Targeted enzyme-responsive drug carriers: Studies on the delivery of a combination of drugs

Article abstract of DOI:10.1002/anie.200353204

Double whammy: Effective and selective killing of E. coli cells containing a specific enzyme was achieved by using targeted enzyme-responsive drug carriers (see picture) that release free drugs in a stringently controlled manner. The carrier lowered the e

Full text of DOI:10.1002/anie.200353204

Angewandte
Chemie
Enzyme-Responsive Drug Carriers
Targeted Enzyme-Responsive Drug Carriers:
Studies on the Delivery of a Combination of
Drugs**
Myung-Ryul Lee, Kyung-Hwa Baek, Hui Juan Jin,
Yong-Gyu Jung, and Injae Shin*
A wide variety of new synthetic compounds have been
screened for chemotherapeutic applications. However, many
drug candidates exhibit poor membrane permeability and/or
severe side effects caused by their lack of selectivity. These
problems can often be overcome by using highly specific
chemotherapeutic agents. On the other hand, considerable
efforts have been devoted to the creation of delivery systems
targeting infected cells (or malignant tumors); these systems
improve membrane permeability, therapeutic efficacy, and
selectivity.[1] In a recent effort aimed at designing novel
systems that release free drugs in a stringently controlled
fashion, we have synthesized and characterized carrier–drug
conjugates connected by linkers that are only cleaved by an
enzyme present in the infected cells. In addition, we have
[*] M.-R. Lee, K.-H. Baek, H. J. Jin, Y.-G. Jung, Prof. Dr. I. Shin
Department of Chemistry
Yonsei University
Seoul 120-749 (Korea)
Fax: (+82)2-364-7050
E-mail: injae@yonsei.ac.kr
[**] This work was supported by a Korea Research Foundation Grant
(KRF-2001-041-D00143).
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2004, 43, 1675 –1678
DOI: 10.1002/anie.200353204
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1675

Communications
developed a new molecular system to deliver a combination
of drugs to the target cells.
As shown in Scheme 1, 4-hydroxymandelic acid was used
as the framework for constructing the enzyme-responsive
Scheme 1. General structure of carrier–drug conjugates linked by 4-
hydroxymandelic acid. The pathway for enzyme-initiated release of free
drugs is depicted.
linkers 1.[2,3] Since the enzyme substrate portion in 1 is
attached to the 4-hydroxy group of the mandelate core, the
enzymatic cleavage processes should not be hindered by
either the carrier or drug moieties attached at the para
position. Enzymatic cleavage of the substrate portion in the
infected cells will result in spontaneous release of the free
drug by a 1,6-elimination process. The quinone methide
intermediate 2 formed by this process will rapidly react with a
nucleophile (for example, water)to generate 3. Diverse drugs
(or physical probes)containing OH, CO 2H, and NH2 moieties
can be coupled to the a-hydroxy group of the mandelate core
through ether, ester, and carbamate linkages, respectively.
Polymer, peptide, and dendrimer carriers can be coupled to
the linker through amide bonds.[4]
To demonstrate the potential of this targeted enzyme-
responsive drug-delivery system we synthesized carrier–
small-molecule conjugates 4a–c and 5, which contain a
phenylacetic acid residue as the substrate for cleavage by
bacterial penicillin G amidase (PGA)[5] and a TAT peptide as
the peptide carrier. TAT peptides are one of a number of
protein-transduction domains that have received considera-
ble attention as efficient delivery systems recently.[4a] These
carriers mediate intracellular uptake of small molecules and
macromolecules in a receptor-independent manner.[4a] We
selected the TAT peptide as the carrier because of its facile
preparation by solid-phase synthesis and its biodegradability,
which will be desirable following drug release.
Scheme 2. Synthesis of conjugates 4. Reagents: a) Cs2CO3, allyl bro-
mide; b) phenylacetyl chloride, DIEA; c) a: nalidixic acid, EDC, DMAP;
b: COCl2, 6-aminoquinoline; c: chlorambucil, EDC, DMAP;
d) [Pd(PPh3)4], AcOH, NMA; e) tert-butyl-6-aminohexanoate, EDC,
DMAP; f) TFA; g) pentafluorophenol, EDC. DIEA=diisopropylethyl-
amine, DMAP=4-dimethylaminopyridine, EDC=3-(3-dimethylamino-
propyl)-1-ethylcarbodiimide, NMA=N-methylalanine, Pfp=penta-
fluorophenyl, TFA=trifluoroacetic acid, TES=triethylsilane.
Nalidixic acid is an inhibitor of DNA gyrase,[6] and chlor-
ambucil is an antitumor agent known to cross-link DNA
strands.[7] 6-Aminoquinoline has an emission maximum at
370 nm in its conjugated form and at 550 nm in its free state.
Thus, conjugates 4b and 5 bearing 6-aminoquinoline were
prepared to probe the enzymatic cleavage in cells.[8]
After removal of the allyl group in 7a–c and condensation
of the exposed acid with tert-butyl-6-aminohexanoate fol-
lowed by deprotection of the tert-butyl group, the resulting
acids were converted into the pentafluorophenyl esters (8a–c)
for facile coupling with the TAT peptide in the next step.
Finally, conjugates 4a–c were obtained by the coupling of 8a–
c with peptide1 (YGRKKRRQRRR), which was prepared on
the solid support by the 9-fluorenylmethoxycarbonyl/tBu
strategy, and cleavage from the solid support.[9]
Conjugate 5 containing two small molecules (nalidixic
acid and 6-aminoquinoline)was prepared for the study of the
delivery of
a combination of drugs into target cells
The synthetic pathways for preparation of the conjugates
containing one (4a, nalidixic acid; 4b, 6-aminoquinoline; 4c,
chlorambucil)or two ( 5, nalidixic acid and 6-aminoquinoline)
small-molecule agents are illustrated in Schemes 2 and 3.
After protection of the carboxylic acid in 4-hydroxymandelic
acid (6)with an allyl group, selective phenylacetylation of the
phenolic hydroxy group and subsequent coupling of nalidixic
acid, 6-aminoquinoline, or chlorambucil (Ra–c moieties,
respectively)to the a-hydroxy group gave 7a–c (Scheme 2).
(Scheme 3). Compound 8a was coupled to the N-terminal
amino group of peptide2 (YGRKKRRQRRRC), which has
cysteine at the C terminus, on a solid support. After cleavage
of conjugate 10 from the solid support, the thiol group of the
cysteine residue in 10 was chemoselectively coupled to 9b,
which was prepared from condensation of 8b and 2-(2-
pyridyldithio)ethylamine, to produce 5.[9] This coupling
method was highly efficient for the synthesis of the conjugate
containing two different molecules.
1676
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 1675 –1678

Angewandte
Chemie
direct conjugation to the TAT peptide, not from a
combined antibacterial effect. Importantly, 4a did not
display an antibacterial effect (up to a concentration of
15.2 mm)on E. coli without a PGA gene. This demon-
strated that release of the free drug from the conjugate
was controlled by intracellular enzymes.
The release of small molecules by intracellular
enzymes was also examined by using the fluorescent
probe-containing conjugate 4b. For this purpose,
fluorescence microscopy was employed to analyze the
incubation mixtures of 4b (7.6 mm)and E. coli HB101
with and without a PGA gene. It was revealed that
PGA-expressing cells exhibited green fluorescence
when incubated with 4b for 20 min, whereas the
characteristic green fluorescence of 6-aminoquinoline
was not observed in E. coli cells without a PGA gene.
To show that the release process is cell selective, 4b was
incubated with a mixture of E. coli cells with and
without the PGA gene (Figure 2a and b). The obser-
vation that 6-aminoquinoline fluorescence was seen
only in cells with the PGA gene clearly demonstrates
that the developed delivery system was cell selective.
We then compared the rate of bacterial cell death
with 4a and nalidixic acid by counting colony-forming
units (CFUs)after a certain time of exposure to the
compounds (Figure 3).[12] Conjugate 4a killed PGA-
containing bacteria more rapidly than free nalidixic
acid did at concentrations of 3.8 and 276 mm, respec-
tively (2 multiples of the MICs of 4a and nalidixic acid,
respectively). However, E. coli cells that did not
Scheme 3. Synthesis of conjugate 5. DMF=N,N-dimethylformamide,
NEM=N-ethylmorpholine.
Initial examination of the enzyme-promoted release of
drugs or physical probes from these conjugates was per-
formed by analyzing mixtures of 4a–c incubated with PGA
from Escherichia coli in sodium phosphate buffer (pH 7.4)at
378C for 30 min. HPLC analysis of each mixture after
incubation revealed the presence of nalidixic acid, 6-amino-
quinoline, or chloroambucil along with phenylacetic acid and
substances that were judged by mass spectrometry to be the
products of water addition to the quinone methide inter-
mediates and cyclized products resulting from intramolecular
reactions of the quinone methide with lysine or arginine side
chains present in the TAT peptide (Figure 1a). Importantly,
these compounds were not produced when the conjugates
were incubated in PGA-free solutions.
The antibacterial properties of 4a were then investigated
by determining the minimum inhibitory concentrations
(MICs)of 4a and nalidixic acid against an E. coli HB101
strain transformed with a PGA gene.[9,10] The conjugate 4a
exhibited an MIC value about 70 times lower (1.9 mm)than
that of free nalidixic acid (138 mm). Many cationic peptides
have been known to have antimicrobial activities.[11] There-
fore, we tested the antibacterial activity of the TAT peptide
and found that it did not exhibit antibacterial activity up to a
concentration of 657 mm. We also examined the possible
synergistic effect of the TAT peptide and nalidixic acid on
antibacterial activity. It was found that the MIC value of
nalidixic acid was not altered by the addition of this peptide at
a concentration of 140 mm. Therefore, the better therapeutic
efficacy of 4a relative to free nalidixic acid resulted from its
Figure 1. HPLC chromatogram of the incubation mixtures of a) 4a or
b) 5 with PGA. a) A=a water-quenched product ([M+1]: m/z=1822);
B=a cyclized product([ M+1]: m/z=1804); C=phenylacetic acid;
D=nalidixic acid. b) A=6-aminoquinoline; B=a monocyclized
product([ M+1]: m/z=2245); C=a dicyclized product([ M+1]:
m/z=2227); D=nalidixic acid.
Angew. Chem. Int. Ed. 2004, 43, 1675 –1678
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1677

Communications
above, we have shown that the newly designed drug–carrier
conjugates effectively and selectively kill only those E. coli
cells containing a specific enzyme (for example, PGA). We
also demonstrated that the carrier lowered the effective
concentration of drug needed to kill the bacteria. Further-
more, we observed that two different compounds (a drug and
a fluorescent probe)were released from a bisconjugate by an
intracellular enzyme. This approach could be used for
targeted combination therapy. The general method developed
in this work may be useful for targeting virus-infected cells by
using carrier–antivirus-drug conjugates connected by linkers
that are viral-enzyme labile; malignant tumors with genes
encoding specific enzymes could be induced to release free
anticancer drugs from conjugates.[14]
Figure 2. Images of the incubation mixture of 4b and E. coli HB101
cells with and without the PGA gene: a) normal microsopic image,
b) fluorescence image, c) fluorescence image of the incubation mixture
of 5 and E. coli HB101 cells with the PGA gene.
Received: October 31, 2003 [Z53204]
Keywords: combination therapy · drug delivery · enzymes ·
.
peptides
[1] a)M. Yokoyama, T. Okano, Adv. Drug Delivery Rev. 1996, 21,
77; b)K. Kataoka in Controlled Drug Delivery: Challenges and
Strategies (Ed.: K. Park), American Chemical Society, Wash-
ington, 1997, pp. 49 – 71.
[2] Enzyme-labile safety-catch linkers have been developed for
solid-phase synthesis: a)R. Reents, D. A. Jeyaraj, H. Waldmann,
Adv. Synth. Catal. 2001, 343, 501; b)U. Grether, H. Waldmann,
Angew. Chem. 2000, 112, 1688; Angew. Chem. Int. Ed. 2000, 39,
1629; c)B. Sauerbrei, V. Jungmann, H. Waldmann, Angew.
Chem. 1998, 110, 1187; Angew. Chem. Int. Ed. 1998, 37, 1143.
[3] During the preparation of this manuscript, a similar approach for
the development of a tumor-targeting delivery system was
reported: A. Gopin, N. Pessah, M. Shamis, C. Rader, D. Shabat,
Angew. Chem. 2003, 115, 341; Angew. Chem. Int. Ed. 2003, 42,
327.
&
Figure 3. Rate of cell death caused by 4a and nalidixic acid. : E. coli
!
containing a PGA gene in the absence of active compounds; : E. coli
*
lacking a PGA gene in the presence of 3.8 mm 4a; : E. coli containing
a PGA gene in the presence of 276 mm nalidixic acid; : E. coli contain-
~
ing a PGA gene in the presence of 3.8 mm 4a.
[4] a)S. R. Schwarze, S. F. Dowdy, Trends Pharm. Sci. 2000, 21, 45;
b)A. T. Florence, N. Hussain, Adv. Drug Delivery Rev. 2001, 50,
S69.
[5] J. Y. Lu, D. A. Lowe, M. D. Kennedy, P. S. Low, J. Drug Targeting
1999, 7, 43; b)V. M. Vrudhula, P. D. Senter, K. J. Fischer, P. M.
Wallace, J. Med. Chem. 1993, 36, 919.
[6] W. A. Goss, W. H. Deitz, T. M. Cook, J. Bacteriol. 1965, 89, 1068.
[7] S. R. Rajski, R. M. Williams, Chem. Rev. 1998, 98, 2723.
[8] R. P. Haugland, Handbook of Fluorescenct Probes and Research
Chemicals, 6th ed., Molecular Probes, Eugene, 1996, chap. 10.
[9] See the Supporting Information.
contain a PGA gene grew at a similar rate in the presence of
3.8 mm 4a as they did in the control experiment.
For over a decade, combination therapies have been
exploited to improve therapeutic efficiency and/or to diminish
the toxicity of drugs.[13] Although a variety of drug-delivery
systems have been developed, these have rarely been
designed to deliver drug combinations. We examined the
enzyme-promoted release properties of conjugate 5, which
contained two different compounds. HPLC analysis of a
mixture of 5 incubated with PGA for 30 min showed that
nalidixic acid and 6-aminoquinoline were produced along
with mono- and dicyclized products (Figure 1b). This obser-
vation indicates that the conjugate underwent efficient
enzyme-responsive release of a combination of compounds.
To further investigate the possibility that two drugs could
be delivered into cells, the same conjugate 5 (3.8 mm)was
incubated with E. coli possessing the PGA gene. Fluorescence
microscopy images of cells after incubation for 20 min showed
that 6-aminoquinoline was released (Figure 2c). In addition,
the bacterial cells were nearly completely killed by 5 after
incubation for 8 h. These results clearly indicated that the
designed systems were capable of delivering a combination of
drugs to target cells.
[10] MIC values were determined with an inoculum of 107 CFUmLÀ1
of E. coli; “Methods for Dilution Antimicrobial Susceptibility
Tests for Bacteria that Grow Aerobically”: Approved Standards
M7-A4, 4th ed., National Committee for Clinical Laboratory
Standards, Wayne, PA, 1997.
[11] R. E. Hancock, R. Lehrer, Trends Biotechnol. 1998, 16, 82.
[12] S. Navon-Venezia, R. Feder, L. Gaidukov, Y. Carmeli, A. Mor,
Antimicrob. Agents Chemother. 2002, 46, 689.
[13] a)E. Gradelski, B. Kolek, D. P. Bonner, L. Valera, B. Minassian,
J. Fung-Tomc, Int. J. Antimicrob. Agents 2001, 17, 103; b)H. C.
Neu, Rev. Infect. Dis. 1989, 11(Suppl. 5), S1025.
[14] a)K. W. Culver, Z. Ram, S. Wallbridge, H. Ishii, E. H. Oldfield,
R. M. Blaese, Science 1992, 256, 1550; b)B. E. Huber, C. A.
Richards, E. A. Austin, Ann. N.Y. Acad. Sci. 1994, 716, 104;
c)T. A. Connors, Gene Ther. 1995, 2, 235; d)G. Cavallaro, G.
Pitarresi, M. Licciardi, G. Giammona, Bioconjugate Chem. 2001,
12, 143.
In conclusion, we have developed an efficient enzyme-
responsive drug-delivery system. In the studies described
1678
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 1675 –1678