Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy for anticancer therapy

Article abstract of DOI:10.1016/j.ejpb.2013.03.006

The aim of this study was to develop clickable prodrugs bearing a tunable pH responsive linker designed for acidic pH-mediated release of histone deacetylase inhibitors. HDACi are an important class of molecules belonging to the epigenetic modulators used for innovative cancer strategies. The behavior of these prodrugs was determined by a bioluminescence resonance energy transfer assay in living tumor cells. This work demonstrated that this innovative type of clickable prodrugs entered cancer cells and showed restored anti proliferative properties attributed to the effective release of the HDAC inhibitors. A correlation between kinetic studies, dose responses, and biological activities was obtained, making such clickable prodrugs good candidates for new strategies in epigenetic-oriented anticancer therapies.

Full text of DOI:10.1016/j.ejpb.2013.03.006

European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
Contents lists available at SciVerse ScienceDirect
European Journal of Pharmaceutics and Biopharmaceutics
journal homepage: www.elsevier.com/locate/ejpb
Research paper
Design of pH responsive clickable prodrugs applied to histone
deacetylase inhibitors: A new strategy for anticancer therapy
Régis Delatouche a, Iza Denis b,c,d, Marion Grinda a, Fatima El Bahhaj a, Estelle Baucher b,c,d
,
a,
Floraine Collette e, Valérie Héroguez e, Marc Grégoire b,c,d, Christophe Blanquart b,c,d, , Philippe Bertrand
⇑
⇑
a Institut de Chimie des Milieux et Matériaux de Poitiers, CNRS, UMR 7285, Université de Poitiers, Poitiers cedex, France
b Inserm U892, Nantes, France
c Université Nantes, Nantes, France
d CNRS 6299, Nantes, France
e Laboratoire de Chimie des Polymères Organiques, CNRS, UMR 5629, Université de Bordeaux, Pessac, France
a r t i c l e i n f o
a b s t r a c t
Article history:
The aim of this study was to develop clickable prodrugs bearing a tunable pH responsive linker designed
for acidic pH-mediated release of histone deacetylase inhibitors. HDACi are an important class of mole-
cules belonging to the epigenetic modulators used for innovative cancer strategies. The behavior of these
prodrugs was determined by a bioluminescence resonance energy transfer assay in living tumor cells.
This work demonstrated that this innovative type of clickable prodrugs entered cancer cells and showed
restored anti proliferative properties attributed to the effective release of the HDAC inhibitors. A correla-
tion between kinetic studies, dose responses, and biological activities was obtained, making such click-
able prodrugs good candidates for new strategies in epigenetic-oriented anticancer therapies.
Ó 2013 Elsevier B.V. All rights reserved.
Received 21 January 2013
Accepted in revised form 1 March 2013
Available online xxxx
Keywords:
Click chemistry
Prodrug
pH response
Histone acetylation
Cancer
Bioluminescence
1. Introduction
of gene transcription and defining the so-called histone code [4].
Epigenetics plays nowadays an important role in cancer researches
Epigenetics has emerged in recent years as an important mech-
anism for the regulation of gene expression through post-transla-
tional modifications (PTMs) on the chromatin [1] with
applications in various human diseases (cancer, neurodegenerative
and metabolic disorders, hepatitis, and others). In eukaryotic cells,
the minimal unit of the chromatin is the nucleosome, built from
the assembly of eight histones (two copies of H2A, H2B, H3, and
H4) and DNA [2]. Each nucleosome is separated by a fifth histone:
H1. This DNA packaging is responsible for the regulation of gene
transcription. PTMs are reversible chemical reactions occurring
on DNA and on histones. While the major reaction on DNA is the
methylation of CpG islands [3], several modifications are known
for histones, mainly on the lysine residues found in the terminal
tails of histones H3 and H4. The combination of these PTMs can
be exclusive or complementary and is recognized by transcription
factors in a selective way, leading to the repression or stimulation
[5]. The most studied histone PTM is lysine acetylation, with his-
tone acetyl transferases (HAT) [6] responsible for the N- acetyla-
tion of lysine while histone deacetylases (HDAC) [7] remove this
acetyl group (Fig. 1). This acetylation equilibrium is part of the
mechanisms controlling gene expression and is recognized by
bromodomain-containing transcription factors [8].
e
HDACs appeared over expressed in several cancer cell lines
resulting in abnormal hypoacetylation of histones, favoring in par-
ticular tumor suppressor gene silencing. Thus, inhibition of HDACs
appeared to be an innovative anticancer strategy and a large num-
ber of studies demonstrated the potential of HDACi on different
cancer cell lines and on animal models of tumors [9]. HDAC inhib-
itors (HDACi) entered clinical trials [10] and two drugs were ap-
proved to treat cutaneous T-cell lymphoma (CTCL): Vorinostat
(ZolinzaÒ) = Suberoyl Anilide Hydroxamic Acid (SAHA) and Romi-
depsin (IstodaxÒ) (Fig. 1). SAHA is a non-specific HDACi while
Romidepsin is considered to be a natural disulfide prodrug cleaved
in cells by glutathione [11]. Entinostat (MS-275) and CI-994 illus-
trate the benzamide group (Fig. 1). Valproic acid (VPA), a carbox-
ylic acid approved to treat epilepsy, is a millimolar HDACi, often
used due to its well-known clinical profile, whereas HDACi have
showed high efficacy on hematological diseases such as myelodys-
plastic syndromes (MDS) and CTCL. Their effects on solid tumors
⇑
Corresponding authors. Inserm U892/CNRS 6299, Nantes, France (C. Blanquart),
Institut de Chimie des Milieux et Matériaux de Poitiers, CNRS, UMR 7285, Université
de Poitiers, 4 rue Jacques Fort, B27, F-86022 Poitiers cedex, France. Tel.: +33
(0)549454192; fax: +33 (0)549453501 (P. Bertrand).
E-mail addresses: christophe.blanquart@inserm.fr (C. Blanquart), philippe.
bertrand@univ-poitiers.fr (P. Bertrand).
0939-6411/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ejpb.2013.03.006
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

2
R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
Thus, pH responsive carriers can be used to transport and release
compounds during endocytosis with typical pH responsive groups
like trityl [19], benzaldehyde [20], and hydrazone [21] depending
on the linked molecules. For Doxorubicin, a hydrazone derivative
on the side chain carbonyl group and a carbamate group on the ami-
no sugar were described as prodrugs for acidic pH-mediated release.
The hydrazone strategy gave better results [22] showing that simple
carbamate are not useful in this case for acidic release. Trytil and
benzaldehyde groups were used for nucleosides derivatives. In con-
trast, we demonstrated in a recent contribution to this field that the
azide–alkyne click chemistry approach leading to triazole rings can
be used as a general ligation strategy producing at the same time pH
responsive prodrugs. Carbamate derivatives and analogs of the trityl
group were prepared by the same synthetic strategy all retaining
acidic pH sensitivity [23]. These results indicated that our generic,
flexible, and readily synthesized tunable pH responsive groups are
valuable alternative to prepare prodrugs of the various functional
groups found in bioactive molecules. Another point of interest is
the protonation of triazole rings, exploited for the preparation of
membranes for proton transfer [24a]. This property could be useful
for vesicle escape during endocytosis. It has been demonstrated that
polyimidazoles used as vector for DNA (so-called polyplexes [24b])
can escape the endocytosis pathway at the endosome level to avoid
later DNA degradation in the more acidic lysosomes. This escape is
the result of modification in proton gradient in the endosomes due
to imidazole protonation. Protonation of the triazole could lead to
the same effect.
We aimed applying these results for the preparation of pH
responsive diaryltriazolyl-based prodrugs of the important thera-
peutic class of HDACi. Our preliminary results on anilines
prompted us to select CI-994, a HDACi member of the benzamide
group previously used in combination trials and related to Entino-
stat, the latter currently in phase I/II clinical trials. SAHA was also
chosen for its clinical importance and rapid clearance. SAHA has
been involved in 2012 in several phase I/II clinical trials in combi-
nation with other anticancer compounds and to fight various can-
cer types. The possibility to obtained controlled release and
possibly for different molecules from the same pH responsive
structure could thus also have high interest in differentiated re-
lease. We envisioned the covalent linking of these two molecules
in a protecting prodrug way through their respective hydroxamic
acid and benzamide functions, both being prone to metabolic
activities. Having also in mind to develop a prodrug system that
could be further functionalized, we modified our initial pH respon-
sive system, introducing a second alkyne–azide click chemistry
site. Our double click synthetic strategy involves the key mono-
propargylated ether 3 (Scheme 1), where a triazol ring is first
formed to produce the pH responsive part while an alkyne ether
part will serve for future modifications by a second click reaction.
In particular, this second click site could be used for further graft-
ing onto drug delivery system (DDS), provided this DDS bears an
Fig. 1. Top: Principle of reversible acetylation of histones by HAT/HDAC and its
recognition by bromodomain-containing proteins, with therapeutically positioned
HDAC inhibitors. Bottom: examples of HDAC inhibitors. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of
this article.)
were limited due to their hematological toxicity, mainly thrombo-
penia, and cardiotoxicity which limit the doses administrated [10].
Another main limitation is their rapid clearance. SAHA is metabo-
lized in less than 1 h by glucuronidation [12] and by beta-oxida-
tion, like VPA [13], and some resistance are known for
hydroxamic and carboxylic acids [14]. Romidepsin is metabolized
as glutathione conjugate [15] or by the action of P450 enzymes
[16]. Resistance has also been demonstrated for Romidepsin [17].
Benzamides are more stable with 1 or 2 days half-life.
All these constrains are responsible for the low level of
compounds reaching the malignant tissues and make HDACi good
candidates for ligation to prodrugs able to release them slowly over
time only in cancer cells in order to reduce side effects and metabo-
lism. An important strategy developed for efficient cancer cell
internalization of bioactive molecules relies on the endocytosis
pathway [18]. To our knowledge, this delivering strategy has never
been applied to HDACi and appeared of high interest to circumvent
the clinical limitations of this class of compounds. Drug delivery
based on endocytosis-mediated internalization exploits the initial
formation of endosomes with moderately acidic pH (pH = 6) leading
after maturation to lysosomes with a more acidic pH (pH = 4–5).
Scheme 1. Design of acidic pH responsive releasing systems for differentiated ‘‘double click’’ strategy.
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
3
azide group or to add detectable properties for in vivo imaging
[25]. The conversion of the alcohols 8 to ethers 3 is expected to in-
crease the hydrophobic character of the molecules, but on the
other hand, this strategy also avoids the protection/deprotection
of the terminal primary hydroxy group prior to drug insertion that
could make the approach more complex. Blocking this primary hy-
droxyl group for drug insertion is necessary as we demonstrated
previously that primary alcohols can be used as drugs with our sys-
tem [23c]. Thus, preparing prodrugs from alcohols 8a–c could give
intermolecular attack of these alcohols to yield undesired polymers
instead of the expected alkyne-free prodrugs. A spacer R2, selected
to be biologically inert and to potentially improve the water solu-
bility, was inserted between the two click sites to account for the
bulkiness of the pH responsive part. Spacers are usually necessary
to make the reactive site of prodrugs easily accessible especially
when enzymes are involved in the releasing mechanism.
We report in this work our synthetic findings to obtain a differ-
entiated double azide–alkyne click chemistry on the same struc-
ture, its application to the preparation of new clickable pH
responsive HDACi prodrugs and their biological evaluations
(Scheme 1, drug = HDAC inhibitor (HDACi)). Three versions were
prepared (various R1 groups) for each HDACi and were evaluated
in cancer cells for their ability to release the inhibitors with resto-
ration of HDAC inhibition measured by a BRET assay. The final im-
pact on cancer cell viability was also measured.
(0.052 g, 0.073 mmol, 17%). The unreacted alcohol 3a is also recov-
ered. Method B: To a solution of 3a (0.11 g, 0.25 mmol, 1 eq.) in dry
toluene under nitrogen atmosphere was added AcCl (0.09 mL,
1.25 mmol, 5 eq.). After 2 h reflux, the solution was evaporated to
give the crude chloride 10a as oil. The addition of SAHA, work-
up, and purification was performed as for method A. Compound
4a was obtained in better yields with method
B (0.075 g,
0.107 mmol, 42%). The unreacted alcohol 3a (0.010 g, 10%) and
SAHA (0.124 g, 36%)) were also recovered.
TLC MeOH/CH2Cl2/Et3N 5:94:1 Rf = 0.33; 1H NMR (400 MHz,
Acetone-D6): d = 9.75 (s, 1H), 9.08 (s, 1H), 7.69–7.65 (m, 3H),
7.46–7.41 (m, 4H), 7.34–7.25 (m, 8H), 7.04–7.00 (m, 1H), 4.57 (t,
2H, J = 5.1 Hz), 4.16 (d, 2H, J = 2.4 Hz), 3.88 (t, 2H, J = 5.2 Hz),
3.62–3.55 (m, 6H), 3.53–3.49 (m, 6H), 2.93 (t, 2H, J = 2.4 Hz), 2.31
(t, 2H, J = 7.5 Hz), 1.93 (t, 2H, J = 7.2 Hz), 1.60 (m, 2H), 1.39 (m,
2H), 1.25 (m, 2H), 1.12 (m, 2H); 13C NMR (100 MHz, Acetone-
D6): d = 26.0, 33.7, 37.7., 50.8, 58.6, 69.8, 70.1, 70.9, 71.1, 71.16,
71.2, 71.22, 75.8, 80.9, 115.7, 119.9, 123.8, 126.7, 128.5, 128.6,
129.2, 129.4, 133.6, 140.8, 161,0, 172,0; HRESI-MS: calcd. for
[M+Na]+ (C40H49N5O7Na): 734.35242, found: 734.3524.
2.2.2. Typical synthesis of compound 5
N-(2-((1-(3,6,9,12-tetraoxapentadec-14-ynyl)-1H-1,2,3-triazol-
4-yl) diphenylmethyl amino) phenyl)-4-acetamidobenzamide 5a.
Chloride 10a was prepared as above and dissolved in minimum
ACN (2–3 mL), and the resulting solution was added to a solution
of CI-994 (0.231 g, 0.90 mmol) in ACN (5 mL) containing Et3N
(0.24 mL, 1.72 mmol). The solution was stirred 12 h at ambient
temperature and concentrated, and the crude material was dis-
solved in acetone from which the unreacted CI-994 was precipi-
tated and filtered off for recycling. The filtrate was purified
(Combi Flash DCM/MeOH/Et3N 97:2:1) to give 5a as an oil
(0.142 mg, 0.17 mmol, 39%). The unreacted alcohol 3a was also
recovered. 1H NMR (400 MHz, Acetone-D6): d = 9.44 (bs, 2H),
8.02 (d, 2H, J = 8.6 Hz), 7.82 (s, 1H), 7.77 (d, 2H, J = 8.7 Hz), 7.65
(m, 4H), 7.24 (m, 5H), 7.18 (m, 2H), 6.70 (dt, 1H, J = 1.6 Hz,
J = 8.1 Hz), 6.61 (dt, 1H, J = 1.3 Hz, J = 7.5 Hz), 6.17 (dd, 1H,
J = 1.1 Hz, J = 8.2 Hz), 6.12 (s, 1H), 4.50 (t, 2H, J = 5.2 Hz), 4.14 (d,
2H, J = 2.4 Hz), 3.79 (t, 2H, J = 4.9 Hz), 3.59 (m, 2H), 3.55 (m, 2H),
3.48 (m, 4H), 3.43 (m, 4H), 2.92 (t, 1H, J = 2.4 Hz), 2.12 (s, 3H);
13C NMR (100 MHz, Acetone-D6): d = 169.3, 152.9, 146.5, 143.5,
141.9, 129.9, 129.5, 128.9, 128.8, 127.7, 127.5, 126.8, 126.5,
125.9, 119.2, 118.6, 118.1, 81.0, 75.8, 71.21, 71.17, 71.0, 70.2,
69.8, 65.7, 58.6, 50.9, 24.4, 22.9, 15.2; HRESI-MS: calcd. for
[M+Na]+ (C41H44N6O6Na): 739.32145, found: 739.3215.
2. Materials and methods
2.1. Materials
All solvents were dried before use when required with classical
methods. Distilled water was used for cycloaddition under Sharp-
less conditions. 250 lm silica gel plates were used for TLC analysis.
All reactions were conducted under nitrogen, and concentrations
were performed under reduced pressure using a rotary evaporator.
Extracted organic layers were dried with MgSO4 and the solvents
removed under reduced pressure. Purifications were made by col-
umn chromatography with silica gel (35–70 lm silica) or using
Combi Flash apparatus for compound 4 and 5. DCM = CH2Cl2,
THF = tetrahydrofuran, EA = ethyl acetate, PE = petroleum ether
35–60°, ACN = acetonitrile. 1H (400 MHz) and 13C (100 MHz)
NMR spectra are given in ppm as referenced to TMS (tetramethyl-
silane) as internal standard. 1H NMR coupling constants are re-
ported in Hertz and refer to apparent multiplicities and not true
coupling constants. Data are reported as follows: chemical shift,
multiplicity (s = singlet, br s = broad singlet, d = doublet, t = triplet,
q = quartet, m = multiplet, dd = doublets of doublets, etc.), integra-
tion, and coupling constant. HRMS were performed in the Centre
Régional de Mesures Physiques de l’Ouest, Université de Rennes
1, Campus de Beaulieu, 35042 Rennes, France.
2.2.3. Determination of hydrolysis rates of prodrugs in mildly acidic pH
The compounds 4a–c and 5a–c were submitted to acidic hydro-
lyzes with buffered solutions. 0.4 mL of a 0.5 mg/mL solution of the
compounds in ACN was added to 1.6 mL of buffer under vigorous
stirring. At this concentration, no precipitation of the starting com-
pound was observed for any of the derivatives analyzed. HPLC of the
starting prodrugs and buffers alone in the mixture of ACN/H2O used
for hydrolysis conditions were used to validate the drug release and
reformation of alcohols 3. No unknown products were formed in
these conditions. The acidity is being probably not strong enough
for Ritter reaction with ACN [26]. The final solutions tested were
thus composed of 20:80 ratio of acetonitrile/buffer with pH values
in the range 4.3, 5.0, 6.0, and 7.3. The clear solution was then in-
jected at different times directly with an automated High Pressure
Liquid Chromatography (HPLC) apparatus equipped with a DAD
Hitachi L-2455 (detection range: 200–400 nm), an autosampler
Hitachi L-2200, a pump Hitachi L-2130, and a reversed phase HPLC
column Lichrocart 150-4,6 purospher STAR (Supplementary infor-
mation for eluting systems). The peaks corresponding to the start-
ing material and the released alcohols were integrated and
2.2. Methods
2.2.1. Typical synthesis of compound 4
N-phenyl-N0-[[1-[2-[2-[2-(2-prop-2-ynoxyethoxy) ethoxy] eth-
oxy] ethyl] triazol-4-yl]-bis(phenyl) methoxy] octanediamide 4a.
Method A: To a solution of 3a (0.20 g, 0.43 mmol) in DCM (5 mL)
was added HCl (2 M in Et2O, 0.3 mL, 0.86 mmol). After 2 h reflux,
the solution was co-evaporated with toluene to give the crude
chloride 10a as oil. To a solution of SAHA (0.340 g, 0.129 mmol)
and dry NEt3 (0.24 mL, 1.72 mmol) in ACN (5 mL) was added the
crude chloride 10a dissolved in a minimum of ACN at ambient
temperature. The resulting solution was stirred overnight at room
temperature, then filtered, and washed with ACN to recover unre-
acted SAHA as a solid. The concentrated filtrate was purified by
Combi Flash (DCM/EtOH/Et3N 94:5:1) to give 4a as a colorless oil
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

4
R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
compared to determine the hydrolysis kinetics. All peak attribu-
tions were confirmed by comparison of the UV spectra from 200
to 400 nm of pure material. A co-injection with the starting mate-
rial at the end of the hydrolysis was also made when necessary to
detect any deviation in retention times during the experiments
and to confirm the disappearance of the starting prodrugs.
Fluorescence was quantified by measuring emission at 595 nm
after excitation at 532 nm using a Typhoon apparatus.
3. Results and discussion
3.1. Synthesis
2.2.4. Drugs
A polyethyleneoxide (PEO) chain with 4 units was chosen as
For biological tests as control, SAHA was purchased from R&D
chemicals and CI-994 was obtained from Sigma–Aldrich. SAHA
and CI-994 used for prodrugs synthesis were prepared according
to literature procedures. All drugs and prodrugs tested were dis-
solved in DMSO at a stock concentration of 10 mM.
spacer R2 (Scheme 1) to prepare the known compound
1
(Scheme 2) [30]. Ketones 7a–c, purchased or prepared from chlo-
ride 6b for 7b, were converted to alkynols 2a–c and reacted with
the azide 1 by copper catalyzed Huisgen’ cycloaddition to give
alkynols 8a–c in high yields (Scheme 2). The propargylation of
alkynols 8a–c gave the expected inseparable mixture of monoether
3 and diether 9. This difficulty was circumvented by a treatment
with two equivalents of base and propargylbromide to give exclu-
sively the diether 9. Treatment with 1 M KHSO4 of the 3:9 mixture
or the diether 9 alone gave only the desired monoether 3a–c. The
expected sensitivity to acidic pH of the prodrugs was exploited
to selectively hydrolyze the ether 9 while the monoether 3 was sta-
ble under such conditions. This result also confirmed the applica-
tion of such system as pH responsive prodrugs for alcohols. The
syntheses of the key intermediates 3 were achieved in 4 steps in
high yields from diols 8a–c. Alkynols 3a–c were then converted
to the intermediate chlorides 10a–c and used directly for substitu-
tion by SAHA and CI-994 (Scheme 2). The final linking of the drugs
to afford the prodrugs 4 and 5 gave moderate yields due to limited
solubility of the drugs in our experimental conditions. Attempts to
improve the yields (added DMSO, various bases (various equiva-
lents of NEt3, collidine, lutidine) and chlorination methods (AcCl/
reflux or HCl/Et2O) gave limited results, but the starting material
can be recovered easily. Unreacted SAHA and CI-994 could be recy-
cled by precipitation from the reaction mixture with acetone or
acetonitrile in 36–73% and 37%, respectively, while the alkynols
3a–c were recovered during purification of the filtrate by chroma-
tography in 10%, 38%, and 40%, respectively. This recycling strategy
allowed preparing up to 300 mg of compound 5a and 240 mg of
compounds 4b.
2.2.5. Cell culture
The pleural mesothelial cell line, MeT-5A, was obtained from
American Type Culture Collection (ATCC). The mesothelioma,
Meso13 and Meso34, and adenocarcinoma (ADCA), ADCA 153
and ADCA 72, cell lines were established from the pleural fluids
of mesothelioma or lung ADCA patients, respectively [27]. All cell
lines were maintained in RPMI medium (Invitrogen) supplemented
with 2 mM L-glutamine, 100 IU/mL penicillin, 0.1 mg/mL Strepto-
mycine, and 10% heat inactivated fetal calf serum (FCS) (Eurobio).
2.2.6. Transfections
MeT-5A cells were seeded at a density of 2.5 ꢀ 105 cells per
35 mm dish. Transient transfections were performed 1 day later
using Attractene (Qiagen), according to the manufacturer’s proto-
col. For BRET experiments, MeT-5A cells were transfected with
0.6 lg Rluc-Brd cDNA and 1 lg YFP-fused histone H3 cDNA [28].
One day after transfection, cells were transferred into 96-well
microplates (microplate 96 well, white, Berthold Technologies) at
a density of 3 ꢀ 104 cells per dish in 180
l
L of culture medium.
L of compounds
Treatments were performed by addition of 20
l
solutions to obtain 200
concentration used for the 20
the expected final concentration in the culture medium to account
for the dilution effect. The addition was made at different times
depending on the experiment. BRET measurements were per-
formed as described below.
l
L of final culture medium. The compound
L addition was 10-fold higher than
l
SAHA and CI-994 were found linked to alcohols 3 as shown in
Scheme 2. No alternative ligations were detected. In the case of
CI-994, only the free NH2 group is reactive, when compared to
the two other amide groups that require more basic conditions
for activation. In the case of SAHA, the hydroxyl group is more
reactive than the nitrogen in both amide groups (benzamide or
hydroxyamide). All compounds 4 and 5 were >95% purity (HPLC
Supplementary information) with no free drug content. The stabil-
ity of prodrugs in neat form and in DMSO solution stored at ꢁ20 °C
was confirmed by several biological experiments over months.
2.2.7. BRET measurements
All BRET measurements were performed at room temperature
using the Mithras LB 940 microplate analyzer (Berthold Technolo-
gies). Cells were pre-incubated for 15 min in PBS in the presence of
2.5 lM coelenterazine, following which light-emission acquisition
at 485 and 530 nm was carried out. Plates were measured five
times. The BRET signal was expressed in milliBRET units (mBu).
The BRET unit has been defined previously as the signal ratio
530/485 nm obtained when the two partners are present, cor-
rected by the ratio 530/485 nm obtained under the same experi-
mental conditions, when only the partner fused to Renilla
luciferase is present in the assay [29].
3.2. Hydrolysis rates
Compounds 4 and 5 were then submitted to buffered solutions
at biologically relevant pH to monitor the HDACi release over time
[23]. As expected, the more electron donating group on the phenyl
rings (R1 groups) the faster the hydrolysis (Table 1). With the more
sensitive compounds having half-life less than minutes the appar-
ent half life was difficult to determine with accuracy due to the
experiment schedule. For this reason, half-life is indicated as being
below minutes. Compounds 5 were all more sensitive to acidic
conditions compared to compounds 4. This result was expected
from our previous studies on alcohols (to be compared to hydrox-
yamides) and anilines [23d], showing the higher stability of the
alcohols [23c]. As a comparison at pH 4.3, compound 4a was stable
while 5a was rapidly hydrolyzed and compound 4b was 1200-fold
more stable than 5b. These differences were reduced at pH 6 with
2.2.8. Determination of cell viability
Cell growth of Meso13, Meso34, ADCA153, and ADCA72 was
monitored using Uptiblue (Interchim). Reduction of this compound
by the cell results in the formation of a fluorescent compound
quantified by measuring fluorescence at 595 nm after excitation
at 532 nm using a Typhoon apparatus. Cells were seeded in 96-well
plates at a density of 5 ꢀ 103 cells/well in 180
lL of culture med-
ium. Twenty-four hours later, 20 lL compounds solutions were
added for 72 h. The compound concentration used for the 20 lL
addition was 10-fold higher than the expected final concentration
in the culture medium to account for the dilution effect. Uptiblue
(5%, v/v) was then added to the culture medium for 2 h at 37 °C.
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
5
Scheme 2. (i) R1APh, CH2Cl2, AlCl3. (ii) a – TMSAC„CAH, nBuLi, THF; b – KOH, MeOH. (iii) CuI, EtOH:CH2Cl2. (iv) THF, NaH 2 eq., BrACH2C„CH 2 eq. (v) KHSO4 1 M. (vi) HCl
in Et2O, or SOCl2. (vii) NEt3, ACN.
for instance prodrug 4b being only 2-fold more stable than 5b.
Both 4c and 5c were rapidly hydrolyzed, highlighting a possible
susceptibility of the c series to nucleophilic attack, a result previ-
ously observed with di- or trianisylmethyl chlorides, very sensitive
to water hydrolysis [31]. In tumor cells, the pH environment was
described as more acidic than in normal tissues while the intracel-
lular pHs for endosomes and lysosomes is around 6 and 5, respec-
tively [24c]. The hydrolysis rate measured with pH 6 buffer
indicated that endosomal hydrolysis (and drug escape) should be
very slow for compounds 4a, b, and 5a and that the process will
be faster at the lysosomal level.
methyl or methoxy substituted trityl hydroxamic acids described
to our knowledge in the literature for comparisons purposes. At
the opposite compound, 4c was rapidly hydrolyzed at all tested
pHs. The best system for the expected endocytosis-mediated re-
lease of SAHA appeared to be the intermediate version 4b, with a
half-life of 20 h at pH 4.3 and with a 4 days half-life at pH 7.3. De-
spite the apparent limitation at physiological pH, this result is an
improvement for SAHA delivery, the free molecule being metabo-
lized in less than 1 h. In the case of CI-994, once again version 5c
was hydrolyzed rapidly at all pHs and the best solution was the
derivative 5a with a 30 min half-life at pH 4.3 and with a 4 days
half-life at pH 7.3. For compound 5c, this result is very different
compared to previous studies with aniline prodrugs stable at pH
7.4. This could be due to the electron withdrawing effect of the
ortho-amide group, influencing the Ar2triazolyl CAN bond strength
or reactivity.
For SAHA (Table 1), derivative 4a (Scheme 2, R1@Ph) was stable
in the pH range selected in agreement with literature data
indicating that trityl protected hydroxamic acids were stable in ba-
sic conditions and required higher acidity for cleavage. There is no
The lysosomal maturation being less than 1 h and the lyso-
somes remaining active for hours, we hypothesized that these
two candidates 4b and 5a should correspond to our needs, as at
pHs 5 and 6, they have long intermediate half-lives, considering
the application for cell delivery through endocytosis and require-
ment for stability in blood plasma.
From a mechanistic point of view, these results may also be ex-
plained by Scheme 3. After protonation of compounds 4 or 5 on the
hetero atom linked to the tertiary carbon, suggested as the more
nucleophilic site, the active molecule is released to give the
Table 1
Half-lives for compounds 4a–c and 5a–c in buffered conditions.
pH 4.3
pH 5.0
pH 6
pH 7.3
4a
4b
4c
5a
5b
5c
Stable
20 h
30 min
30 min
<1 min
<1 min
Stable
–
80 min
400 min
12 min
<5 min
Stable
Stable
4 days
34 min
3.6 days
2 h 30
2 h 30
80 h (3.3 days)
108 min
48 h
10 min
<5 min
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

6
R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
Table 2
EC50 values for compound-induced BRET signal.
reaction are not favored for tertiary alkyl and that alternative
Sn1 mechanisms could be involved, with pre-association mecha-
nisms and intimate or solvent ion pairing formation [33]. It can
be suggested for CI-994 that the electron attractive effect of the
ortho carbonyl group gives a less basic NA atom, favoring a transi-
tion state where a water molecule is interacting with the N3 atom
prior to hydrogen transfer to the NA nitrogen atom. Alternatively,
we showed in a related work that aniline carbamate prodrugs
bearing a electron withdrawing group in ortho position react in
nucleophilic conditions to yield oxazolidinone by intramolecular
cyclization as well as aniline release [34]. Such mechanisms could
be involved in the present work for water dissociation assistance.
This suggested mechanism can also be applied to SAHA, with the
formation of an intermediate anion. This anion can be stable to
some extent, favoring solubility or diffusion, but eventually also
the re-attack of TAr2C+, if present.
Compounds
EC50 (lM)
SAHA
4a
4b
4c
CI-994
5a
0.42 0.08
7.17 3.19
0.59 0.18
0.75 0.07
0.86 0.08
6.20 1.37
0.68 0.12
8.88 3.89
5b
5c
EC50 values were determined using GraphPad prism,
Prism 5 for Windows, by curve fitting using a sigmoidal
dose–response model. Results are the means SEM. of
three independent experiments.
corresponding carbocation TAr2C+, stabilized by the two phenyl
groups. This carbocation is finally trapped in a Sn1 mechanism by
water to give alcohols 3 after proton release. If the released mole-
cules are sufficiently nucleophilic, a retro attack of the carbocation
will limit the apparent release kinetics. In acidic conditions, this
equilibrium between carbocation stability and the nucleophilic
character of the active molecule are orienting the reaction. Once
protonated, the CI-994 is released and the free aniline can be itself
protonated, limiting retro attack and accounting for apparent fas-
ter kinetics compared to SAHA. Alternatively, a possible direct
water substitution on the protonated prodrug intermediate can
be proposed leading to drug release and formation of protonated
alcohols 3. A final deprotonation will give the free alcohols. For ser-
ies c, a mechanism can also be proposed involving the participation
of carbonyl groups [32] or the triazole ring favoring local water dis-
sociation. In our previous work, we demonstrated that the N3
nitrogen atom can be protanated first if the bioactive molecule is
linked as a carbamate [23b,d]. It has been postulated that Sn2
3.3. Dose-dependent restoration of HDAC inhibition measured by BRET
assay
We previously developed a practical BRET assay to determine
the ability of molecules to induce histone H3 acetylations based
on HDAC inhibition [28]. This assay allows accelerated biological
evaluations of potential HDACi directly in living cells and avoids
complex protocols like Western blotting. Briefly, interactions be-
tween bromodomain and acetylated histones were exploited to
engineer phRluc-C1-BrD and pEYFP C1-histone H3 cDNA transfec-
ted in cells. phRluc-C1-BrD cDNA encodes a bromodomain fused to
Renilla Luciferase (RLuc) and pEYFP C1-histone H3 cDNA encodes
histone H3 fused to Yellow Fluorescent Protein (YFP). An increase
in BRET signal reflects an increase in histone H3 acetylation, indi-
cating that free HDACi enter the cells and block the nuclear HDACs
activities. In this work, several cancer cell lines were used for
comparing the compound activities. The immortalized pleural
Scheme 3. Postulated release mechanisms for compounds 4 and 5 in acidic and moderately basic conditions.
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
7
mesothelial cell line MeT-5A, was obtained from American Type
Culture Collection (ATCC), the malignant pleural mesothelioma,
Meso13 and Meso34, and adenocarcinoma cell lines (ADCA: ADCA
153 and ADCA 72) were established from the pleural fluids of
mesothelioma or lung ADCA patients, respectively [29]. The BRET
assay was used to determine dose responses and kinetics of the
releasing systems. Compounds 3, 4a–c, and 5a–c were evaluated
using our BRET assays and for antiproliferative activities on MPM
and ADCA cancer cell lines.
good correlation is observed between acidic sensitivity and BRET
induction.
CI-994 and its derivatives 5 gave also dose-dependent BRET
induction with a sigmoid profile. (Fig. 2, CI-994 top left, compound
5 bottom right). Again prodrugs 5 are supposed to be inactive until
CI-994 is released as CI-994 is attached by the orthobenzamide
group suggested as the zinc binding group for this group of HDACi.
CI-994 and compounds 5a–c presented EC50 (0.86 0.08
lM,
6.20 1.37 M and 0.68 0.12 M, 8.88 3.89 M respectively)
l
l
l
BRET-based dose response experiments were performed first to
determine the half maximal effective concentration (EC50) for
SAHA, CI-994 and their derivatives 4 and 5 and the released alco-
hols 3 (Table 2).
Compounds 3 were expected to have no HDAC inhibition as
confirmed by Fig. 2 (top right) where no BRET induction was ob-
served. They were also found inactive in our cancer cell lines as
shown by Fig. 5 (bottom).
SAHA and prodrugs 4 gave all BRET signals as a sigmoid dose re-
sponse profile (Fig. 2: SAHA top left, compounds 4 bottom left).
This confirmed the expected HDACi activity for SAHA and that pro-
drugs 4 were able to release SAHA to obtain restored HDAC inhibi-
tion. The prodrugs 4 by themselves are supposed to have no
activity on HDACs as the HDACi binding group is blocked
(hydroxamic acid). Some differences were observed between the
three releasing systems a, b and c and the free HDACi. For com-
pounds 4c with the HDACi release observed at all pH (Table 1), re-
lease in cells could be obtained at any stage of the penetration
steps from direct water attack. SAHA and its derivatives 4b–c pre-
not clearly in agreement with pH response in Table 1. This result
could be due to difference in permeability for compounds 5a and
5c compared to 5b which could be equilibrated by the respective
pH sensitivity once the compounds are in the cells. The BRET signal
being obtained from the nucleus, the diffusion across the nucleus
membrane should also be considered.
As opposed to SAHA which is acidic and can give anionic form in
biological pH, the free amino group of CI-994 is moderately basic
and may be partly protonated in acidic part of the cells. Thus, if
compounds 5 effectively release CI-994 in acidic compartments,
CI-994 can possibly be retained in these compartments. Such acidic
protonation of basic therapeutic agents is known and endosomal
trapping can be an issue.
3.4. Kinetics of restoration of HDAC inhibition measured by BRET assay
The kinetics of the HDACi release in living cells were also deter-
mined by BRET assays after different times of incubation with the
compounds. Fig. 3A shows the results obtained with free or linked
SAHA. For all SAHA derivatives, maximal BRET signal inductions
were obtained after 24–36 h of treatments. However, intensities of
the induced BRET signals were different. Indeed, SAHA induced the
higher increase in BRET signal (93.9 7.9 mBu) compared to 4c
(66.2 10.3 mBu), 4b (44.3 8.2 mBu), and 4a (15.8 4.9 mBu).
sented equivalent EC50 (0.42 0.08
0.75 0.07 M, respectively) indicating an effective release of the
inhibitor from 4b–c as expected while compound 4a presented
higher EC50 (7.17 3.19 M: Table 2) in agreement with the ob-
served stability at all pH (Table 1). Thus, for SAHA derivatives, a
lM, 0.59 0.18 lM, and
l
l
Fig. 2. Dose-dependent pharmacological characterization using BRET for free drugs, their derivatives 4 and 5, and the expected released alcohols 3. MeT-5A cells were
transfected with phRluc-C1-BrD and pEYFP-C1 histone H3 and treated for 16 h with increasing doses of the different compounds. Then, BRET signals were measured as
described in the materials and methods section. 100% corresponded to the maximal induced BRET signal measured. Results are the means SEM of three independent
experiments.
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

8
R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
Fig. 4. Kinetic of induction of BRET signal by free drugs, their derivatives 4 and 5.
Data are presented as percent of maximal induced BRET. Left: SAHA and derivatives
4. Right: CI-994 and derivatives 5. The method was the same as for Fig. 2. Graphic
represents the percent of maximal induced BRET signal measured for each
compounds. Results are the means SEM of three independent experiments.
CI-994 derivatives (CI-994: 53.3 8.2 mBu, 5a: 62.3 10.2 mBu, 5b:
54.3 13.3, 5c: 50.6 7.3 mBu).
As shown in Fig. 3A, the BRET induction for SAHA is faster and
higher than the other derivatives 4, and only compound 4c seemed
to have a maximum induction at shorter time. This can be the re-
sult of a partial early release of SAHA from 4c out of the cells with
subsequent diffusion inside cells, but in a smaller extent than
SAHA alone, combined with a fast cell penetration of 4c and subse-
quent SAHA release. For compound 4b, a softer profile is found
regarding the induction of the BRET signal, in agreement with a
prolonged release over time in cells. Considering the penetration
properties of the compound 4b, this delayed BRET induction,
reaching a plateau in our experiment, could also be explained by
slower penetration in cells. The low BRET signal for compound
4a could be due to its stability or its very slow cell penetration.
Obviously, for SAHA, acidic susceptibility and cell penetration
properties can be two ways to consider the prodrugs activities.
CI-994 and its derivatives demonstrated equivalent BRET re-
sponses, indicating that the release observed in acidic conditions
for all derivatives (Table 1) is maintained in cells leading to fast re-
lease of CI-994 outside cells or in acidic compartments of the cells.
Therefore, we hypothesized that an acidification-mediated release
can be involved based on the time required for signal induction of
the prodrugs when compared to their stability in simulated phys-
iological pH conditions (>3 days for compounds 4a and 5b). Com-
parison between all the compounds in relative BRET induction
mode is presented in Fig. 4, showing that the relative activity pro-
files are similar for all compounds. Only compound 5a presented a
relative slower release of the HDAC inhibitors in accordance with
its hydrolysis rate determined in buffered solution.
Fig. 3. Kinetic evaluation of the induction of histone H3 acetylation using BRET by:
(A) SAHA and 4a–c, (B) CI-994 and 5a–c. (Top) Results are expressed as the induced
BRET signal. (Bottom) Graphics represent the maximal induced BRET signal
measured for each compounds. Results are the means SEM of three independent
experiments. MeT-5A cells were transfected with phRluc-C1 BrD and pEYFP-C1
histone H3. Cells were treated during 8, 24 h, 36 h, or 48 h with 5
lM SAHA and 4a–
c (A) and with 10 M CI-994 and 5a–c (B). BRET signals were measured at the end of
l
the treatment as described in the BRET measurements section.
Fig. 3B shows the results obtained with free or linked CI-994. Maxi-
mal BRET signal inductions were obtained for 5c, 5b, and CI-994
after 24–36 h of treatment. 48 h of treatment was needed with 5a
to obtain the maximal induction of BRET signal. No significant differ-
ences were observed for the maximal induction of BRET signal for all
A good correlation between BRET induction over time and pH
response was thus observed for SAHA derivatives while for CI-
994 prodrugs, the differences between the 3 versions were not
significant.
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
9
Fig. 5. Dose-dependent effect on MPM and lung ADCA cell viability for free drugs and derivatives 4, 5, and 3. MPM (Meso 13 and Meso 34) and lung ADCA (ADCA 72 and ADCA
153) cells were treated with increasing doses of CI-994, 5a–c, SAHA, 4a–c, or 3a–c for 72 h. Cell viability was evaluated using Uptiblue cell counting reagent. Cell viability was
expressed as the percent of control. Results are the means SEM of four independent experiments performed on two MPM or two lung ADCA cell lines.
3.5. Determination of cancer cell viability after treatments with HDAC
inhibitors and their prodrugs
observed between MPM and lung ADCA cell lines for all com-
pounds tested (Table 3).
No toxicity was obtained with compounds 3 at the IC50 concen-
tration found for the free inhibitors (Fig. 5), confirming that the
activities are only due to the release of HDACi from prodrugs.
Lower IC50 was obtained with SAHA on MPM and ADCA cell
lines (1.47 0.65 lM and 1.71 0.94 lM, respectively). Intermedi-
ates IC50 were obtained with 4c and 4b, whereas 4a did not affect
The correlation between histone H3 acetylation and toxic effect
was determined for all compounds by viability tests on MPM and
lung ADCA cells treated with increasing doses for 72 h (Fig. 5 and
Table 3). All compounds, except 4a, were toxic for MPM and lung
ADCA cells depending on the concentration used. No major differ-
ences in half maximal inhibitory concentration (IC50
)
were
cells viability. These results are in agreement with the three types
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

10
R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
Table 3
Acknowledgments
IC50 values for HDACi-induced inhibition of cell growth.
Compounds
IC50
(
l
M) MPM
IC50
(
l
M) ADCA
Authors thank Agence Nationale de la Recherche (ANR), Centre
National de la Recherche Scientifique (CNRS), Ligue Interregional
Contre le Cancer (Comités Départementaux du Grand Ouest:
CD16, CD44, CD22 and CD56) the Ligue Nationale Contre le Cancer
for ID Grant, ARSMESO44, Nantes University Hospital and COST Ac-
tion TD0905. Authors thank ANR for RD, and FC Grants (ANR-08-
PCVI-030) and Ministère de la Recherche for FEB grant.
SAHA
4a
4b
4c
CI-994
5a
1.47 0.65
1.71 0.94
/
/
9.39 2.68
4.11 0.59
14.49 6.29
61.69 29.72
12.46 3.91
40.25 12.85
17.32 6.93
5.22 1.71
27.22 10.35
29.05 10.58
24.46 13.66
29.4 7.96
5b
5c
IC50 values were determined using GraphPad prism, Prism 5 for Windows, by curve
fitting using a sigmoidal dose–response model. Results are the means SEM of four
independent experiments performed on two MPM or two lung ADCA cell lines.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ejpb.2013.03.006.
of prodrugs for SAHA, considering both pH response and cell
penetration.
CI-994 and compounds 5a–c presented similar IC50 on the cell
lines tested in agreement with maximum BRET induction and fast
release.
References
[1] S. Choudhuri, Y. Cui, C.D. Klaassen, Molecular targets of epigenetic regulation
and effectors of environmental influences, Toxicol. Appl. Pharmacol. 245
(2010) 378–393.
4. Conclusions
[2] S. Henikoff, Nucleosome destabilization in the epigenetic regulation of gene
expression, Nat. Rev. Genet. 9 (2008) 15–26.
[3] M.J. Fazzari, J.M. Greally, Epigenomics: beyond CpG islands, Nat. Rev. Genet. 5
(2004) 447.
[4] S.S. Oliver, J.M. Denu, Dynamic interplay between histone H3 modifications
A synthesis of clickable pH responsive releasing system has
been achieved, applied to two different functional groups found
in HDAC inhibitors. The compounds were found stable when stored
neat or when stored in DMSO solutions at ꢁ20 °C. Except for the
series c, which is not stable whatever the pH conditions, the other
series gave interesting prolonged release over time in buffered
conditions. This was correlated with the in vitro slower restoration
of HDAC inhibition as determined by BRET assays. Due to the type
of releasing system and to the evident penetration of the prodrugs
into cells (BRET is measured in cells), endocytosis could be in-
volved, although co-localization of the molecules with acidic vesi-
cles could not be determined with such molecules due to lack of
any detectable properties. This internalization is also correlated
with toxic activities toward cancer cell lines. Systems 4 and 5 are
by themselves interesting intermediates as therapeutic agents
and the released alcohols 3 appeared inactive in our tests at the
usual IC50 concentrations for SAHA and CI-994. Indeed, because
linked HDACi are not active, prodrugs should display no or weak
hematological toxicities due to intravenous administration and re-
lease of the HDACi should occur mainly in tumor cells or in tumoral
tissues where acidity is usually higher. Therefore, these com-
pounds should decrease side effects associated with the use of
HDACi in clinical trials [10] and increase their clinical benefits.
These hypotheses would be verified on mouse models of tumor
to compare efficacy of free or linked HDACi. The prolonged effect
appeared very predictive when applied to SAHA. Considering the
releasing mode of these prodrugs, it is clear that the oral route will
not be feasible regarding the stomach acidity, except considering
molecules where stomach release is wanted and provided that
the released alcohols 3 have no effect in the stomach. In their cur-
rent form, such prodrugs could be used for parenteral administra-
tion and the remaining alkyne group used to obtain specific
applications. In particular, in vivo imaging could be of high interest
to determine the biodistribution of the prodrugs. We also planned
to use these alkyne prodrugs associated with azido-vectors (nano-
particles) able to reach the tumor site by the enhanced permeabil-
ity and retention effect and also able to enter cells by endocytosis,
where the acidity of the endosomes/lysosomes will be exploited to
release the active compounds. The in vivo evaluation of such func-
tional nanoparticles (biodistribution, pharmacokinetic, toxicity,
and anti-tumoral activity) will be the next challenge. Alternatively,
these molecules 4 and 5 could be used for in vitro fundamental
epigenetic researches when longer HDACi effects (or other epige-
netic targets) are required.
and protein interpreters: emerging evidence for
ChemBioChem 12 (2011) 299–307.
a ‘‘histone language’’,
[5] P.A. Jones, S.B. Baylin, The fundamental role of epigenetic events in cancer, Nat.
Rev. Genet. 3 (2002) 415–428.
[6] F. Dekker, H.J. Haisma, Histone acetyl transferases as emerging drug targets,
Drug Discov. Today 14 (2009) 942–948.
[7] P. Bertrand, Inside HDAC with HDAC inhibitors, Eur. J. Med. Chem. 45 (2010)
2095–2116.
[8] D.S. Hewings, T.P.C. Rooney, L.E. Jennings, D.A. Hay, C.J. Schofield, P.E. Brennan,
S. Knapp, S.J. Conway, Progress in the development and application of small
molecule inhibitors of bromodomain–acetyl-lysine interactions, J. Med. Chem.
http://dx.doi.org/10.1021/jm300915b.
[9] A.A. Lane, B.A. Chabner, Histone deacetylase inhibitors in cancer therapy, J.
Clin. Oncol. 27 (2009) 5459–5468.
[10] N. Martinet, P. Bertrand, Interpreting clinical assays for histone deacetylase
inhibitors, Cancer Manage. Res. 3 (2011) 117–141.
[11] R. Furumai, A. Matsuyama, N. Kobashi, K.H. Lee, M. Nishiyama, H. Nakajima, A.
Tanaka, Y. Komatsu, N. Nishino, M. Yoshida, S. Horinouchi, FK228
(depsipeptide) as a natural prodrug that inhibits class I histone deacetylases,
Cancer Res. 62 (2002) 4916–4921.
[12] L. Du, D.G. Musson, A.Q. Wang, Stability studies of vorinostat and its two
metabolites in human plasma, serum and urine, J. Pharm. Biomed. Anal. 42
(2006) 556–564.
[13] R.A. Parise, J.L. Holleran, J.H. Beuner, S. Ramalingam, M.J. Egorin, A liquid
chromatography–electrospray ionization tandem mass spectrometric assay
for quantitation of the histone deacetylase inhibitor, vorinostat
(suberoylanilide hydroxamicacid, SAHA), and its metabolites in human
serum, J. Chromatogr. B 840 (2006) 108–115.
[14] K.J. Dedes, I. Dedes, P. Imesch, A.O. von Bueren, D. Fink, A. Fedier, Acquired
vorinostat resistance shows partial cross-resistance to ‘second-generation’
HDAC inhibitors and correlates with loss of histone acetylation and apoptosis
but not with altered HDAC and HAT activities, Anticancer Drugs 20 (2009)
321–333.
[15] J.J. Xiao, J. Byrd, G. Marcucci, M. Grever, K.K. Chan, Identification of thiols and
glutathione conjugates of depsipeptide FK228 (FR901228), a novel histone
protein deacetylase inhibitor, in the blood, Rapid Commun. Mass Spectrom. 17
(2003) 757–766.
[16] T. Shiraga, Z. Tozuka, R. Ishimura, A. Kawamura, A. Kagayama, Identification of
cytochrome P450 enzymes involved in the metabolism of FK228, a potent
histone deacetylase inhibitor, in human liver microsomes, Biol. Pharm. Bull. 28
(2005) 124–129.
[17] (a) R.W. Robey, Z. Zhan, R.L. Piekarz, G.L. Kayastha, T. Fojo, S.E. Bates, Increased
MDR1 expression in normal and malignant peripheral blood mononuclear
cells obtained from patients receiving depsipeptide (FR901228, FK228,
NSC630176), Clin. Cancer Res. 12 (2006) 1547–1555;
(b) J.J. Xiao, A.B. Foraker, P.W. Swaan, S. Liu, Y. Huang, Z. Dai, J. Chen, W. Sadée,
J. Byrd, G. Marcucci, K.K. Chan, Efflux of depsipeptide FK228 (FR901228, NSC-
630176) is mediated by P-glycoprotein and multidrug resistance-associated
protein 1, J. Pharmacol. Exp. Ther. 313 (2005) 268–276.
[18] G. Scita, P.P. Di Fiore, The endocytic matrix, Nature 463 (2010) 464–473.
[19] (a) V.F. Patel, J.N. Hardin, J.M. Mastro, K.L. Law, J.L. Zimmermann, W.J.
Ehlhardt, J.M. Woodland, J.J. Starling, Novel acid labile COL1 trityl-linked
difluoronucleoside immunoconjugates: synthesis, characterization, and
biological activity, Bioconjug. Chem. 7 (1996) 497–510;
(b) E. Leikauf, F. Barnekow, H. Köster, Heterobifunctional trityl derivatives as
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006

R. Delatouche et al. / European Journal of Pharmaceutics and Biopharmaceutics xxx (2013) xxx–xxx
11
linking reagents for the recovery of nucleic acids after labeling and
immobilization, Tetrahedron 51 (1995) 3793–3802.
(c) E.S. Lee, Y.H. Bae, Recent progress in tumor pH targeting nanotechnology, J.
Control. Release 132 (2008) 164–170.
[20] J.M.J. Fréchet, E.R. Gillies, A.P. Goodwin, Acetals as pH-sensitive linkages for
drug delivery, Bioconjug. Chem. 15 (2004) 1254–1263.
[25] C.D. Hein, X.-M. Liu, D. Wang, Click chemistry, a powerful tool for
pharmaceutical sciences, Pharm. Res. 25 (2008) 2216–2230.
[26] J.J. Ritter, P.P. Minieri, A new reaction of nitriles. I. Amides from alkenes and
mononitriles, J. Am. Chem. Soc. 70 (1948) 4045–4048.
[27] F. Gueugnon, S. Leclercq, C. Blanquart, C. Sagan, L. Cellerin, M. Padieu, C.
Perigaud, A. Scherpereel, M. Gregoire, Identification of novel markers for the
diagnosis of malignant pleural mesothelioma, Am. J. Pathol. 178 (2011) 1033–
1042.
[21] (a) M. Hruby, C. Konak, K. Ulbrich, Polymeric micellar pH-sensitive drug
delivery system for doxorubicin, J. Control. Release 103 (2005) 137–148;
(b) L.M. Kaminskas, B.D. Kelly, V.M. McLeod, G. Sberna, D.J. Owen, B.J. Boyd,
C.J.H. Porter, Characterisation and tumour targeting of PEGylated polylysine
dendrimers bearing doxorubicin via a pH labile linker, J. Control. Release 152
(2011) 241–248.
[22] F. Wang, Y.-C. Wang, S. Dou, M.-H. Xiong, T.-M. Sun, J. Wang, Doxorubicin-
tethered responsive gold nanoparticles facilitate intracellular drug delivery for
overcoming multidrug resistance in cancer cells, ACS Nano 5 (2011) 3679–
3692.
[28] C. Blanquart, M. François, C. Charrier, P. Bertrand, M. Gregoire, Pharmacological
characterization of histone deacetylase inhibitor and tumor cell-growth
inhibition properties of new benzofuranone compounds, Curr. Cancer Drug
Targets 11 (2011) 919–928.
[23] (a) P.
Bertrand,
J.P.
Gesson,
Click
chemistry
with
O-
[29] N. Boute, S. Boubekeur, D. Lacasa, T. Issad, Dynamics of the interaction
between the insulin receptor and protein tyrosine-phosphatase 1B in living
cells, EMBO Rep. 4 (3) (2003) 313–319.
[30] M. Gonçalves, K. EStieu-Gionnet, T. Berthelot, G. Laïn, M. Bayle, X. Canron, N.
Betz, A. Bikfalvi, G. Déléris, Design, synthesis, and evaluation of original
carriers for targeting vascular endothelial growth factor receptor interactions,
Pharm. Res. 22 (2005) 1411–1421.
[31] N. Streidl, A. Antipova, H. Mayr, Suppression of common-ion return by amines:
a method to measure rates of fast SN1 reactions, J. Org. Chem. 19 (2009) 7328–
7334.
[32] J.K. Young, S. Pazhanisamy, R.L. Schowen, Energetics of carbonyl addition and
elimination. Kinetic manifestations of acyl substituent effects in anilide
hydrolysis, J. Org. Chem. 49 (1984) 4148–4152.
dimethylpropargylcarbamate for preparation of pH-sensitive functional
groups. a case study, J. Org. Chem. 72 (2007) 3596–3599;
(b) R. Delatouche, M. Mondon, A. Gil, G. Frapper, C. Bachmann, P. Bertrand,
Novel triazolyl derivatives for acidic release of amines, Tetrahedron 67 (2011)
401–407;
(c) M. Mondon, R. Delatouche, C. Bachmann, G. Frapper, C. Len, P. Bertrand,
Triazolyl derivatives for acidic release of alcohols, Eur. J. Org. Chem. (2011)
2111–2119;
(d) R. Delatouche, G. Frapper, C. Bachmann, P. Bertrand, Extending triazolyl-
based release under mildly acidic conditions to give aniline derivatives,
Synthesis 44 (2012) 1090–1094.
[24] (a) Z. Zhou, S. Li, Y. Zhang, M. Liu, W. Li, Promotion of proton conduction in
polymer electrolyte membranes by 1H-1,2,3-triazole, J. Am. Chem. Soc. 127
(2005) 10824–10825;
[33] A.R. Katritzky, B.E. Brycki, The mechanisms of nucleophilic substitution in
aliphatic compounds, Chem. Soc. Rev. 19 (1990) 83–105.
(b) P. Midoux, C. Pichon, J.J. Yaouanc, P.A. Jaffrès, Chemical vectors for gene
[34] R. Delatouche, A. Lesage, F. Collette, V. Heroguez, P. Bertrand, Investigations
into competitive cycloaddition/cyclization or elimination from 1,1-dimethyl-
propargylcarbamates of anilines, Aust. J. Chem. 64 (2011) 166–173.
delivery:
a current review on polymers, peptides and lipids containing
histidine or imidazole as nucleic acids carriers, Br. J. Pharmacol. 157 (2009)
166–178;
Please cite this article in press as: R. Delatouche et al., Design of pH responsive clickable prodrugs applied to histone deacetylase inhibitors: A new strategy
for anticancer therapy, Eur. J. Pharm. Biopharm. (2013), http://dx.doi.org/10.1016/j.ejpb.2013.03.006