
were added to the solution. After stirring for 30 minutes at 22 °C, 1,2ꢀbis(chloromethyl)benzene
ACS Paragon Plus Environment
25

Page 27 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(569 mg, 3.25 mmol) was added. The mixture was stirred for 18 hours at 22 °C. Then, the
solvent was removed in vacuum (0.01 mbar) and the product was purified by column
chromatography on silica gel using petroleum ether/acetone (6/1, v/v) as eluent. Further
purification of the crude orange solid was carried out by recrystallization from aceton. Yield: 222
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
mg, 370 µmol (34%). TLC (SiO2, eluent petroleum ether/acetone, 6/1, v/v) Rf = 0.34; HꢀNMR
(400 MHz, acetonꢀd6): δ (ppm) 7.72 (d, 3J = 7.2 Hz, 2 H), 7.50 (d, 3J = 7.2 Hz, 1 H), 7.35 (m, 4
H), 7.20 (d, 3J = 7.6 Hz, 1 H), 5.28 (s, 2 H), 5.25 (s, 2 H), 4.91 (s, 2 H), 4.47 (s, 2 H), 4.19 (s, 5
13
H), 4.01 (s, 2 H), 1.35 (s, 12 H); CꢀNMR (101 MHz, CDCl3): δ (ppm) 154.24, 138.70, 136.50,
134.80, 134.46, 134.18, 133.25, 129.99, 128.96, 126.68, 125.01, 83.41, 69.76, 68.76, 67.21,
63.94, 62.14, 50.51, 43.82, 24.43; HRꢀMS (ESI+), m/z: calculated for C32H35BClFeNO4 [Mꢀeꢀ]+
599.1698, found 599.1706: C, H, N analysis: calculated for C32H35BClFeNO4 (%) – C 64.09, H
5.88, N 2.34; found – C 64.19, H 5.51, N 2.28.
Compound 12. Compound 3 (500 mg, 1.08 mmol) was dissolved in anhydrous DMF (5 mL)
under a nitrogen atmosphere. Then Cs2CO3 (883 mg, 2.71 mmol) and TBAI (799 mg,
2.17 mmol) were added to the solution. After stirring for 30 minutes at 22 °C,
1,2ꢀbis(chloromethyl)benzene (114 mg, 651 µmol) was added. The mixture was stirred for 18
hours at 22 °C. Then, the solvent was removed in vacuum (0.01 mbar) and the product was
purified by column chromatography on silica gel using petroleum ether/acetone (6/1, v/v) as
eluent. Yield: 198 mg, 193 µmol (30%). TLC (SiO2, eluent petroleum ether/acetone, 4/1, v/v) Rf
= 0.32; 1HꢀNMR (400 MHz, acetonꢀd6): δ (ppm) 7.74 (d, 3J = 7.6 Hz, 4 H), 7.29 (m, 8 H), 5.24
(s, 4 H), 5.10 (s, 4 H), 4.45 (s, 4 H), 4.16 (s, 10 H), 4.00 (s, 4 H), 1.35 (s, 24 H); 13CꢀNMR (101
MHz, CDCl3): δ (ppm) 154.68, 140.35, 135.41, 135.16, 127.29, 127.18, 125.66, 102.05, 84.10,
69.33, 67.42, 64.68, 62.91, 51.38, 24.76; HRꢀMS (ESI+), m/z: calculated for C56H62B2Fe2N2O8
ACS Paragon Plus Environment
26

Journal of Medicinal Chemistry
Page 28 of 41
1
2
[Mꢀeꢀ]+ 1024.3405, found 1024.3398; C, H, N analysis: calculated for C56H62B2Fe2N2O8 (%) – C
65.66, H 6.10, N 2.73; found – C 65.78, H 6.42, N 2.70.
3
4
5
6
7
8
9
Compound 14. Compound 139 was dissolved in anhydrous DMF (20 mL) under a nitrogen
atmosphere. Then Cs2CO3 (1.10 g, 3.40 mmol) and TBAI (1.25 g, 3.40 mmol) were added to the
solution. After stirring for 30 minutes at 22 °C, benzyl chloride (429 mg, 391 µL, 3.40 mmol)
was added. The mixture was stirred for 18 hours at 22 °C. Then, the solvent was removed in
vacuum (0.01 mbar) and the product was purified by column chromatography on silica gel using
petroleum ether/acetone (4/1, v/v) as eluent. Yield: 345 mg, 376 µmol (55%). TLC (SiO2, eluent
petroleum ether/acetone, 4/1, v/v) Rf = 0.46; 1HꢀNMR (300 MHz, acetonꢀd6): δ (ppm) 7.74 (d, 3J
= 6.8 Hz, 4 H), 7.24ꢀ7.35 (m, 14 H), 5.27 (s, 4 H), 4.98 (s, 4 H), 4.50 (s, 4 H), 3.95 (s, 4 H), 1.35
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13
(s, 24 H); CꢀNMR (101 MHz, acetonꢀd6): δ (ppm) = 140.35, 139.16, 135.16, 128.82, 128.46,
127.53, 127.24, 126.87, 126.70, 102.22, 84.08, 67.40, 65.72, 63.49, 53.04, 24.75; HRꢀMS
(ESI+), m/z: calculated for C52H58B2FeN2O8 [Mꢀeꢀ]+ 916.3740, found 916.3743; C, H, N
analysis: calculated for C52H58B2FeN2O8 (%) – C 68.15, H 6.38, N 3.06; found – C 68.09, H
6.45, N 3.11.
Cellular Assays
Cells and Cell Culture. The human promyelocytic leukemia cell line HLꢀ60 was obtained from
SigmaꢀAldrich (Germany). Burkitt lymphoma cell lines RAJI, BL2 and human mantle cell
lymphoma cell line JVMꢀ2 were obtained from DSZM (Germany). Cells were cultured
according to recommendations of DSMZ. In particular, RAJI and JVMꢀ2 cells were grown in
RPMI 1640 medium supplemented with 10% FBS, 1% Lꢀglutamine, and 1%
penicillin/streptomycin. BLꢀ2 cells were grown in RPMI 1640 medium supplemented with 20%
ACS Paragon Plus Environment
27

Page 29 of 41
Journal of Medicinal Chemistry
1
2
3
4
5
6
FBS, 1% Lꢀglutamine, and 1% penicillin/streptomycin. All cells studied here are nonꢀadherent.
Stock suspensions of the cells were grown up to 0.5ꢀ1.5 x 106 cells/mL and diluted as required.
7
8
9
Estimation of oxidative stress in cell cultures. An aliquot of cells was taken from the cultivation
medium. The medium was replaced with PBS buffer to obtain a cell suspension containing 106
cells/mL. A solution of 2’,7’ꢀdichlorodihydrofluorescin diacetate (DCFHꢀDA, 1 ꢁL, 5 mM in
DMSO) was added to the cell suspension (1 mL) and incubated in the dark for 30 min at room
temperature. Afterwards, the fluorescence of the samples (λex = 488 nm, λem = 530 nm) was
determined by using the flow cytometer.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Assay for Determination of Cell Permeability of the Prodrugs
Cells grown in RPMI 1640 medium supplemented with 10% FBS (20% for BLꢀ2), 1%
glutamine, and 1% penicillin/streptomycin were centrifuged, and the medium was replaced with
RPMI 1640 medium (5% FCS, 1% Lꢀglutamine, 1% penicillin/streptomycin) to obtain
suspensions containing 106 cells/mL. Solutions of prodrugs (10 ꢁL, solvent DMSO) were added
to the suspensions (1 mL) and incubated for 1 h. The final concentration of the prodrugs in the
suspensions was 50 ꢁM. Afterwards, cells were washed three times with PBS buffer (3 × 500
ꢁL) and treated with concentrated H2O2 solution (200 ꢁL, 1 M) for 30 min, and all volatiles were
removed by lyophilization. Dry, lysed cells were washed with water (200 ꢁL), and aqueous
solution obtained was acidified with HCl (400 ꢁL, 0.1 M). Then this solution was extracted with
2ꢀethylꢀ1,3ꢀhexanediol (100 ꢁL, 10% in CHCl3, v/v), and a portion of the organic phase obtained
(70 ꢁL) was mixed with H2SO4/CH3CO2H (400 ꢁL, 1/1, v/v). Curcumin solution in methyl
isobutyl ketone (250 ꢁL, 2 mg in 1 mL of the solvent) was added and allowed to react for 2 h.
The reaction was quenched by addition of water (1 mL). The light absorbance at 550 and 780 nm
of the organic phase was measured. The former value corresponds to absorbance of curcuminꢀ
ACS Paragon Plus Environment
28