Real-time monitoring of drug release

Article abstract of DOI:10.1039/b919329d

A new prodrug system, assembled using a distinctive coumarin linker, was demonstrated to report real-time activation and drug release in vitro.

Full text of DOI:10.1039/b919329d

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Real-time monitoring of drug releasew
a
b
Roy Weinstain,z Ehud Segal,z Ronit Satchi-Fainarob and Doron Shabat*a
Received (in Cambridge, UK) 17th September 2009, Accepted 25th November 2009
First published as an Advance Article on the web 10th December 2009
DOI: 10.1039/b919329d
A new prodrug system, assembled using a distinctive coumarin
linker, was demonstrated to report real-time activation and drug
release in vitro.
drug molecule as the end-unit, such a system can serve as a
reporting drug-delivery system. Coupling of the ON–OFF
fluorescent signal to the event of drug release provides
real-time information about the release process that can be
detected by fluorescent imaging techniques in a non-invasive
manner. As a model system, we synthesized molecule 5 with a
phenylacetamide trigger designed for cleavage by the enzyme
penicillin-G-amidase (PGA) and with the chemotherapeutic
drug melphalan as the end-unit (Fig. 2). Cleavage of
the phenylacetamide moiety by PGA generates an aniline
derivative, which undergoes spontaneous 1,6-elimination
followed by decarboxylation. The exposed secondary amine
then undergoes an intramolecular cyclization to release the
phenolate of coumarin. 1,8-Elimination eventually results in
the formation of free melphalan and fluorophore 4 (see ESIw
for disassembly mechanism).
Drug-delivery systems (DDSs) are being developed to
improve the therapeutic index of small molecule drugs.1–3
In order to evaluate the therapeutic effect of DDSs, their
pharmacokinetics,4 pharmacodynamics,5 cell permeation
efficiencies and pathways,6 and mechanisms of activation7
are being extensively studied. To date, however, crucial data
on time and location of drug release from the delivery system
in vivo, or even in living cells, cannot be directly obtained.
Real-time information on drug release would enable in vivo
kinetic studies of the release process. Therefore, DDSs that
instantaneously report on the release of their active drug could
be of great benefit, especially if the reported signal could be
detected in a non-invasive manner. Latent fluorophores (LFs)
are attractive candidates for this type of reporter.8,9 By
coupling LF activation to the drug release event in a delivery
system, real-time information about the release process can be
obtained using non-invasive fluorescence detection techniques.
Herein, we describe the design, synthesis and characterization
of the first reporting drug-delivery system (RDDS) for
in vitro use.
The release of melphalan from RDDS 5 upon incubation
with PGA in PBS (pH 7.4) was monitored by RP-HPLC
(Fig. 3A). The fluorescence generated from the linker was
monitored using a spectrophotometer (lex = 315 nm, lem
=
460 nm, Fig. 3B). When RDDS 5 was incubated with PGA,
free melphalan was released, accompanied by a gradual
increase in fluorescence emitted at 460 nm. In contrast, no
melphalan and no fluorescence were observed in the absence of
PGA. The observed short time-gap between melphalan release
and fluorescence generation is probably due to an additional
step required for fluorophore 4 formation (see Fig. 1). These
results demonstrate that upon specific activation of RDDS 5,
fluorescent signal is directly correlated to drug release.
The design of our RDDS was based on a 7-hydroxycoumarin
linker (compound 4) with a hydroxymethyl substituent
(Fig. 1). The phenolic alcohol of 4 is connected to a triggering
group (that can be activated either chemically or enzymatically)
and the hydroxymethyl substituent serves as a ‘‘handle’’ for
attachment of an end-unit (in this case, the drug).10–12 The
release of the end-unit is initiated by removal of the trigger of
molecule 1 and formation of phenolate 2. A spontaneous
1,8-elimination13 reaction then takes place, leading to the
release of the end-unit and generation of coumarin quinone-
methide derivative 3. Addition of a water molecule to the
reactive quinone-methide 3 leads to the formation of the
highly fluorescent coumarin derivative 4.
With these results in hand, we evaluated both the ability of
RDDS 5 to report drug release and the efficacy of the released
drug in cells. Human T-lineage acute lymphoblastic leukemia
MOLT-3 cells were treated with various concentrations of
RDDS 5 in the presence or absence of 1 mM PGA.
A colorimetric assay based on tetrazolium salt XTT was used
to evaluate cytotoxicity. The data from the cell proliferation
assay are presented in Fig. 4A. In the absence of PGA,
molecule 5 showed only minor cytotoxicity, even at relatively
high concentrations (IC50 4 100 mM). However, in the
presence of PGA, molecule 5 had cytotoxicity equivalent to
When the trigger of molecule 1 is attached, no fluorescence
is emitted.14 However, when the system undergoes specific
activation, the end-unit is released and fluorescence is
generated through formation of coumarin 4. By employing a
a School of Chemistry, Raymond and Beverly Sackler Faculty of
Exact Sciences, Tel-Aviv University, Tel Aviv 69978, Israel.
E-mail: chdoron@post.tau.ac.il; Fax: +972 (0)3 640 5761;
Tel: +972 (0)3 640 8340
b Department of Physiology and Pharmacology, Sackler Faculty of
Medical Sciences, Tel-Aviv University, Tel Aviv 69978, Israel
w Electronic supplementary information (ESI) available: Experimental
details, synthetic schemes, additional supporting data including NMR
spectra of all new compounds. See DOI: 10.1039/b919329d
z These authors contributed equally.
Fig. 1 Proposed disassembly mechanism of coumarin-based RDDS.
Chem. Commun., 2010, 46, 553–555 | 553
ꢀc
This journal is The Royal Society of Chemistry 2010

View Article Online
Cathepsin B expression is elevated in cancer cells and in tumor
endothelial cells,16,17 making it useful for prodrug18 and DDS
activation.19 A low molecular weight RDDS equipped with a
suitable trigger can be taken into cells via a non-specific
pathway, for example through pinocytosis, and will then
undergo specific activation by the targeted enzyme in the
endolysosomal environment.
To test this approach, we synthesized molecule 6, equipped
with the dipeptide Phe-Lys as the triggering-substrate for
cathepsin B20,21 and melphalan as the end-unit (Fig. 2).
Cleavage of the amide bond at the C-terminus of the lysine
initiates the disassembly cascade, resulting in the release of
free melphalan and the formation of fluorophore 4. Two
populations of MOLT-3 cells were subjected to treatment with
RDDS 6 in a cell growth inhibition assay. The first was grown
in complete medium (RPMI supplemented with 10% fetal
bovine serum) before treatment and the other was grown in
starvation medium (RPMI supplemented with 2% fetal bovine
serum) during the 24 hours prior to treatment. Under these
stress conditions, cells are known to increase the expression of
proteolytic enzymes.22,23 Cell starvation was thus used to
mimic the elevated expression of proteolytic enzymes observed
in cancerous tissue. When cells were treated with free
melphalan, there was no difference in its toxicity between
starved or non-starved MOLT-3 cells (data not shown). In
contrast, RDDS 6 was more toxic toward the starved cells
(IC50 = 4 mM) by 7.5 fold than towards non-starved cells
(IC50 = 30 mM) (Fig. 5A). We presume that the increased
cytotoxicity of RDDS 6 towards starved MOLT-3 cells
resulted from elevated expression of proteolytic enzymes,
including cathepsin B, which specifically activates the RDDS
trigger. The fluorescence emitted from the treated MOLT-3
cells was monitored in real-time. The fluorescence signal from
starved MOLT-3 cells treated with RDDS 6 was significantly
higher than the signal from non-starved cells (see ESIw). The
strong correlation between cytotoxicity and emitted fluores-
cence (Fig. 5B) demonstrates the ability of RDDSs, like 6, to
report on their activity. In order to determine whether any of
the fragments of the RDDS are toxic to the cells, a control
experiment was performed with model RDDS 7 (Fig. 2). This
model RDDS was equipped with the end-unit tryptophan
instead of melphalan and was thus not expected to be toxic.
Indeed as shown in Fig. 5A, RDDS 7 was not toxic toward
either MOLT-3 population.
Fig. 2 Chemical structures of synthetic coumarin-based RDDSs.
Fig. 3 (A) Melphalan released from RDDS 5 [300 mM] in 100 mL PBS
(pH 7.4) in the presence (blue squares) or absence (red circles) of
0.1 mg mLꢁ1 PGA. (B) Emitted fluorescence (lex = 315 nm, lem
=
460 nm) from 300 mM RDDS 5 in 100 mL PBS (pH 7.4) in the presence
(blue squares) or absence (red circles) of 0.1 mg mLꢁ1 PGA.
that of free melphalan (IC50 = 2.5 mM). No cytotoxicity was
observed for the linker coumarin 4. The fluorescence emitted
from the treated MOLT-3 cells over this same time course was
monitored in real-time. The data for MOLT-3 cells treated
with 5 mM RDDS 5 and PGA are shown in Fig. 4B. A gradual
increase in fluorescence was observed over 1 hour. No increase
in fluorescence was observed in the absence of PGA. These
results suggest that fluorescence emitted due to drug release
from RDDS 5 can be used to monitor growth inhibition
activity.
In order to achieve selective activation of a RDDS in cancer
cells,15 we chose the proteolytic enzyme cathepsin B to activate
the trigger. This enzyme is a member of the cathepsin family;
these proteases are abundant in endosomes and lysosomes.
Fig.
4
(A) Viability of leukemia MOLT-3 cells treated with
Fig. 5 (A) Viability of non-starved leukemia MOLT-3 cell line
treated with RDDS 7 (’) or RDDS 6 (m) and starved leukemia
MOLT-3 cells treated with RDDS 7 (&) or RDDS 6 (W). Cells were
incubated for 72 hours. (B) Correlation between viability of starved
leukemia MOLT-3 cells treated with varying concentrations of RDDS
6 and emitted fluorescence (lex = 315 nm, lem = 460 nm).
melphalan (J), RDDS 5 + PGA [1 mM] (&), RDDS 5 (’), and
coumarin 4 (B), at the indicated concentrations in RPMI medium. Cells
were incubated for 72 hours. (B) Emitted fluorescence (lex = 315 nm,
l
em = 460 nm) from MOLT-3 cells treated with 5 mM RDDS 5 in the
presence (blue diamonds) and absence (red squares) of 1 mM PGA.
ꢀc
This journal is The Royal Society of Chemistry 2010
554 | Chem. Commun., 2010, 46, 553–555

View Article Online
To further support our conclusions, we monitored the
activation of RDDS 6 in human umbilical vein endothelial
cells (HUVEC) by confocal microscopy. Cathepsin B has been
associated with angiogenesis processes and is observed at high
levels in lysosomes of tumor endothelial cells. HUVEC were
serum-starved, treated with RDDS 5 or RDDS 6, then
washed, stained, and prepared for microscopy using standard
procedures. When HUVEC were treated with RDDS 5, no
coumarin 4 fluorescence was observed, indicating that RDDS
5 is not activated (Fig. 6A–D). This result is consistent with
the observed inability of RDDS 5 to inhibit HUVEC growth
in a cell proliferation assay (data not shown). In contrast, the
confocal images indicated the presence of fluorophore 4 in the
HUVEC treated with RDDS 6 (Fig. 6E–H). The location of
the coumarin 4 molecules inside the cells was confirmed by
confocal Z-stack images (see ESIw). Intracellular coumarin
indicates that RDDS 6 molecules were internalized and then
were specifically activated by cathepsin B, releasing concomi-
tantly fluorophore 4 and free melphalan. As indicated in the
channel overlay (Fig. 6H), HUVEC incubated with RDDS 6
showed cytoplasmic accumulation of activated coumarin.
Although there is a short time-gap between the release of the
free drug and the released fluorescence (Fig. 2), the observed
signal can be calibrated to report the prodrug activation in
real-time. While, there are many examples of DDSs labeled with
fluorophores to allow pharmacokinetic evaluation,24 these
systems do not report release of the drug from the delivery
vehicle. The concept presented in this study describes the first
system in which the generation of the active free drug is
visualized and reported by a fluorescent signal. The emitted
fluorescence of the coumarin linker at a wavelength of 460 nm
was sufficient to monitor the DDS activation in vitro. However,
in order to monitor such DDSs in vivo, analogous linkers with
fluorescence emitted at longer wavelengths will be required.
In conclusion, we have introduced a novel coumarin-based
linker with latent fluorescence into reporting drug-delivery
systems. Coupling of latent fluorophore activation with a drug
release event resulted in DDSs that report cargo release
through an ON–OFF fluorescent signal. We showed that
PGA- and cathepsin-B-activated RDDSs signal their cytotoxic
activity toward MOLT-3 cells and HUVEC, respectively, by
emitting fluorescence. This allowed us to monitor in real-time
the drug release. We observed a strong direct correlation
between tumor cell growth inhibition activity and emitted
fluorescence in MOLT-3 cells. Using confocal microscopy,
we showed that the drug release event in HUVEC occurred
in the cytoplasm. The amount of drug release can be calculated
by quantifying the emitted fluorescence; this should allow
prediction of a DDS’s therapeutic effect and potential
side effects. Other RDDSs can be similarly designed by
introducing appropriate reporting agents and a variety of
potent anticancer drugs.
D. S. thanks the Israel Science Foundation (ISF)
for financial support and R. W. thanks Ian Seiple from
Phil S. Baran’s laboratory for helpful discussions.
Notes and references
1 M. E. Davis, Z. Chen and D. M. Shin, Nat. Rev. Drug Discovery,
2008, 7, 771–782.
2 (a) R. Haag and F. Kratz, Angew. Chem., Int. Ed., 2006, 45,
1198–1215; (b) D. Shabat, C. Rader, R. A. Lerner and
C. F. Barbas, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 6925–6930.
3 S. H. Medina and M. E. H. El-Sayed, Chem. Rev., 2009, 109,
3141–3157.
4 R. Satchi-Fainaro, M. Puder, J. W. Davies, H. T. Tran,
D. A. Sampson, A. K. Greene, G. Corfas and J. Folkman,
Nat. Med., 2004, 10, 255–261.
5 Y. Cheng and T. Xu, Eur. J. Med. Chem., 2008, 43, 2291–2297.
6 E. A. Goun, T. H. Pillow, L. R. Jones, J. B. Rothbard and
P. A. Wender, ChemBioChem, 2006, 7, 1497–1515.
7 R. Satchi-Fainaro, R. Mamluk, L. Wang, S. M. Short, J. A.
Nagy, D. Feng, A. M. Dvorak, H. F. Dvorak, M. Puder,
D. Mukhopadhyay and J. Folkman, Cancer Cell, 2005, 7, 251–261.
8 D. W. Domaille, E. L. Que and C. J. Chang, Nat. Chem. Biol.,
2008, 4, 168–175.
9 N. Johnsson and K. Johnsson, ACS Chem. Biol., 2007, 2, 31–38.
10 A. Gopin, N. Pessah, M. Shamis, C. Rader and D. Shabat, Angew.
Chem., Int. Ed., 2003, 42, 327–332.
11 A. Gopin, C. Rader and D. Shabat, Bioorg. Med. Chem., 2004, 12,
1853–1858.
12 D. Shabat, R. J. Amir, A. Gopin, N. Pessah and M. Shamis,
Chem.–Eur. J., 2004, 10, 2626–2634.
13 M. Shamis and D. Shabat, Chem.–Eur. J., 2007, 13, 4523–4528.
14 E. Leroy, N. Bensel and L. Reymond Jean, Bioorg. Med. Chem.
Lett., 2003, 13, 2105–2108.
15 L. Stern, R. Perry, P. Ofek, A. Many, D. Shabat and R. Satchi-
Fainaro, Bioconjugate Chem., 2009, 20, 1783–1791.
16 J. A. Foekens, J. Kos, H. A. Peters, M. Krasovec, M. P. Look,
N. Cimerman, M. E. Meijer-Van Gelder, S. C. Henzen-Logmans,
W. L. J. Van Putten and J. G. M. Klijn, J. Clin. Oncol., 1998, 16,
1013–1021.
17 T. Strojnik, J. Kos, B. Zidanik, R. Golouh and T. Lah, Clin.
Cancer Res., 1999, 5, 559–567.
18 S. C. Jeffrey, M. T. Nguyen, J. B. Andreyka, D. L. Meyer,
S. O. Doronina and P. D. Senter, Bioorg. Med. Chem. Lett.,
2006, 16, 358–362.
19 K. Miller, R. Erez, E. Segal, D. Shabat and R. Satchi-Fainaro,
Angew. Chem., Int. Ed., 2009, 48, 2949–2954.
20 G. M. Dubowchik, R. A. Firestone, L. Padilla, D. Willner,
S. J. Hofstead, K. Mosure, J. O. Knipe, S. J. Lasch and
P. A. Trail, Bioconjugate Chem., 2002, 13, 855–869.
21 M. A. Walker, G. M. Dubowchik, S. J. Hofstead, P. A. Trail and
R. A. Firestone, Bioorg. Med. Chem. Lett., 2002, 12, 217–219.
22 C. Cuvier, A. Jang and R. P. Hill, Clin. Exp. Metastasis, 1997, 15,
19–25.
23 B. F. Sloane, K. Moin, E. Krepela and J. Rozhin, Cancer Metastasis
Rev., 1990, 9, 333–352.
24 A. Agarwal, P. K. Tripathi, S. Tripathi and N. K. Jain, Curr. Drug
Targets, 2008, 9, 895–898.
Fig. 6 Subcellular confocal imaging of HUVEC treated with RDDS
5 (panels A–D) and RDDS 6 (panels E–H). HUVEC were incubated
with RDDS 5 or RDDS 6 and were fixed and stained with propidium
iodide (red) for nuclei and phalloidin-FITC (green) for actin fibers.
Activated coumarin (blue) was not detected in HUVEC treated with
RDDS 5 but was observed in cells treated with RDDS 6. Scale bars
represent 50 mm.
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 553–555 | 555