Arylboronate prodrugs of doxorubicin as promising chemotherapy for pancreatic cancer

Article abstract of DOI:10.1016/j.bioorg.2019.103158

This study describes the synthesis of arylboronate-based ROS-responsive prodrugs of doxorubicin and their biological evaluation as anticancer agents. The determination of the most sensitive cancer type toward arylboronate prodrugs is crucial for further consideration of these molecules in clinical phase. To address this goal, an arylboronate-based profluorescent probe was used to compare the capacity of various cancer cell lines to efficiently convert the precursor into the free fluorophore. On the selected MiaPaCa-2 pancreatic cancer cells, a benzeneboronate prodrug exhibited 67% of the cytotoxicity obtained with the free doxorubicin. The prodrug was also able to induce tumor regression on MiaPaCa-2 pancreatic tumor model in ovo. Using this model, the amount of free doxorubicin liberated from this prodrug into the tumor was equivalent to the quantity measured after direct intratumoral injection of the same concentration of doxorubicin.

Full text of DOI:10.1016/j.bioorg.2019.103158

Journal Pre-Proof
Arylboronate prodrugs of doxorubicin as promising chemotherapy for pancre-
atic cancer
Charles Skarbek, Silvia Serra, Hichem Maslah, Estelle Rascol, Raphaël
Labruère
PII:
S0045-2068(19)30938-1
DOI:
https://doi.org/10.1016/j.bioorg.2019.103158
YBIOO 103158
Reference:
Bioorganic Chemistry
To appear in:
Received Date:
Accepted Date:
11 June 2019
24 July 2019
Please cite this article as: C. Skarbek, S. Serra, H. Maslah, E. Rascol, R. Labruère, Arylboronate prodrugs of
doxorubicin as promising chemotherapy for pancreatic cancer, Bioorganic Chemistry (2019), doi: https://doi.org/
10.1016/j.bioorg.2019.103158
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will
undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing
this version to give early visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier Inc.

JOURNAL PRE-PROOF
Arylboronate prodrugs of doxorubicin as promising
chemotherapy for pancreatic cancer
Charles Skarbek, Silvia Serra, Hichem Maslah, Estelle Rascol, Raphaël Labruère*
Institut de Chimie Moléculaire et des Matériaux d’Orsay (ICMMO), CNRS, Univ Paris Sud, Université
Paris-Saclay, 15 rue Georges Clemenceau, 91405 Orsay Cedex, France
Keywords: Prodrug, doxorubicin, reactive oxygen species, anticancer chemotherapy
* Corresponding author
E-mail address: raphael.labruere@u-psud.fr (R. Labruère).
Abstract: This study describes the synthesis of arylboronate-based ROS-responsive prodrugs
of doxorubicin and their biological evaluation as anticancer agents. The determination of the
most sensitive cancer type toward arylboronate prodrugs is crucial for further consideration of
these molecules in clinical phase. To address this goal, an arylboronate-based profluorescent
probe was used to compare the capacity of various cancer cell lines to efficiently convert the
precursor into the free fluorophore. On the selected MiaPaCa-2 pancreatic cancer cells, a
benzeneboronate prodrug exhibited 67 % of the cytotoxicity obtained with the free doxorubicin.
The prodrug was also able to induce tumor regression on MiaPaCa-2 pancreatic tumor model
in ovo. Using this model, the amount of free doxorubicin liberated from this prodrug into the
tumor was equivalent to the quantity measured after direct intratumoral injection of the same
concentration of doxorubicin.
Introduction
Cancer caused more than 9 million deaths worldwide in 2018 and the number of new cases is
constantly increasing.[1] Among different approaches, chemotherapy is largely used in medical
oncology. Nonetheless, most of the currently employed chemotherapies against cancer are
almost as toxic for the normal cells as for the tumor cells. Thus, there is an urgent need for the
development of new and specific chemotherapies, capable of exploiting the differences between
normal and tumor cells in order to increase the potency and therapeutic index by improving the
balance between efficacy and toxicity.[2] During the last decades, the prodrug strategy has
emerged, involving the development of activity-masked molecules designed for being rapidly
activated through enzymatic or chemical reactions inside the tumors or within the tumor
microenvironment. This approach ideally ensures a high selectivity against targets and low
toxicity, since drugs are liberated specifically at the tumor area, generally by targeting the
overexpression of enzymes or reactive oxygen species (ROS) surrounding the tumor, or the
hypoxic environment of the tumor. Furthermore, the prodrug strategy remains faster than the
development of novel therapeutic agents with appropriate drug-like characteristics.[3]
Among various metabolic pathways, the activation of antitumor prodrugs by ROS seems
particularly promising.[4] Biological study of tumors have revealed that cancer cells have a higher
level of oxidative stress compared to healthy cells. This oxidative stress results in the

JOURNAL PRE-PROOF
overproduction of ROS including hydrogen peroxide (H2O2), hydroxyl radical (HO●) and
●-
superoxide anion (O2 ). These ROS contribute significantly in different tumor progression
processes such as cell proliferation and angiogenesis.[5] Neovascularization is a critical element
of tumor metastasis since it allows access of certain cancer cells to the bloodstream leading to
colonization of new organs.[6] In addition, H2O2 molecules produced in tumor cells will generate
migration of these cells, which is also an important factor in the metastatic process.[5]
Arylboronic acids and their corresponding esters have been widely used as temporary drug
masking groups since activation is made possible by ROS (Scheme 1).[7,8] These temporary
moieties are triggered by H2O2, the most abundant and stable ROS produced by cancer cells,
to further afford non-toxic boric acid.[9] Boronic acid groups were either directly linked to the drug
or branched through a quinone methide-based self-immolative spacer. Oxidation of the carbon-
boron bond will allow the formation of a hydroxy group. In the case of the self-immolative spacer,
the latter electron donor function will cause electronic delocalization within the aromatic nucleus
leading to the release of the drug.[10] Remarkably, arylboronic acid spacers are employed to link
either alcohol or amine function-bearing drugs through ether, carbonate or carbamate bonds.
The spacer is also commonly used to hinder the drug and allow a better masking of the biological
activity.
Boronic acid/boronate prodrug
RO
H2O2
Masked
drug
Active
drug
B
HO
RO
B(OH)3 + 2 ROH
R = H or alkyl
Arylboronic acid/boronate prodrug
RO
OR
O
B
OH
OH
H2O2
Spontaneous
H2O
B(OH)3 + 2 ROH
Masked
drug
X
Masked
drug
OH
X
Active
drug
R = H or alkyl
X = -O-
X = -O(C=O)O-
X = -O(C=O)NH-
Scheme 1. Conversion of boron-containing prodrug into active drug.
This strategy has led to the development of probes for tumor imaging,[11,12] ROS amplifiers[13-15]
DNA cross-linking agents[16,17] and prodrugs of matrix metalloproteinase inhibitors,[18] nitrogen
mustards,[19,20] camptothecin derivatives,[21,22] endoxifen,[23,24] methotrexate,[25] cisplatin,[26]
HDAC inhibitors,[27,28] 5-fluorouracil[29] and doxorubicin.[30]
From this literature, we observed that the different arylboronic acid and ester precursors were
evaluated on various cell lines with disparate cytotoxic results. The cancer cell line sensitivity to
arylboronic prodrugs does not follow a clear trend from one study to another. In regards to a
particular cell line, the structure of the arylboronic acid derivative has therefore a strong
influence on the biological activity. The clear identification of the type of cancer sensitive to these

JOURNAL PRE-PROOF
boron-based prodrugs is central since none of these molecules has entered clinical trial so far.
In this work, we decided to develop potent arylboronate-drug conjugates and further select the
most receptive target among several cancer cell lines in order to identify the type of cancer that
responds well towards these precursors. First, we selected doxorubicin (DOX), a good
candidate for the design of prodrug. Indeed, DOX is a very effective antitumor drug with limited
therapeutic efficacy due to stochastic repartition in the patient organs resulting in important
cardiotoxicity.[31] Arylboronate-based self-immolative spacers were then used in order to branch
the free amine function of DOX (Figure 1). Several arylboronate moieties were used in order to
study their influence on the kinetic of DOX release in presence of H2O2. Three different scaffolds
were envisioned: (i) the unsubstituted benzeneboronate; (ii) a fluorinated benzene homolog
since electron withdrawing atom were shown to increase the kinetic of oxidation;[32] (iii) a furan
ring, described as effective self-immolative spacer,[33] to add a strong structural modification in
comparison with the two aforementioned benzene moieties.
OH
O
O
OH
OH
Dororubicin (DOX)
H
OMe O
OH
O
O
OH
O
HN
O
O
Arylboronate promoiety
B
O
Figure 1. General structure of the arylboronate prodrugs of doxorubicin.
Six cancer cell lines originating from different organs (breast cancer MCF-7 and the resistant
counterpart MCF-7 MDR, hepatocellular carcinoma Hep G2, lung adenocarcinoma A549,
glioblastoma U87, pancreatic cancer cell line MiaPaCa-2) were compared. An arylboronate
profluorescent probe was used to determine the most effective cell line for the oxidation of the
carbon-boron bond. Prodrugs of doxorubicin were then prepared and their cytotoxicity was
assessed on the different cell lines. The most promising compound was further tested using in
ovo tumor model via the HET-CAM (Hen’s Egg Test-Chorioallantoïc Membrane) assay. This
test has been widely used both for screening anticancer drugs[34-36] and for human tumor
growth.[37,38]
Results and Discussion
Synthesis
The arylboronate profluorescent probe 3 was prepared in one step by reacting coumarin 2 with
4-bromomethylphenylboronic acid pinacol ester 1 in the presence of Cs2CO3 at room
temperature for 4h (Scheme 2).

JOURNAL PRE-PROOF
O
O
O
B
B
O
HO
O
O
a
O
O
O
CF3
CF3
Br
3
2
1
Scheme 2. Synthesis of the arylboronate profluorescent probe 3. Reagents and conditions: (a)
Cs2CO3, DMF, r.t., 4h.
Prodrugs 8a-c were synthesized in two steps, starting from aryl alcohols 6a-c(Scheme 3).
Firstly, boronate 6b was prepared through palladium catalyzed bromine-boron exchange
starting from compound 4.[39] Ester 6c was obtained in 79 % yield by reacting commercially
available boronic acid 5 with pinacol in THF. Alcohol activation of 6a-c was carried out in
presence of 4-nitrophenyl chloroformate to provide activated compounds 7a-c. DOX was then
linked to the arylboronate promoiety through its amine group using an addition-elimination
reaction to afford prodrug 8a-c. A similar reaction was used to prepare the control molecule 9,
a benzeneboronate precursor of non-toxic phenylalanine. Another control molecule, the
carboxylic acid 12 which is a bioisostere derivative of 8a, unable to liberate DOX was prepared:
(i) activated carbonate 11 was obtained from commercially available 4-(hydroxymethyl)benzoic
acid 10 in 74 % yield; (ii) DOX was then branched according to an addition-elimination reaction
(Scheme 3).

JOURNAL PRE-PROOF
O
O
B
O
O
O
B
OH
6a
O
O
O
B
O
OH
HO
OH
NH
O
O
O
Br
B
O
H
O
d
c
a
OH
O
HO
O
F
F
O
O
OH
OH
OH
4
6b
OMe
NO2
8a
8b
8c
(phenyl)
(fluorophenyl)
(furanyl)
7a
7b
7c
(phenyl)
(fluorophenyl)
(furanyl)
HO
O
O
O
B
b
B
O
e
OH
OH
6c
5
O
HO
HN
O
O
O
B
O
9
COOH
COOH
O
OH
HO
O
OH
O
COOH
N
H
d
O
c
H
O
O
HO
O
O
O
OH
O
OH
10
OMe
NO2
12
11
Scheme 3.Synthesis of boronate prodrugs and controls. Reagents and conditions: (a) AcOK,
.
Pd(dppf)Cl2 CH2Cl2, B2(Pin)2, DMF, 80°C; (b) pinacol, THF, 40°C; (c) 4-nitrophenyl
chloroformate, base, DCM, 0°C to r.t.; (d) DOX.HCl, HOBt, Et3N, DMF, r.t.; (e) phenylalanine,
HOBt, Et3N, DMF, r.t.
H2O2 activation of the profluorescent probe
The in vitro activation of arylboronate profluorescent probe 3 was carried out using increasing
equivalent of H2O2 (1, 5, 10 and 50 equivalents). We recorded the temporal evolution of the free
coumarin 2 fluorescence after oxidation and further self-immolation from precursor 3 (Figure

JOURNAL PRE-PROOF
2A). We observed that high concentrations of H2O2 led to the fastest kinetic of coumarin release.
The use of 5, 10 and 50 equivalents led to the highest fluorescence release with half time values
of 5.5, 5 and 2.5 hours, respectively. As expected, these results showed a bimolecular kinetic
of oxidation for our benzeneboronate profluorescent probe 3. The use of equimolar quantity of
H2O2 leads to slow release of coumarin in line with the known sluggish kinetic of reaction
between benzeneboronate and H2O2.[40] Moreover, the release is slowed down since the
coumarin is linked to the self-immolative spacer via an ether link.[41]
Cell line screening
The capacity of each tested cell line to produce ROS was measured using the benzeneboronate
profluorescent probe 3. The slow kinetic profile of disassembly is essential to generate enough
contrast between the different cell lines toward their ability to produce different level of ROS.
After six hours of incubation with each cell line, the release of coumarin from 3 was quantified
by spectrofluorimetry (Figure 2B).
A
B
2500
2000
1500
1000
500
***
***
10000
8000
6000
4000
2000
0
***
***
0
R
U87
Ca-2
a
A549
MCF-7
P
Hep G2
Mia
0
2
4
6
8
10
12
MCF-7 MD
Time H
Cell line
Figure 2.(A) Coumarin release from the arylboronate profluorescent probe 3 in presence of
H2O2 (▼ 50 eq., ▲ 10 eq., ◼ 5 eq. and ● 1eq.). (B) Investigation of the capacity of ROS
production by cell lines (*statistically significant difference compared to U87 cell line with p <
0.0001).
This intensity relates the ability of each tested cell line to produce the ROS responsible for the
oxidation of the boron-carbon bond leading to the release of coumarin. Since oxidation and
subsequent self-immolation of precursor 3 is rather slow and dependent of the ROS
concentration, we chose to: (i) incubate 3 in the presence of a large number of cells in order to
produce as much ROS as possible using six-well plates; and (ii) maintain contact for 6 hours to
obtain a significant amount of liberated coumarin. The A549 and U87 cell lines demonstrated a
low production of ROS as only a weak fluorescence intensity was observed. On the other hand,
Hep G2, MiaPaCa-2, MCF-7 and MCF-7 MDR revealed a higher production of ROS as the
fluorescence intensity was much higher compared to the other tested cell lines. Further, Hep
G2 and MiaPaCa-2 were the most active cell lines for conversion of precursor 3 into coumarin
2 and these two cell lines seem to be the most suitable for further investigation of ROS-
activatable prodrugs. The comparison of the ROS production by these cancer cells versus the
cell line producing the least ROS (U87) was unequivocal and confirmed using a Kruskall-Wallis
test with p < 0.0001.

JOURNAL PRE-PROOF
In vitro cytotoxicity
As opposed to the probe 3, the prodrugs are built of an aryl self-immolative spacer bonded to
DOX through a carbamate linker to enhance the kinetic of disassembly[41] as it is beneficial for
the biological activity to release high level of drug rapidly (data not shown). The IC50 of all the
designed prodrugs (8a-c) and controls (9 and 12) were evaluated on the panel of cell line and
the results are depicted in Table 1. No activity was observed on the multidrug resistant cell line
MCF-7 MDR. DOX showed an important cytotoxic activity on all the other tested cell lines with
nanomolar IC50 ranging from 10 to 600 nM. The cell lines U87, MCF-7, Hep G2 and MiaPaCa-
2 were sensitive to DOX withIC50 around 100 nM. A549 cell line on the other hand demonstrated
a lower activity with an IC50 of 600 nM. The control compounds 9 and 12 had no activity on the
studied cell lines even at our highest tested concentration of 50 µM. These results are coherent
as these control molecules have no trigger unit (12) or deliver a non-toxic motif (9).
Table 1. In vitro cytotoxicity of 8a-c, 9 and 12 on the different cell lines. Percentage of restored
activity is displayed in bracket. Nd: Not determined.
IC50 (M) ± SD
Cell line
DOX
8a
8b
8c
9 / 12
4.50 ± 0.40
(2 %)
11.10 ± 0.80
(5 %)
3.50 ± 0.20
(3 %)
4.30 ± 0.35
(14 %)
U-87
0.10 ± 0.02
0.60 ± 0.10
Nd
> 50
> 50
8.40 ± 0.90
(7 %)
A-549
3.00 ± 0.20
(3 %)
> 50
0.30 ± 0.01
(67 %)
7.80 ± 0.50
(1 %)
> 50
8.30 ± 0.70
(2 %)
4.90 ± 0.60
(2 %)
> 50
1.20 ± 0.20
(17 %)
MCF-7
0.09 ± 0.01
> 50
> 50
> 50
> 50
MCF-7 MDR
MiaPaca-2
0.20 ± 0.00
4.00 ± 0.40
(4 %)
7.70 ± 0.40
(2 %)
3.40 ± 0.20
(4 %)
Hep G2
0.15 ± 0.02
> 50
All the ROS-activatable doxorubicin prodrugs (8a-c) demonstrated activities on the DOX-
sensitive cell line and variable micromolar activities were obtained ranging from 0.1 to 10 µM
underlining moderate activities of the tested prodrugs. The activity of the different prodrugs on
a particular cell line was comparable except in the case of MiaPaca-2 where 8a was more
cytotoxic than 8b and 8c. Interestingly, 8a showed activity on the MiPaCa-2 cell line with an IC50
close to the one obtained with the free DOX, 0.3 µM for 8a versus 0.2 µM for DOX. Although
the conversion of 8a is not complete: ~ 67 %, as deducted from the activity obtained with the
same concentration of free doxorubicin, we observe a greater conversion than those obtained
for related structures (~ 40 %).[22]8a was however not as active against the Hep G2 cells (IC50
= 4 µM, ~ 4 % estimated conversion), a cell line able to produce similar level of ROS as the
MiPaCa-2 cell line. As described by Doskey et al.,[42] the H2O2 removal rate constants differ
greatly between cell lines. In exponential growth phase, Hep G2 cells were shown to remove
H2O2 four times faster than the MiaPaCa-2 cells. In this study, the cytotoxic activities were

JOURNAL PRE-PROOF
determined after 72 hours of incubation when cells have reached their exponential growth phase
and this can therefore rationalize the strongest activity of 8a against MiaPaCa-2 in comparison
with Hep G2 cell lines.
Chick Chorioallantoic Membrane (CAM) Assay and HPLC Detection of DOX
The HET-CAM assay was developed to investigate the efficacy of our most active arylboronate
doxorubicin prodrug 8a using the MiaPaCa-2 pancreatic cancer cell line for tumor production on
the CAM as it showed the most promising cytotoxicity and an elevated release of ROS. Different
protocol of intratumoral injection of 8a or DOX were performed. In our case and according to
the literature, we first determined the highest dose of DOX that was not lethal for the embryo.[36]
Our maximum dose of 184 nmol (100 µg) of DOX in a single injection was well tolerated in a
statistically significant sample size of chick embryos. For the evaluation of the tumor regression,
we chose to enhance the number of injections rather than testing a high concentration as a
single dose taking into account the poor solubility of 8a. Optimized protocol was found with the
administration of DOX or 8a (20 nmol) intratumorally twice a day and two days in a row followed
by an observation of the tumor size after three days. This protocol led us to observe a significant
decrease of the tumor in DOX group (~ 50 % regression) and a comparable result was obtained
for 8a (~ 50 % regression) (Figure 3). Statistically relevant images of MiaPaCa tumor implanted
on the CAM are shown in Figure 3. Necrosis seen in Figure 3B was noticed in several
experiments as the result of DOX/8a action. In this vascularized tumor model, the tested drug
has also some antivascular effect often leading to partial necrosis of the tumor.
Figure 3. In ovo efficacy of 8a compared to DOX after two intratumoral injections of each drug
(20 nmol) twice a day for two consecutive days, white arrows indicate the tumors, scale bar: 0.5
cm. (A) left: tumor before injection of DOX; right: tumor three days post injection of DOX (50 %
tumor regression). (B) left: tumor before injection of 8a; right: tumor three days post injection of
8a (tumor necrosis and 50 % regression). Image treatment and tumor measurements were
fulfilled using FIJI software.[43]
In order to confirm that the prodrug activity was due to its conversion into free DOX, a HPLC
method for the measurement of DOX concentration in the tumors was elaborated. Tumors were
extracted from the chorioallantoic membrane and sonicated to recover the free DOX. The same
amount of DOX was found in both groups (30 % of the initial dose) which is coherent with the

JOURNAL PRE-PROOF
tumor regression observed in ovo whereas prodrug 8a was not detected. These results show
that, in presence of a well-developed tumor, the prodrug 8a is converted into DOX.
In this study, no strong tumor regression or total resection was observed. Indeed, in the HET-
CAM assay, the dose escalation is difficult to study as it can lead to the death of the embryo
resulting in a false positive result. Taking into account the ethical regulation and the small
window between the drug injection and the end point before the embryo’s birth, observation of
tumor evolution over a long period is not possible. Therefore, we were able to only monitor
partial regression of the tested tumors.
Conclusion
A profluorescent probe was used to select a cancer cell line strongly producing ROS. Among
the six cell lines tested, Hep G2 and MiaPaCa-2 were the most effective at oxidizing the
benzeneboronate probe. Prodrugs of doxorubicin and controls were then synthesized and their
cytotoxic activity was determined in vitro on our panel of cell lines. Prodrug 8a was selectively
active on pancreatic cancer cells MiaPaca-2 with a recovery of 67 % of the activity of the free
doxorubicin. Further testing in ovo allowed us to observe the behavior of this prodrug on a
MiaPaCa-2 tumor model. Similar tumor regression was observed using the same dose of both
the prodrug 8a and the free drug counterpart on the available timescale of reading.
Quantification of the doxorubicin released from 8a by oxidation in pancreatic tumor in ovo
confirmed the latter result. The development of benzeneboronate-containing prodrugs must be
pursued for the treatment of certain cancer types and we herein highlighted the promising
potential of such prodrug in the treatment of pancreatic adenocarcinoma.
Experimental Section
Synthesis
All chemical reagents were of analytical grade, obtained from Acros, Alfa Aesar, or Aldrich, and
used without further purification. Solvents were obtained from SDS or VWR-Prolabo.
Dichloromethane (DCM) was dried on molecular sieves and used immediately. Chromatography
was performed using silica gel (35-70 μm, Merck). Concentration of solutions was performed
under reduced pressure at temperature below 40°C using rotary evaporator. Analytical TLC was
performed using Silica Gel 60 F254 pre-coated aluminum plates (Merck). Spots were visualized
by treatment with ninhydrine revelator followed by heating and/or by absorbance of UV light at
254 nm or fluorescent light at 360 nm. NMR spectra were collected on Bruker DPX 250 (1H at
250 MHz and 13C at 62.5 MHz), AV 300 (1H at 300 MHz and 13C at 75 MHz) or AV 360 (1H at
360 MHz and 13C at 90 MHz) spectrometers and analyzed using MestReNova software.
Chemical shifts are reported in ppm (δ) and coupling constants in Hz (J). NMR spectra were
performed in CDCl3 or (CD3)2SO. High-resolution mass spectrometry (HRMS) analyses were
performed by electrospray with positive or negative (ESI+ or ESI-). The purity of all compounds
used for biological activity test was checked by reverse phase analytical HPLC and confirmed
to be ≥ 95 %.
Compound 3

JOURNAL PRE-PROOF
To a solution of bromo derivative 1 (41 mg, 0.13 mmol), coumarin 2 (30 mg, 0.13 mmol) in 5 mL
of DMF was added Cs2CO3 (46 mg, 0.14 mmol). The reaction was kept under agitation at r.t.
for 4 h before being quenched with a saturated solution of NaHCO3. The suspension was
extracted with EtOAc (3 x 20mL), the organic layers were combined, washed with brine, dried
with MgSO4, and concentrated under reduced pressure to give a crude product. The crude
product was purified by column chromatography over silica gel with Cyclohexane-EtOAc: 100-
0 to 60-40 to give the expected product as a white solid (35 mg, yield: 60 %).
1H NMR (300 MHz, CDCl3) δ 7.85 (d, J = 8.0 Hz, 2H), 7.63 (dd, J = 9.0, 1.8 Hz, 1H), 7.43 (d, J
= 8.0 Hz, 2H), 6.99 (dd, J = 9.0, 2.6 Hz, 1H), 6.93 (d, J = 2.5 Hz, 1H), 6.62 (s, 1H), 5.19 (s, 2H),
13
1.35 (s, 12H). C NMR (75 MHz, CDCl3) δ 162.45, 161.23, 159.37, 156.20, 138.35, 135.23,
126.54, 126.38, 123.43, 119.77, 113.99, 112.38, 107.23, 102.49, 83.92, 70.51, 24.84. HRMS
(ESI): m/z Calcd for C23H22BF3O5Na, [M+Na]+: 469.1410, found: 469.1409.
Compound 7a
To a solution of 4-nitrophenyl chloroformate (264 mg, 1.26 mmol) in 3 mL of DCM at 0°C was
added dropwise Et3N (177 µL, 1.26 mmol). The mixture was stirred during 20 min under argon
then a solution of 6a (100 mg, 0.42 mmol) in 4 mL of DCM was added dropwise to the first
solution and the mixture was stirred during 3 h at r.t. The reaction was quenched with brine,
extracted with DCM (3 x 20 mL), the organic layers were combined, dried with MgSO4 and
concentrated under reduced pressure to give a crude product. The crude product was purified
by column chromatography over silica gel with cyclohexane-EtOAc: 90-10 to 60-40 to give the
expected product (135 mg, yield: 80 %).
1H NMR (300 MHz, CDCl3) δ 7.88 – 7.82 (m, 2H), 7.46 – 7.41 (m, 2H), 7.40 – 7.35 (m, 2H), 5.31
13
(m, 2H), 1.35 (s, 12H). C NMR (75 MHz, CDCl3) δ 155.47, 152.38, 145.36, 137.02, 135.15,
127.58, 127.23, 125.25, 121.74, 83.95, 70.75, 24.82.
HRMS (ESI): m/z Calcd for C20H22BNO7Na, [M+Na]+ :422.1387, found: 422.1367.
Compound 8a
To a solution of doxorubicine hydrochloride (10 mg, 0.017 mmol) in 1 mL of DMF was added
Et3N (2.4 µL, 0.017 mmol). The mixture was stirred for 30 min at r.t. before adding HOBt (2.3
mg, 0.017 mmol) and a solution of 7a (6.8 mg, 0.017 mmol) in 1 mL of DMF and the resulting
mixture was stirred a r.t. during 3 hours. The reaction was quenched with brine, extracted with
DCM (3 x 20 mL), the organic layers were combined, dried with MgSO4 and concentrated under
reduced pressure. The crude product was purified over silica gel with DCM-MeOH: 90-10 to
give the expected product (10 mg, yield: 78 %).
1H NMR (300 MHz, CDCl3) δ 13.96 (s, 1H), 13.20 (s, 1H), 8.02 (d, J = 7.6 Hz, 1H), 7.82 – 7.73
(m, 3H), 7.42 – 7.36 (m, 1H), 7.29 (d, J = 9.5 Hz, 2H), 5.51 (d, J = 2.5 Hz, 1H), 5.23 (dd, J =
17.3, 8.8 Hz, 2H), 5.04 (s, 2H), 4.77 (s, 2H), 4.57 (s, 1H), 4.15 (d, J = 6.5 Hz, 1H), 4.08 (s, 3H),
3.88 (s, 1H), 3.68 (s, 1H), 3.25 (d, J = 18.9 Hz, 1H), 2.97 (d, J = 7.5 Hz, 1H), 2.91 (d, J = 10.7
Hz, 1H), 2.33 (d, J = 14.4 Hz, 1H), 2.16 (dd, J = 14.7, 3.7 Hz, 2H), 1.33 (s, 12H), 1.24 (s, 3H).
13C NMR (75 MHz, CDCl3) δ 214.03, 187.14, 186.74, 161.13, 156.29, 155.71, 155.60, 139.46,
135.88, 135.55, 135.07, 133.68, 127.23, 120.90, 119.94, 118.55, 111.65, 111.48, 100.87, 83.93,
69.79, 69.65, 67.41, 66.74, 65.68, 56.77, 47.12, 45.94, 35.72, 34.07, 30.27, 29.81, 24.95, 16.97.
HRMS (ESI): m/z Calcd for C41H46BNO15Na, [M+Na]+: 826.2858, found: 826.2860.

JOURNAL PRE-PROOF
Compound 6c
To a solution of boronic acid 5 (3.5 mmol, 500 mg) in anhydrous THF (15 mL) was added and
pinacol (3.5 mmol, 416 mg). The resulting solution was evaporated under reduced pressure at
40 °C. This procedure was repeated until TLC analysis indicated a complete conversion. The
crude product was purified by column chromatography over silica gel with hexane-EtOAc: 90-
10 to 50-50 to give the expected product (618 mg, yield: 79 %).
1H NMR (300 MHz, CDCl3) δ 7.02 (d, J = 3.3 Hz, 1H), 6.34 (d, J = 3.3 Hz, 1H), 4.66 (s, 2H),
1.34 (s, 12H). 13C NMR (75 MHz, CDCl3) δ 159.07, 124.32, 108.44, 84.17, 75.02, 57.56, 24.62.
HRMS (ESI): m/z Calcd for C11H17BO4Na, [M+Na]+: 247.1117, found: 247.1114.
Compound 7b
To a solution of 4-nitrophenyl chloroformate (175 mg, 0.87 mmol) in 5 mL of DCM at 0°C was
added pyridine (68 µL, 0.87 mmol), the solution was stirred during 5 min at 0°C then a solution
of 6b[39] (73 mg, 0.29 mmol) in 1 mL of DCM was added dropwise to the mixture and the reaction
was stirred at r.t. overnight. The reaction was then quenched with a saturated solution of
NaHCO3, the aqueous layer was extracted with EtOAc (3 x 20 mL), the organic layers were
combined, washed with brine, dried with MgSO4, and concentrated under reduced pressure to
give a crude product. The crude product was purified by column chromatography over silica gel
with Cyclohexane-EtOAc: 95-5 to 60-40 to give the expected product (25 mg, yield: 20 %).
1H NMR (300 MHz, CDCl3) δ 8.34 (dd, J = 1.7; 7.2 Hz, 2H), 8.27 (dd, J = 1.7; 7.2 Hz, 2H), 7.63-
7.59 (m, 1H), 7.52-7.50 (m, 1H), 7.40 – 7.38 (m, 1H), 5.38 (s, 2H), 1.35 (s, 12H). 13C NMR (75
MHz, CDCl3) δ 162.05, 159.20, 152.30, 145.40, 130.10, 125.40, 125.30, 124.10, 121.70,
121.60, 121.20, 64.60, 24.70. HRMS (ESI): m/z Calcd for C20H21BFNO7Na, [M+Na]+: 440.1292,
found: 440.1276.
Compound 7c
To a solution of 4-nitrophenyl chloroformate (504 mg, 2.5 mmol) in 15 mL of DCM was added
DMAP (54 mg, 0.44 mmol) and the reaction was stirred et 0°C during 15 min. In another round
bottom flask, a solution 6c (500 mg, 2.2 mmol) and Et3N (612 µL, 4.4 mmol) in 15 mL of DCM
was stirred at 0°C. The second mixture was added dropwise on the first mixture to give a yellow
solution. The solution was stirred during 6 hours at r.t. before being quenched with water. The
aqueous layer was extracted with DCM (3 x 20 mL), the organic layers were combined, washed
with brine, dried with MgSO4, and concentrated under reduced pressure to give a crude product.
The crude product was purified by column chromatography over silica gel with DCM-MeOH:
100-0 to 90-10 to give the expected product (250 mg, yield: 30 %)
1H NMR (300 MHz, CDCl3) δ 8.24 – 8.17 (m, 2H), 7.07 (d, J = 3.3 Hz, 1H), 7.05 – 6.98 (m, 2H),
6.51 (d, J = 3.3 Hz, 1H), 5.14 (s, 2H), 1.36 (s, 12H). 13C NMR (75 MHz, CDCl3) δ 163.12, 153.56,
126.22, 125.90, 124.04, 115.62, 114.69, 113.20, 111.12, 84.48, 62.91, 24.72.HRMS (ESI): m/z
Calcd for C18H20BNO8Na, [M+Na]+: 412.1179, found: 412.1161
Compound 8b
To a solution of doxorubicin hydrochloride (35 mg, 0.06 mmol) in 7 mL of DMF was added Et3N
(8 µL, 0.06 mmol). The mixture was stirred for 30 min at r.t. before adding HOBt (8 mg, 0.06
mmol) and a solution of 7b (25 mg, 0.06 mmol) in 2 mL of DMF and the resulting mixture was

JOURNAL PRE-PROOF
stirred a r.t. during 6 hours. The reaction was quenched with brine, the aqueous layer was
extracted with extracted with DCM (3 x 20 mL), the organic layers were combined, dried with
MgSO4 and concentrated under reduced pressure. The crude product was purified over silica
gel with DCM-MeOH: 90-10 to give the expected product (30 mg, yield: 60 %).
1H NMR (300 MHz, CDCl3) δ 13.94 (s, 1H), 13.25 (s, 1H), 8.02 (d, J = 7.4 Hz, 1H), 7.80 (t, J =
8.6 Hz, 1H), 7.41 (d, J = 8.6 Hz, 1H), 7.23 (d, J = 10.4 Hz, 3H), 5.60 (t, J = 7.2Hz, 1H), 5.35 (s,
1H), 5.21 (s, 1H), 4.77 (s, 2H), 4.54 (d, J = 8.6 Hz, 1H), 4.25 (d, J = 10.3Hz, 1H), 4.09 (s, 4H),
3.68 (s, 1H), 3.28 (d, J = 21.8 Hz, 2H), 3.10-2.90 (m, 2H), 2.55 (d, J = 13.2 Hz, 1H), 2.24-2.11
(m, 2H), 1.65– 1.50 (m, 6H), 1.35 – 1.20 (m, 12H). 13C NMR (75 MHz, CDCl3) δ 161.05, 156.14,
15.70, 135.85, 135.50, 133.80, 133.30, 120.86, 119.90, 118.50, 111.60, 99.00, 75.80, 69.30,
65.45, 64.60, 56.70,47.60, 35.90, 33.80, 29.70, 29.50,15.80. HRMS (ESI): m/z Calcd for
C41H45BFNO15Na, [M+Na]+: 844.2764, found: 844.2749.
Compound 8c
To a solution of doxorubicin hydrochloride (19 mg, 0.033 mmol) in 4 mL of DMF was added Et3N
(4 µL, 0.033 mmol). The mixture was stirred for 30 min at r.t. before adding HOBt (5 mg, 0.033
mg) and a solution of 7c (13 mg, 0.033mmol) in 1 mL of DMF. The resulting solution was stirred
at r. t. overnight before being quenched with brine, extracted with DCM (3 x 20mL), the organic
layers were combined, dried with MgSO4 and concentrated under reduced pressure. The crude
product was purified over silica gel with DCM-MeOH: 90-10 to give the expected product (15
mg, yield: 60 %)
1H NMR (300 MHz, CDCl3) δ 13.95 (s, 1H), 13.22 (s, 1H), 8.07 (d, J = 9.1 Hz, 1H), 7.78 (t, J =
8.1 Hz, 1H), 7.38 (d, J = 8.1 Hz, 1H), 6.87 (d, J = 9.0 Hz, 1H), 5.49 (d, J = 3.4 Hz, 1H), 5.27 (s,
1H), 5.10 (d, J = 8.7 Hz, 1H), 4.76 (s, 2H), 4.56 (s, 1H), 4.07 (s, 6H), 3.84 (s, 1H), 3.72 – 3.59
(m, J = 9.2, 5.9 Hz, 1H), 3.26 (dd, J = 18.9, 1.4 Hz, 1H), 2.97 (s, 3H), 2.89 (s, J = 4.3 Hz, 3H),
2.33 (d, J = 14.6 Hz, 1H), 2.22 – 2.11 (m, 2H), 1.91 – 1.70 (m, 2H), 1.37 – 1.13 (m, 12H). 13C
NMR (75 MHz, CDCl3) δ 212.84, 186.08, 185.64, 161.86, 161.73, 160.01, 155.16, 154.93,
154.60, 139.80, 134.79, 134.44, 132.54, 125.04, 119.79, 118.84, 117.43, 114.59, 110.53,
110.36, 99.71, 68.57, 66.30, 64.52, 59.99, 55.65, 52.70, 45.84, 35.62, 34.62, 32.83, 30.55,
29.18, 28.66, 28.20, 23.80, 23.69, 21.66, 15.82, 13.50. HRMS (ESI): m/z Calcd for
C39H44BNO16Na, [M+Na]+: 816.2650, found: 816.2832.
Compound 9
To a solution of 7a (30 mg, 0.075 mmol) in 7 mL of DMF was added Et3N (11 µL, 0.075 mmol).
The mixture was stirred for 30 min at r.t. before adding HOBt (11 mg, 0.075 mg) and a solution
of phenylalanine (12.4 mg, 0.075 mmol) in 3 mL of DMF and the resulting mixture was stirred at
r.t. overnight. The reaction was quenched with brine, extracted with DCM (3 x 20 mL), the
organic layers were combined, dried with MgSO4 and concentrated under reduced pressure.
The crude product was purified over silica gel with DCM-MeOH: 90-10 to afford the expected
product (220 mg, yield: 89 %)
1H NMR (360 MHz, CDCl3) δ 7.79 (d, J = 7.8 Hz, 2H), 7.41 – 7.19 (m, 5H), 7.14 (d, J = 7.8 Hz,
2H), 5.10 (m, 2H), 4.81 – 4.60 (m, 1H), 3.36 – 3.00 (m, 2H), 1.34 (s, 12H).13C NMR (91 MHz,
CDCl3) δ 155.80, 139.40, 135.81, 135.42, 132.23, 129.81, 128.58, 127.19, 127.07, 83.88, 66.91,
54.65, 37.81, 29.70, 24.84. HRMS (ESI): m/z Calcd for C23H29NO6, [M+H]+: 426.2086, found:
426.2068

JOURNAL PRE-PROOF
Compound 11
To a solution of 4-nitrophenyl chloroformate (745 mg, 3.6 mmol), Et3N (505 µL, 3.6 mmol) in 4
mL of DCM at 0°C was added dropwise a solution of 4-(hydroxymethyl)benzoic acid 10 (186
mg, 1.2 mmol) in 4 mL of DCM and 2 mL of DMF. The reaction was stirred at r. t. for 7h30. The
reaction was quenched with NaHCO3, the aqueous layer was extracted with DCM (3 x 20 mL),
the organic layers were combined, dried with MgSO4, and concentrated under reduced
pressure. The crude product was purified over silica gel with cyclohexane-EtOAc: 100-0 to 60-
40 to give the expected product (217 mg, yield: 74 %)
1H NMR (300 MHz, CDCl3) δ 8.33-8.26 (m, 4H), 7.46-7.36 (m, 4H), 5.42 (s, 2H). 13C NMR (75
MHz, CDCl3) δ 163.69, 155.52, 152.36, 145.52, 140.56, 130.80, 129.00, 128.30, 125.33,
122.58, 69.76. HRMS (ESI): m/z Calcd for C15H10NO7, [M-H]+: 316.0457, found: 316.0456
Compound 12
To a solution of doxorubicin hydrochloride (20 mg, 0.034 mmol) in 5 mL of DMF was added Et3N
(5 µL, 0.034 mmol). The mixture was stirred for 30 min at r.t. before adding HOBt (5 mg, 0.034
mg) and a solution of 11 (11 mg, 0.034 mmol) in 2 mL of DMF and the resulting mixture was
stirred a r.t. overnight. The reaction was quenched with brine, extracted with DCM (3 x 20 mL),
the organic layers were combined, dried with MgSO4 and concentrated under reduced pressure.
The crude product was purified over silica gel with DCM-MeOH: 90-10 to afford the expected
product (19 mg, yield: 73 %).
1H NMR (300 MHz, (CD3)2SO) δ 13.97 (d, J = 7.4Hz, 1H), 13.20 (d, J = 5.0Hz, 1H), 7.93 (d, J =
7.8Hz, 1H), 7.83 (m, 2H), 7.75 (d, J = 7.8Hz, 1H), 7.58 (d, J = 7.4 Hz, 1H), 7.32 (d, J = 7.4 Hz,
1H), 6.97 (s, 1H), 5.44 (d, J = 13.3 Hz, 1H), 5.22 (d, J = 13.3 Hz, 1H), 4.97 (s, 1H), 4.94-4.80
(m, 2H), 4.73 (d, J = 7.4 Hz, 1H), 4.65-4.50 (m, 2H), 4.30-4.10 (m, 2H), 3.94 (s, 3H), 4.97 (s,
1H) 13C NMR (75 MHz, (CD3)2SO) δ 213.90, 188.36, 186.45, 186.36, 160.74, 156.13, 155.20,
154.52, 154.06, 140.23, 136.20, 135.51, 134.58, 134.05, 133.92, 127.44, 126.99, 119.90,
119.69, 118.97, 110.72, 110.59, 101.67, 100.53, 100.36, 74.96, 70.01, 69.89, 67.94, 66.67,
63.75, 56.56, 17.06. HRMS (ESI): m/z Calcd for C36H35NO15Na, [M+Na]+: 744.1904, found:
744.1901.
H2O2 activation of the profluorescent probe
Carbon-boron bond oxidation by ROS (i.e. H2O2) was studied using the arylboronate
profluorescent probe 3 in presence of increasing H2O2 concentration at r.t. The study was
fulfilled in 96 well black plates (Corning) by mixing 3 in DMSO / 10 mM Hepes buffer (pH 7.5)
(1/99) at a final concentration of 50 µM in presence of increasing amount of H2O2: 1, 5, 10 and
50 equivalents (i.e. 50, 250, 500 and 2 500 µM, final concentration). The coumarin release
kinetic was studied over 12 h by reading its fluorescence intensity at λexcitation / λemission: 390 / 500
nm every 5 minutes for the first hour and every 30 minutes during the eleven remaining hours
(Infinite M200 Pro, Tecan trading AG, Switzerland). The final fluorescent intensity was
calculated by subtracting the minor fluorescence of 3 in absence of H2O2 to the fluorescence
obtained in presence of H2O2. The curves relating the release kinetic of coumarin were obtained
using GraphPad Prism software (GraphPad Software Inc, San Diego, CA, USA).

JOURNAL PRE-PROOF
Cell line and culture
Several cancer cell lines have been used for the in vitro evaluation of the designed prodrugs in
comparison to the parent drug: doxorubicin (DOX). Six cell lines were selected for the
cytotoxicity assay as follows: U87 (glioblastoma), A549 (lung carcinoma), MCF-7 (breast
cancer), MCF-7 MDR (multi drug resistant breast cancer), MiaPaCa-2 (pancreatic cancer) and
Hep G2 (liver cancer). All cells were grown in DMEM (Dulbecco’s Modified Eagle Medium)
supplemented with 10 % of fetal calf serum (FCS) and 100 U/mL penicillin and 100 µg/mL
streptomycin (Invitrogen) in a 5 % CO2 and 95 % hygrometry environment at 37 °C.
In vitro Screening
All seven cell line were investigated for their ability to produce ROS using profluorescent probe
3. Six well plates were used for this study, each well was seeded with 5.105 cell per well and
treated with 10 µM of 3 (sixplicate). After a 6h incubation period at 37 °C in presence 5 % CO2
and 95 % hygrometry, each well was sonicated three times during 5s (Sonicator FB505, Fisher
Scientific, USA) and 100 µL of the supernatant was transferred in a 96 well black plate. The
fluorescence of each well was then read at the coumarin fluorescence wavelength λexcitation
/
λemission: 390 / 500 nm to measure its release reflecting the ability of each cell line to produce
ROS. This study was carried out in triplicate and the results of each cell line was compared to
the cell line producing the least ROS (U87) results using a non-parametric Kruskall-Wallis test
(p < 0.001) with GraphPad Prism software (GraphPad Software Inc, San Diego, CA, USA). A p
value of < 0.0001 was accepted as statistically significant.
In vitro cell-proliferation assay
The cytotoxic activity of all the designed prodrugs and controls (8a, 8b, 8c, 9 and 12) was
investigated using the MTS [(3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-
sulfophenyl)-2H-tetrazolium)] method (Promega). The cells (2 500, 7 000, 7 000, 5 000, 5 000,
7 000 and 10 00 cell per well for A549, MCF-7, MCF-7 MDR, MiaPaCa-2, U87 and Hep G2,
respectively) (100 µL) were seeded in 96-well plates and incubated overnight at 37 °C in
presence 5 % CO2 and 95 % hygrometry. Then, the cells were treated with 100 µL of each
compound at different concentrations (i.e. 0.0025, 0.005, 0.025, 0.05, 0.25, 0.5, 2.5, 5, 25, 50
µM, final concentration) for all the tested cell lines. After 72 h, 20 µL (1/10) of MTS reagent were
added in each well. Depending on the cell line, 2 to 5 h incubations at 37 °C in presence 5 %
CO2 and 95 % hygrometry were needed to obtain the optimum optical density which was
measured at 490 nm wavelength using a microplate reader (Infinite M200 Pro, Tecan trading
AG, Switzerland). Untreated cells were used as control. Each concentration was tested in six
replicates and the experiment was fulfilled in triplicates. The concentration inhibiting 50 % of the
cell proliferation (IC50) was determined using GraphPad Prism software (GraphPad Software
Inc, San Diego, CA, USA).
Chick chorioallantoic membrane assay (CAM assay)

JOURNAL PRE-PROOF
Fertilized white leghorn chicken eggs (E.A.R.L. Morizeau, Dangers, France) were maintained at
24°C before hatching. Eggs were subsequently hatched in a humidified incubator (Sailnovo 56A
incubator) at 37 °C and 70 % hygrometry. The initial day of incubation is considered day 0. At
day 10, all eggs were mired and the air sac was visualized and delimited. The shell was
cautiously cut near the air sac position to visualize the egg membrane. The egg membrane was
homogeneously moisturized with NaCl 0.9 % solution and left 5 min. The solution was then
aspirated and the egg membrane was carefully removed revealing the chorioallantoic
membrane (CAM). Prior to implantation on the CAM, MiaPaCa-2 cell suspensions were
prepared by detaching them with trypsin / EDTA followed by their counting. Cells were
centrifuged at 1200 rpm for 5 min, washed twice using DMEM without FCS and suspended in a
mix of Matrigel (ECM Matrigel, Sigma) and DMEM (75/25) at a final concentration of 5.107 cell
per mL. 1.106 MiaPaCa-2 cells (20 µL) were deposited onto the CAM, without allowing the
pipette tip to touch the CAM. The window was then sealed using adhesive plaster to prevent
dehydration. The eggs were incubated (without shaking) awaiting the tumor to grow. After
approximately three days, tumor were visible and the treatment could take place. The tumors
were treated by intratumoral injections of DOX and 8a at different doses either as a single
injection (184 nmol) or twice a day for two consecutive days (20 nmol/injection). In ovo tumors
were photographed three days post-treatment using an Olympus digital camera 15.9 mega pixel
using the macro feature (OM-D E-M5). Image treatment and tumor measurements were carried
out using FIJI software.[42] The tumor were then taken out carefully and suspended in 1 mL of
PBS before being sonicated to dissociate the tumor. After centrifugation at 1200 rpm for 5 min,
the supernatant was filtered before being analyzed by HPLC.
HPLC Detection of doxorubicin
HPLC analysis was used to estimate the doxorubicin release within the tumor collected for the
HET-CAM assay using a gradient method composed of acetonitrile as the phase A and water
containing 0.1 % of formic acid as phase B. The HPLC separation was carried on a 1260 infinity
HPLC system (Agilent Technologies, United Kingdom). The chromatographic separation was
carried out on a Kinetex® 5µm C18 (250 x 4.6 mm) LC column at a flow rate of 1.0 mL/min. A
gradient elution was used with initially 10 % acetonitrile which was increased linearly to 30 %
over 4 min, then increased to 80 % over 2 min and remained at this condition for 6 min before
being finally decreased to 10 % in 3 min and was held until the end of the 20 min run. All analyses
were performed at 30 °C. The mobile phase were filtered and degassed in an ultrasonic bath
before use. The retention times of DOX and 8a in these conditions of analysis were 9.03 and
9.85 min, respectively.
Acknowledgements
This work was funded by the “Comité Essonne de la Ligue Nationale Contre le Cancer” and
the “Fondation ARC pour la recherche sur le cancer”. C. Skarbek is financed by a postdoctoral
training grant from the “Fondation ARC pour la recherche sur le cancer”. S. Serra was a “Ligue
Nationale Contre le Cancer” postdoctoral fellow. H. Maslah is supported by the “Ministère de
l’Education Nationale de la Recherche et de la Technologie”.Pierre Milcendeau is
acknowledged for technical assistance on organic synthesis.

JOURNAL PRE-PROOF
References:
[1] Cancer Research UK, https://www.cancerresearchuk.org/health-professional/cancer-
statistics/worldwide-cancer#heading-One, Accessed in May 2019.
[2]
J. Rautio, N. A. Meanwell, L. Di, M. J. Hageman. Nat. Rev. Drug Discov. 2018, 17, 559-
587.
[3]
[4]
[5]
[6]
V. Taresco, C. Alexander, N. Singh, A. K. Pearce. Adv. Therap. 2018, 1, 1800030.
D. Trachootham, J. Alexandre, P. Huang. Nat. Rev. Drug Discov. 2009, 8, 579-591.
G. Y. Liou, P. Storz. Free Radical Res. 2010, 44, 479-496.
N. Nishida, H. Yano, T. Nishida, T. Kamura, M. Kojiro. Vasc. Health Risk Manag. 2006,
2, 213-219.
[7]
X. Peng, V. Gandhi. Ther. Deliv. 2012, 3, 823-833.
[8]
J. P. Cadahía, V. Previtali, N. S. Troelsen, M. H. Clausen. Med. Chem. Commun. 2019,
DOI: 10.1039/C9MD00169G
[9] S. J. Baker, C. Z. Ding, T. Akama, Y. K. Zhang, V. Hernandez, Y. Xia. Future Med. Chem.
2009, 1, 1275-1288.
[10] A. Alouane, R. Labruère, T. Le Saux, F. Schmidt, L. Jullien. Angew. Chem. Int. Ed. 2015,
54, 7492-7509.
[11] G. C. van de Bittner, E. A. Dubikovskaya, C. R. Bertozzi, C. J. Chang. Proc. Natl. Acad.
Sci. USA 2010, 107, 21316-21321.
[12] D. Andina, J.-C. Leroux, P. Luciani, Chem. Eur. J. 2017, 23, 13549-13573.
[13] H. Hagen, P. Marzenell, E. Jentzsch, F. Wenz, M.R. Veldwijk, A. Mokhir. J. Med. Chem.
2012, 55, 924-934.
[14] S. Daum, M. S. V. Reshetnikov, M. Sisa, T. Dumych, M. D. Lootsik, R. Bilyy, E. Bila, C.
Janko, C. Alexiou, M. Herrmann, L. Sellner, A. Mokhir. Angew. Chem. Int. Ed. Engl. 2017, 56,
15545-15549.
[15] J. Noh, B. Kwon, E. Han, M. Park, W. Yang, W. Cho, W. Yoo, G. Khang, D. Lee. Nat.
Commun. 2015, 6, 6907.
[16] S. Cao, Y. Wang, X. Peng. Chem. Eur. J. 2012, 18, 3850-3854.
[17] Y. Wang, H. Fan, K. Balakrishnan, Z. Lin, S. Cao, W. Chen, Y. Fan, Q. A. Guthrie, H.
Sun, K.A. Teske, V. Gandhi, L. A. Arnold, X. Peng. Eur. J. Med. Chem. 2017, 133, 197-207.
[18] J. L. Major Jourden, S. M. Cohen. Angew. Chem. Int. Ed. Engl. 2010, 49, 6795-6797.
[19] Y. Kuang, K. Balakrishnan, V. Gandhi, X. Peng. J. Am. Chem. Soc. 2011, 133, 19278-
19281.

JOURNAL PRE-PROOF
[20] W. Chen, H. Fan, K. Balakrishnan, Y. Wang, H. Sun, Y. Fan, V. Gandhi, L.A. Arnold, X.
Peng. J. Med. Chem. 2018, 61, 9132-9145.
[21] E. J. Kim, S. Bhuniya, H. Lee, H. M. Kim, C. Cheong, S. Maiti, K. S. Hong, J. S. Kim. J.
Am. Chem. Soc. 2014, 136, 13888-13894.
[22] L. Wang, S. Xie, L. Ma, Y. Chen, W. Lu. Eur. J. Med. Chem. 2016, 116, 84-89.
[23] Q. Jiang, Q. Zhong, Q. Zhang, S. Zheng, G. Wang. ACS Med. Chem. Lett. 2012, 3, 392-
396.
[24] C. Zhang, Q. Zhong, Q. Zhang, S. Zheng, L. Miele, G. Wang. Breast Cancer Res. Treat.
2015, 152, 283-291.
[25] J. Peiró Cadahía, J. Bondebjerg, C.A. Hansen, V. Previtali, A. E. Hansen, T.L. Andresen,
M.H. Clausen. J. Med. Chem. 2018, 61, 3503-3515.
[26] V. Reshetnikov, S. Daum, A. Mokhir. Chem. Eur. J. 2017, 23, 5678-5681.
[27] S. Zheng, S. Guo, Q. Zhong, C. Zhang, J. Liu, L. Yang, Q. Zhang, G. Wang. ACS Med.
Chem. Lett. 2018, 9, 149-154.
[28] Y. Liao, L. Xu, S. Ou, H. Edwards, D. Luedtke, Y. Ge, Z. Qin. ACS Med. Chem. Lett.
2018, 9, 635-640.
[29] Y. Ai, O. N. Obianom, M. Kuser, Y. Li, Y. Shu, F. Xue. ACS Med. Chem. Lett. 2018, 10,
127-131.
[30] M. Ye, Y. Han, J. Tang, Y. Piao, X. Liu, Z. Zhou, J. Gao, J. Rao, Y. Shen, Adv. Mater.
2017, 29, 1702342.
[31] Y. Octavia, C. G. Tocchetti, K. L. Gabrielson, S. Janssens, H. J. Crijns, A. L. Moens. J.
Mol. Cell Cardiol. 2012, 52, 1213-1225.
[32] S. Cao, R. Christiansen, X. Peng. Chem. Eur. J. 2013, 19, 9050-9058.
[33] M. P. Hay, R. F. Anderson, D. M. Ferry, W. R. Wilson, W. A. Denny, J. Med. Chem. 2003,
46, 5533-5545.
[34] C. S. Kue, K. Y. Tan, M. L. Lam, H. B. Lee. Exp. Anim. 2015, 64, 129-138.
[35] N. A. Lokman, A. S. Elder, C. Ricciardelli, M. K. Oehler. Int. J. Mol. Sci. 2012, 13, 9959-
9970.
[36] M. Taizi, V. R. Deutsch, A. Leitner, A. Ohana, S. Goldstein. Exp. Hematol. 2006, 34,
1698-1708.
[37] M. Hagedorn, S. Javerzat, D. Gilges, A. Meyre, B. De Laforge, A. Eichmann, A. Bikfalvi.
Proc. Natl. Acad. Sci. U S A. 2005, 102, 1643-1648.
[38] M. Balke, A. Neumann, C. Kersting, K. Agelopoulos, C. Gebert, G. Gosheger, H. Buerger,
M. Hagedorn. BMC research notes, 2010, 3, 58.
[39] G. B. Atallah, W. Chen, D. Khrakovsky, L. Wang, Z. J. Jia, F. Jr Deanda. U.S. Pat. Appl.
Publ. (2018), US 20180194762 A1 20180712.

JOURNAL PRE-PROOF
[40] J. Zielonka, A. Sikora, M. Hardy, J. Joseph, B. P. Dranka, B. Kalyanaraman. Chem. Res.
Toxicol. 2012, 25, 1793-1799.
[41] A. Alouane, R. Labruère, T. Le Saux, I. Aujard, S. Dubruille, F. Schmidt, L. Jullien. Chem.
Eur. J. 2013, 19, 11717-11724.
[42] C. M. Doskey, V. Buranasudja, B. A. Wagner, J. G. Wilkes, J. Du, J. J. Cullen, G. R.
Buettner. Redox Biol. 2016, 10, 274-284.
[43] J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, Pietzsch, S. Preibisch,
C. Rueden, S. Saalfeeld, B. Schmid, J.-Y. Tinevez, D. J. White, V. Hartenstein, K. Eliceiri, P.
Tomancak, A. Cardona. Nature methods, 2012, 9, 676-682.

JOURNAL PRE-PROOF

JOURNAL PRE-PROOF
Arylboronate-doxorubicin conjugates as anticancer ROS-responsive prodrugs were synthesized.
In vitro cytotoxicity of the prodrugs and controls was determined on a panel of cancer cells.
MiaPaCa-2 pancreatic cancer cells were selected for their ability to efficiently produce ROS species.
Prodrug 8a (80 nmol) was able to induce 50% regression of MiaPaCa-2 tumor in ovo.