Novel inhibitors of carboxypeptidase G2 (CPG2): Potential use in antibody-directed enzyme prodrug therapy

Article abstract of DOI:10.1021/jm990004i

The design and synthesis of potent thiocarbamate inhibitors for carboxypeptidase G₂ are described. The best thiocarbamate inhibitor N-(p- methoxybenzenethiocarbonyl)amino-L-glutamic acid 6d, chosen for preliminary investigations of in vitro ant

Full text of DOI:10.1021/jm990004i

J . Med. Chem. 1999, 42, 951-956
951
Expedited Articles
Novel In h ibitor s of Ca r boxyp ep tid a se G2 (CP G2): P oten tia l Use in
An tibod y-Dir ected En zym e P r od r u g Th er a p y
Tariq H. Khan,*,† Ebun A. Eno-Amooquaye,† Frances Searle,† Pat J . Browne,§ Helen M. I. Osborn,‡ and
Philip J . Burke§
Imperial College of Science, Technology and Medicine, Charing Cross Site, Medical Oncology, St. Dunstan’s Road, London W6
8RF, U.K., Department of Chemistry, University of Reading, Whiteknights, Reading RG6 6AD, U.K., and Enzacta Ltd.,
Building 115, Porton Down Science Park, Salisbury, Wilts SP4 OJ Q, U.K.
Received J anuary 4, 1999
The design and synthesis of potent thiocarbamate inhibitors for carboxypeptidase G2 are
described. The best thiocarbamate inhibitor N-(p-methoxybenzenethiocarbonyl)amino-L-
glutamic acid 6d , chosen for preliminary investigations of in vitro antibody-directed enzyme
prodrug therapy (ADEPT), abrogated the cytotoxicity of a combination of A5B7-carboxypep-
tidase G2 conjugate and prodrug PGP (N-p-{N,N-bis (2-chloroethyl)amino}phenoxycarbonyl-
L-glutamate) toward LS174T cells. This is the first report of a small-molecule enzyme inhibitor
proposed for use in conjunction with the ADEPT approach.
In tr od u ction
enzyme conjugate to the liver by galactose receptors.3b
This successful approach has allowed significant early
reduction of enzyme activity in the serum, with minimal
reduction of enzyme activity at the tumor site,4 and
consequently the prodrug may be administered 1 h after
the galactosylated antibody (SB43-gal). A recent ADEPT
clinical trial utilizing SB43-gal as a clearing agent
confirms the importance of the assisted clearing step,
by reducing the toxicity of the therapy. In addition, the
greater sensitivity of the detection system used high-
lights the efficiency of this plasma clearing process with
SB43-gal.5
In this paper we describe a possible cheaper, alterna-
tive, complementary refinement to ADEPT which may
also allow early administration of prodrugs and result
in minimal normal host tissue toxicity and maximum
cancer therapy. This novel method of increasing the
differential concentration of active enzyme between the
tumor site and other normal host tissues involves the
use of small-molecule enzyme-inactivating agents. Our
strategy aims to chemically inhibit (preferably irrevers-
ibly) residual levels of non-tumor-localized enzyme while
sustaining a sufficient quantity of the enzyme activity
at the tumor site for therapy.
Nonspecific enzyme activity in the recently developed
cancer treatment strategy, antibody-directed enzyme
prodrug therapy (ADEPT), can be a serious limitation
to such a potentially powerful method of cancer treat-
ment. ADEPT has the advantage of being suitable for
a wide range of cancers1a-c and was pioneered to enable
selective delivery of cytotoxic agents to targeted tumors,
via the use of antibody-enzyme conjugates.2 In mice
xenograft models,3a after optimum localization of the
A5B7-carboxypeptidase G2 (CPG2) conjugate to the
targeted tumors (typically 20 h), followed by slow
plasma clearance of the excess conjugate (typically 52
h), a prodrug, which is a specific substrate for only the
targeted enzyme, is administered. In theory, given
sufficient time to allow the conjugate in the plasma and
normal tissues to be removed, the subsequent genera-
tion of the cytotoxic drug occurs solely at the tumor site,
eliminating nonlocalized toxicity.
In reality, after maximum tumor penetration by the
antibody-enzyme conjugate, any attempt to administer
the prodrug earlier than 72 h results in toxicity,
particularly myelosuppression. Consequently, assisted
strategies to remove excess circulating conjugate over
the natural clearance mechanism are required. A suc-
cessful approach would have the benefits of allowing
early prodrug administration, at higher concentrations
and for a much longer period of time. Assisted clearance
has been achieved in animal models and clinically using
an antibody (SB43) directed to the CPG2 component of
the antibody-enzyme conjugate. This antibody has also
been galactosylated (SB43-gal), thus ensuring a more
rapid and efficient plasma clearance of any antibody-
For this approach to be investigated inhibitors for
CPG2 had to be obtained. Consequently, we designed
and synthesized a series of novel compounds which we
hoped would inhibit the enzyme CPG2, in both the free
state and the conjugated state. The preparation and
inhibitory properties of one successful series of com-
pounds are presented in this paper.
Ch em istr y
A series of potential inhibitors for CPG2 were initially
designed, and these are shown in Table 1. A general
chemical synthesis of the thiocarbamate derivatives is
given in Scheme 1.
* To whom all correspondence should be sent.
† Imperial College of Science, Technology and Medicine.
‡ University of Reading.
§ Enzacta Ltd.
10.1021/jm990004i CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/04/1999

952 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 6
Khan et al.
Ta ble 1. Structure and Biological Data for Potential Inhibitors
of Carboxypeptidase G2
Sch em e 1. General Synthesis of Thiocarbamate
Derivatives 6 as Potential Inhibitors for CPG2
c
The synthetic strategy chosen involved N-activation
with p-nitrophenyl chloroformate 1 of the R,γ-di-tert-
butyl-L-glutamic acid hydrochloride derivative 2, in the
presence of triethylamine (TEA), to result in carbamate
3. The crude activated glutamate derivative 3 was
sufficiently pure to be used directly in the next stage,
and subsequent treatment with a range of thiol com-
pounds 4, under argon or nitrogen, yielded, after column
chromatography on silica gel, the purified coupled
products 5. The derivatives 5 were deprotected under
acidic conditions to result in white precipitates 6. These
were usually sufficiently pure to be tested in the
biological systems. Some of the derivatives were recrys-
tallized from a mixture of hexane and ethyl acetate, and
these proved hygroscopic and formed gums very easily
after filtration or on standing. The final deprotected
products 6 were characterized by 1H NMR and thin-
layer chromatography and were shown to be pure by
combustion analysis. Table 1 shows the targets success-
fully synthesized together with their biological proper-
ties.
a
Apparent inhibition. b Cytotoxicity toward LS174T cells. c NDI,
no detectable inhibition; NC, not classical linear plot; ND, not
determined.
F igu r e 1. Structures of phenol mustard glutamate prodrug
(PGP) and phenol mustard active drug (PM).
x-axis giving rise to different Km values for each plot.
These results suggest that 6d exhibits noncompetitive
inhibition with respect to methotrexate.
The ability of CPG2 to activate a phenol mustard
prodrug (PGP), Figure 1, was demonstrated by exposing
LS174T cells to PGP in the presence of F(ab′)2-A5B7-
CPG2. Figure 3 shows the dose-response curves for
PGP against LS174T. The IC50 values obtained from the
cytotoxicity assays are shown in Table 3. The LS174T
cell line showed differential susceptibility to PGP alone
(IC50 value of 23.22 ( 0.47 µM) and F(ab′)2-A5B7-CPG2
enzyme generated phenol mustard active drug (IC50
value of 0.8 ( 0.03 µM). The ability of 6d to inhibit CPG2
activity was evaluated by treating the cells with the
conjugate F(ab′)2-A5B7-CPG2 and then simultaneously
with PGP and a fixed concentration of 6d . It was found
that the cytotoxic activity derived from the prodrug in
the presence of F(ab′)2-A5B7-CPG2 and 6d was greatly
diminished; however, in the absence of the conjugate
F(ab′)2-A5B7-CPG2 co-administration of a mixture of
Biologica l Eva lu a tion
The derivatives in Table 1 were tested in an in vitro
cytotoxicity assay (IC50) and also a Dixon plot (Ki) assay
system. Generally these thiocarbamate derivatives were
found to be nontoxic toward LS174T cells (>500 µM).
Furthermore the apparent Ki ranged from 0.3 µM for
6d to 165 µM for 6a . Derivative 6d was found to be the
most potent CPG2 inhibitor and had low toxicity to
LS174T cells. Therefore, this was chosen as our opti-
mum candidate for studying the detailed enzyme kinet-
ics profile and an in vitro ADEPT investigation. The
Lineweaver-Burk plots (Figure 2) and the apparent
Vmax and apparent Km values listed in Table 2 show that
Vmax decreases with increasing concentration of 6d and
that the family of reciprocal plots intersects above the

Novel Inhibitors of CPG2
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 6 953
Ta ble 3. Cytotoxicity Data for PGP Alone and Due to
Conjugate Turnover of Prodrug PGP in the Presence of
Inhibitor 6d against LS174T Cells
IC50 (µM)
PGP
23.22 ( 0.47
0.8 ( 0.03
F(ab′)2-A5B7-CPG2 + PGP
F(ab′)2-A5B7-CPG2 + PGP + 6d (25 µM)
F(ab′)2-A5B7-CPG2 + PGP + 6d (50 µM)
PGP + 6d (50 µM)
3.3 ( 0.19
16.2 ( 0.66
24.01 ( 0.46
Discu ssion
At the outset of this work no reports of any inhibitors
for CPG2 were known.6 In addition no high-resolution
X-ray crystallography data was available for this en-
zyme.7 However, limited information was available
concerning the closely related enzymes CPG and CPG1.
For example, it had already been demonstrated that
certain peptides, notably N-(benzyloxycarbonyl)glutamate
and N-benzoylglutamate, were moderate competitive
inhibitors of CPG.8 Our initial approach was to design
inhibitors for CPG2 which incorporated important struc-
tural features present in substrates for this enzyme, as
displayed in recent publications.9-11 A survey of the
literature suggested that a free R-carboxylate moiety on
the L-glutamate residue was essential for substrate
specificity in the case of CPG2; a benzene ring should
lie in close proximity to the carbonyl group of the amide
bond, as in folates and methotrexate,12 but the linkage
to the cleaved amide bond could vary between amide,
carbamate, thiocarbamate, and urethane moieties.13 It
is well-established that two zinc ions are associated with
each of the protein monomers of the dimeric enzyme,
and these zinc ions are essential for the activity of this
enzyme.14 Since zinc ions have a high affinity for sulfur,
protected sulfur moieties were incorporated within our
compounds to potentially result in active-site-directed
irreversible inhibitors.15 Recent high-resolution crystal-
lographic evidence, published during the course of this
work,16 confirmed that the zinc ions were indeed part
of the active site of the enzyme.
F igu r e 2. Lineweaver-Burk plot of the reciprocal of the
initial reaction velocity (in units of min/µM) versus the
reciprocal of methotrexate concentration at b, methotrexate
alone; 9, 0.5 µM 6d ; and [, 1.0 µM 6d . CPG2 concentration in
the reaction solution was 0.2866 µg/mL.
Ta ble 2. Apparent Vmax and Apparent Kmax Data from
Lineweaver-Burk Plots of MTX Turnover by CPG2 in the
Presence of Inhibitor 6d
concn of 6d (µM)
Vmax
Km
0
0.5
1.0
79.99 ( 5.69
47.66 ( 5.03
42.28 ( 5.61
10 ( 0.76
28.57 ( 0.64
31.13 ( 0.89
Biodistribution studies on tissue concentrations of
antibody-CPG2 conjugate, prodrugs and active drugs
of the benzoic acid mustard series in xenograft models,17
have identified encouraging biodistribution properties
which validate exploring small-molecule inhibitors for
use in ADEPT. For example, the natural biodistribution
of a benzoic acid mustard glutamate prodrug illustrates
that, between 15 min and 4 h, higher concentrations of
prodrugs occur in the liver, lung, and kidney than in
the tumor xenografts.3a Furthermore, recent findings in
tumor vascular architecture research have illustrated
that due to local flow-limited transport, plasma-borne
agents are poorly delivered to tumors.18 Therefore, these
results argue that, provided the biological distribution
of the inhibitor is similar to that of the prodrug, greater
concentrations of the inhibitor will be available for
inhibition of the residual enzyme in normal tissues than
are required to fully suppress all the enzyme activity at
the tumor. This should allow administration of a high
concentration of prodrugs without causing any myelo-
suppression, immediately after inhibition of the conju-
gate in the plasma, thus allowing a completely intact
immune system to function together with the ADEPT
strategy to eradicate the tumor.
F igu r e 3. Dose-response curves of PGP on LS174T cells. The
cells were treated as described in the Experimental Section.
All cells were exposed to PGP for 1 h: 9, PGP alone; b, F(ab′)2-
A5B7-CPG2 + PGP; 4, F(ab′)2-A5B7-CPG2 + PGP + 6d (25
µM); 0, F(ab′)2-A5B7-CPG2 + PGP + 6d (50 µM).
6d and PGP had little effect on the IC50 value of PGP
alone (Table 3). In the presence of 25 µM 6d the IC50
increased to 3.3 ( 0.19 µM and a 20-fold increase (IC50
16.2 ( 0.66 µM) was observed in the presence of 50 µM
6d . This shift toward higher IC50 values suggests that
activation of prodrug is inefficient in the presence of 6d
and illustrates complete inhibition of CPG2 activity by
6d .

954 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 6
Khan et al.
corrected. Microanalyses were performed by C.H.N. Analysis
Ltd., Alpha House, Countesthorpe Road, South Wigston,
Leicester LE8 2PJ , U.K. Chemicals were purchased from
Aldrich, The Old Brickyard-New Road-Gillingham-Dorset SP8
4XT, England. Di-tert-butyl-L-glutamic acid hydrochloride was
obtained from BaChem or Sigma, Poole.
Our first series of potential inhibitors incorporated a
thiocarbamate functionality (Scheme 1). This work with
the thiocarbamate derivatives, Table 1, has shown that
the ability to inhibit CPG2 is sensitive to substituent
variations on the benzene ring and also the chirality of
glutamic acid. On comparing the lipophilic para-
substitution (6d ,g) with the lipophilic meta-substitution
(6f,h ), we find that the para-substituted derivatives are
consistently better inhibitors than the meta-substituted
derivatives. By placing hydrophilic functional groups in
the para-position of the benzene ring of the thiocarbam-
ate derivative, inhibition efficiency is reduced by a factor
of at least 6 compared to the lipophilic moieties (6d ).
Furthermore, inhibitory properties cannot be improved
by replacing the phenyl ring by a pyridyl ring, as in 6a .
On forming the N-oxide 6a of the pyridyl compound, the
inhibitory properties are again further reduced by a
factor of 60 compared to 6d . The presence of electron-
withdrawing groups (Br 6i, Cl 6j) on the benzene ring
destabilizes the thiocarbamate bond to such an extent
that no enzyme inhibition is observed. It is concluded
that the rate of hydrolysis is too rapid in the aqueous
media for these particular derivatives to exert an
inhibitory effect on CPG2. The rate of hydrolysis of the
thiocarbamate linkage may partly explain the low
inhibitory properties of the pyridyl N-oxide derivative.
The electron-deficient pyridyl ring activates the thio-
carbamate bond to hydrolysis in aqueous media. This
would suggest electron-donating substituents on the
phenyl ring would stabilize the thiocarbamate bond,
thereby making the derivative a good inhibitor. Electron-
donating moieties on the phenyl ring, such as 6d -h ,
were found to have much better inhibitor properties. In
addition, the chirality of glutamic acid, for example, in
6e, the D-isomer of 6d , is very important for obtaining
the correct interaction with the enzyme active site. The
p-methoxy derivative 6d was found to have the best
overall properties from all the thiocarbamate derivatives
synthesized.
Solvents were used as commercially available. However,
where dry solvents were required a drying agent, such as
calcium hydride, was used to dry the solvent by refluxing the
solvent in the presence of the drying agent for 1-2 h in an
inert atmosphere, followed by distillation at atmospheric
pressure.
Gen er a l Syn th etic P r oced u r e for Th ioca r ba m a te De-
r iva tives, Sch em e 1. Di-tert-butyl-L-glutamic acid HCl salt
2 (2.05 g, 6.95 mmol) was activated by addition of p-nitrophe-
nyl chloroformate 1 (1.4 g, 6.95 mmol) in dry dichloromethane
(20 mL) in the presence of triethylamine (2 mL, 14.4 mmol).
The reaction mixture was heated under reflux for 25 min and
then stirred for 1 h at room temperature. The reaction mixture,
containing mainly the N-activated glutamic acid derivative 3,
was flushed with argon, and then a solution of the thiol 4 (7.70
mmol) in dichloromethane (20 mL) was added and heated to
reflux for 10 min. The cooled solution was then stirred for 5 h
at room temperature and monitored by TLC for disappearance
of starting material. The reaction mixture was concentrated
in vacuo and then stirred with ethyl acetate for 30 min; the
precipitate was filtered and the filtrate concentrated in vacuo.
The crude residue was chromatographed on silica gel (CH2-
Cl2). The fully protected derivative 5 was isolated as a colorless
oil, which was dissolved in hexane or CH2Cl2 (30 mL), treated
with HCl(g), and stirred at room temperature overnight. The
diacid product 6, a white precipitate, was filtered, washed with
hexane or CH2Cl2, and dried in vacuo to result in the
analytically pure thiocarbamate derivatives 6 (Scheme 1).
N-(p-Th iop yr id in ylca r bon yl-N-oxid e)-L-Glu ta m ic Acid
(6a ). p-Thiopyridine was dissolved in dichloromethane (10 mL)
and added to the p-nitrophenyl-activated di-tert-butylglutamic
acid intermediate 3. The reaction mixture was stirred at room
temperature and monitored by TLC; after all the starting
material had reacted the reaction mixture was treated with 1
equiv of m-chloroperbenzoic acid and stirred for 24 h. The
reaction mixture was concentrated in vacuo and chromato-
graphed on silica. The purified protected product 5a was
treated with gaseous hydrogen chloride and stirred for 48 h.
The solution was concentrated in vacuo to result in a white
precipitate. Trituration with ethyl acetate (to remove m-
chlorobenzoic acid) resulted in a hygroscopic white solid 6a .
Anal. (C11H12N2O6S) C,H,N.
Incubation of 50 µM 6d in a cell culture assay in the
presence of CPG2 (0.2866 µg/mL) has shown that the
inhibitor can remove practically all toxic effects due to
the turnover of the potent phenol mustard prodrug
(PGP).
N-(m -Am in op h en ylth ioca r bon yl)-L-glu ta m ic Acid (6b).
Modification to the general procedure: The final acidification
deprotection step involved a mixture of 1:5 methanol:dichlo-
romethane and gaseous dry hydrogen chloride. A yellow solid
precipitated out of the reaction mixture; this was filtered and
dried. The impurity was removed by crystallization from hot
methanol and ethyl acetate. The required product 6b was
isolated from the filtrate on concentration in vacuo, as an
orange gum. The product was found to be hygroscopic. Anal.
(C12H14N2O5S‚0.2EtOAc) C,H,N.
N-(p-Am in op h en ylth ioca r bon yl)-L-glu ta m ic Acid (6c).
Modification to the general procedure: The final acidification
deprotection step involved a mixture of 1:5 methanol:dichlo-
romethane and gaseous hydrogen chloride. A yellow solid
precipitated out of the reaction mixture; this was filtered and
dried. The impurity was removed by crystallization from hot
methanol and ethyl acetate. The required product 6c was
isolated from the filtrate on concentration in vacuo, as an
orange gum. The product 6c was found to be hygroscopic. Anal.
(C12H14N2O5S‚0.2EtOAc) C,H,N.
Con clu sion s
In this work a novel noncompetitive inhibitor 6d of
CPG2 has been discovered. This suggests that 6d is an
ideal candidate to test the novel hypothesis of using
small-molecule inhibitors as an enzyme-clearing strat-
egy in antibody-directed enzyme prodrug therapy in
vivo.
Exp er im en ta l Section
NMR spectra were determined on a 250-MHz Brucker
instrument (ULIRS, Queen Mary and Westfield College) with
chemical shifts (δ) reported relative to Me4Si. TLC was
performed on Merck aluminum sheets silica gel F254 (Art.
5554); spots were visualized under 254- and 365-nm UV light
and with the aid of ninhydrin (0.1% w/v in acetone) or
bromophenol blue (0.1% w/v in ethanol) solutions. In some
cases the use of sodium periodate (1% w/v in distilled water)
and heat was used to identify easily oxidizable species. Column
chromatography was carried out on Merck 1.09385 silica gel
60 (230-400 mesh). Melting points were measured on a
Gallenkamp capilliary melting point apparatus and are not
N-(p-Meth oxyph en ylth iocar bon yl)-L-glu tam ic acid (6d):
1H NMR (250 MHz, DMSO-d6) δ 8.20 (1H, d, J ) 8.0 Hz, NH),
7.10 (2H, d, J ) 8.0 Hz, ArH), 6.7 (2H, d, J ) 8.0 Hz, ArH),
3.9 (1H, m, CH), 3.5 (3H, s, OMe), 2.1 (2H, m, CH2), 1.6 (2H,
m, CH2); mp 115 °C; HPLC stability studies indicated a t1/2 of

Novel Inhibitors of CPG2
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 6 955
34 min in phosphate buffer at pH 7.3 and 37 °C.19 Anal.
(C13H15NO6S) C,H,N.
N-(p-Meth oxyph en ylth iocar bon yl)-D-glu tam ic acid (6e):
1H NMR (250 MHz, DMSO-d6) δ 8.20 (1H, d, J ) 8.0 Hz, NH),
7.10 (2H, d, J ) 8.0 Hz, ArH), 6.7 (2H, d, J ) 8.0 Hz, ArH),
3.9 (1H, m, CH), 3.5 (3H, s, OMe), 2.1 (2H, m, CH2), 1.6 (2H,
m, CH2); mp 110 °C. Anal. (C13H15NO6S) C,H,N.
N-(m -Met h oxyp h en ylt h ioca r b on yl)-L-glu t a m ic a cid
(6f): 1H NMR (250 MHz, DMSO-d6) δ 8.80-6.60 (5H, m, ArH,
NH), 4.10 (1H, m, CH), 3.65 (3H, s, OMe), 2.40-1.60 (4H, m,
[CH2]2). Anal. (C13H15NO6S) C,H,N.
N-(p-Meth ylp h en ylth ioca r bon yl)-L-glu ta m ic a cid (6g):
1H NMR (250 MHz, DMSO-d6) δ 12.4-12.5 (2H, b, COOH),
8.55(1H, d, CONH), 7.35 (2H, d, J ) 8 Hz, ArH), 7.24 (2H, d,
J ) 8 Hz, ArH), 4.27-4.14 (1H, m, CH), 2.40-2.22 (5H, m,
CH3 + CH2), 2.09-1.92 (1H, m, CH2), 1.91-1.73 (1H, m, CH2).
Anal. (C13H15NO5S) C,H,N.
N-(m -Meth ylp h en ylth ioca r bon yl)-L-glu ta m ic a cid (6h ):
1H NMR (250 MHz, DMSO-d6) δ 12.4-12.5 (2H, b, COOH),
8.60 (1H, d, CONH), 7.34-7.19 (4H, m, ArH), 4.26-4.14 (1H,
m, CH), 2.38-2.24 (5H, m, CH3 +CH2), 2.09-1.92 (1H, m,
CH2), 1.90-1.73 (1H, m, CH2). Anal. (C13H15NO5S‚H2O) C,H,N.
N-(p-Br om op h en ylth ioca r bon yl)-L-glu ta m ic a cid (6i):
1H NMR (250 MHz, DMSO-d6) δ 12.4-12.5 (2H, b, COOH),
8.71 (1H, d, CONH), 7.61 (2H, d, J ) 8 Hz, ArH), 7.42 (2H, d,
J ) 8 Hz, ArH), 4.27-4.14 (1H, m, CH), 2.39-2.29 (2H, m,
CH3 +CH2), 2.09-1.91 (1H, m, CH2), 1.91-1.73 (1H, m, CH2).
Anal. (C13H12NO5SBr) C,H,N.
N-(p-Ch lor op h en ylth ioca r bon yl)-L-glu ta m ic a cid (6j):
1H NMR (250 MHz, DMSO-d6) δ 12.4-12.5 (2H, b, COOH),
8.71 (1H, d, CONH), 7.49 (4H, s, ArH), 4.28-4.16 (1H, m, CH),
2.38-2.28 (2H, m, CH3 + CH2), 2.09-1.91 (1H, m, CH2), 1.91-
1.73 (1H, m, CH2). Anal. (C13H12NO5SCl) C,H,N.
buffer). CPG2 (38.13 mg/mL, 350 U/mg) was diluted in assay
buffer to make a stock solution of 28.66 µg/mL; 10 U/mL
concentrations shown in the text are final concentrations after
all the incubation solutions are combined. One unit of CPG2
activity corresponds to 1 µmol of methotrexate hydrolyzed
min-1 21
. All values for the initial velocity were calculated from
the progress curve using the computer-assisted spectropho-
tometer (molar extinction coefficient of methotrexate is 8300).
Hyperbolic saturation curves were obtained by measuring
CPG2 activity at varying methotrexate concentration and
different fixed 6d concentration. Methotrexate concentration
was varied between 3.077 and 70 µM, while 6d was at fixed
concentrations of 0.5 and 1.0 µM; 10 µL of stock CPG2 was
placed in a 1-mL cuvette, the reaction was initiated by adding
990 µL of methotrexate alone or methotrexate containing 6d ,
and the decrease in absorbance at 320 nm was recorded over
a time interval of 0-60 s. To obtain values for apparent Km
and apparent Vmax, the data was fitted by the least-squares
fit to the Michaelis-Menten equation using Fig.P (Biosoft)
computer program. The pattern of 6d inhibition with respect
to methotrexate was illustrated by using a double-reciprocal
(Lineweaver-Burk) plot of the data. Concentrations shown in
the text are final concentrations after all the incubation
solutions are combined.
Cytotoxicity Stu d ies. 1. Gen er a l Meth od for Cytotox-
icity of P oten tia l CP G2 In h ibitor s. To determine cytotox-
icity of the inhibitor, cells were incubated with graded con-
centration of inhibitor for 1 h at 37 °C. The cytotoxicity of each
potential inhibitor against LS174T cells was also determined
by exposing the cells to graded concentrations of particular
inhibitor (0.01-500 µM in medium) for 1 h at 37 °C. After each
treatment the cells were washed once with medium and
incubated for a further 6 days. Cell survival for all the
experiments was measured by the SRB assay.22
2. In Vitr o In h ibition of CP G2 Activa tion of P r od r u g
(P GP ) by 6d . To demonstrate in vitro inhibition of CPG2
turnover of prodrug (PGP), Figure 1, in the presence of 6d ,
LS174T cells were treated as follows: (a) To determine the
cytoxicity of parent drug (phenol mustard) produced in situ
by the action of the antibody-enzyme conjugate, the cells were
incubated with 100 µL of F(ab′)2-A5B7-CPG2 (3.32 µg/mL,
0.505 U/mL) at 37 °C for 1 h. The cells were then washed two
times in 200 µL of medium to remove excess conjugate which
had not bound to the cells, followed by incubation with 200
µL of graded concentration of prodrug (PGP) (0.01-500 µM
in medium) for 1 h at 37 °C. (b) To determine inhibition of
cytotoxicity in the presence of 6d , the cells were incubated with
100 µL of F(ab′)2-A5B7-CPG2 (3.32 µg/mL, 0.505 U/mL) for 1
h at 37 °C. The cells were then washed two times in 200 µL of
medium to remove unbound conjugate, followed by incubation
with graded concentration of PGP plus 25 or 50 µM 6d for 1 h
at 37 °C. (c) To determine cytotoxicity of the PGP alone, this
was carried out as described above for the potential inhibitors.
All of the synthesized compounds (Table 1) were then tested
to determine their CPG2 inhibitory properties using the Dixon
plot. The Dixon plot was chosen for the preliminary screen to
determine the apparent Ki.20 In addition, the cytotoxicity (IC50
)
of each compound was also determined.
Deter m in a tion of Ap p a r en t Ki. Assay solutions were set
up with varying inhibitor concentrations (0-3.0 µM) at dif-
ferent concentrations of methotrexate; 10 µL (0.1 U) of stock
enzyme solution was dispensed into 1-mL cuvettes, and
reaction was started by the addition of 990 µL of assay
solutions.
Enzyme solution (10 µL) was dispensed into 1-mL cuvettes,
and inhibitor solutions (990 µL) were added. The initial rate
of decrease of absorbance was measured spectrophotometri-
cally at 320 nm in a Shimadzu UV 1601 spectrophotometer.
The reciprocal of the initial velocity, 1/V, was plotted against
the micromolar (µM) concentrations of inhibitor to give the
20
linear Dixon plot from which apparent Ki was obtained, as
shown in Table 1.
Biologica l Stu d ies. 1. Cell Lin e. The human colon ad-
enocarcinoma cell line, LS174T, which expresses carcinoem-
bryonic antigen (CEA) was obtained from the European
Animal Cell Culture Collection.
2. En zym e a n d An tibod y-En zym e Con ju ga te. Bacterial
enzyme CPG2 and a conjugate of the F(ab′)2 fragment of the
mouse anti-CEA antibody, A5B7, and the antibody enzyme
conjugate (F(ab′)2-A5B7-CPG2) were kindly provided by Dr.
R. G. Melton, Division of Biotechnology, Centre for Applied
Microbiology and Research, Porton Down, Salisbury, Wilts,
SP4 OJ G, U.K.
The most promising of the compounds listed in Table 1 were
then subjected to screening against LS174T cell lines for their
capacity to abrogate in vitro cytotoxicity in the presence of
enzyme and the prodrug N-p-[bis(2-chloroethyl)amino]phe-
noxycarbonyl-L-glutamic acid (PGP), Figure 1.
Compound 6d was chosen due to its low toxicity and best
inhibition of CPG2 and was thus studied in depth, to determine
its mode of inhibition and its capacity to suppress enzyme
activity in vitro.
Ack n ow led gm en t. We are very grateful to Dr. G.
Rogers, Dr. R. G. Melton, Dr. R. Sherwood, and Prof.
R. J . Knox for their kind assistance. We also thank
Enzacta Ltd. for financial support.
Refer en ces
(1) (a) Melton, R. G.; Sherwood, R. F. Antibody-Enzyme conjugates
for therapy (Review). J . Natl. Cancer Inst. 1996, 88, 153-165.
(b) Melton, R. G.; Connors, T.; Knox, R. J . The use of Prodrugs
in Targeted Anticancer Therapies. STP Pharma Sci. 1999, in
press. (c) Enzyme-prodrug strategies for cancer therapy; Melton,
R. G., Knox, R. J ., Eds.; Plenum: London, 1999; in press.
(2) Bagshawe, K. D. Antibody directed enzymes revive anti-cancer
prodrugs concept. Br. J . Cancer 1987, 56, 531-532.
(3) (a) Antoniw, P. A study of the pharmacology of different prodrugs
in conjunction with targeted antibody-enzyme conjugate in
human cancer therapy. Ph.D. Thesis, University of London,
Charing Cross and Westminter Medical School, London, U.K.,
1991. (b) Sharma, S. K.; Bagshawe, K. D.; Burke, P. J .; Boden,
J . A.; Rogers, G. T.; Springer, C. J .; Melton, R. G.; Sherwood, R.
F. Galactosylated antibodies and antibody-enzyme conjugates
in antibody-directed enzyme prodrug therapy. Cancer 1994, 73,
1114-1120.
3. Kin etics. All kinetics were performed at 37 °C in 0.1 M
Tris-HCl, pH 7.3, buffer containing 0.1 mM zinc chloride (assay

956 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 6
Khan et al.
(4) Rogers, G. T.; Burke, P. J .; Sharma, S. K.; Koodie, R.; Boden, J .
A. Plasma clearance of an antibody-enzyme conjugate in ADEPT
by monoclonal anti-enzyme: its effect on prodrug activation in-
vivo. Br. J . Cancer 1995, 72(6), 1357-1363.
(5) Martin, J .; Stribbling, S. M.; Poon, G. K.; Begent, R. H. J .;
Napier, M.; Sharma, S. K.; Springer, C. J . Antibody-directed
enzyme prodrug therapy: pharmacokinetics and plasma levels
of prodrug and drug in a phase I clinical trial. Can. Chem.
Pharmacol. 1997, 40, 189-201.
(6) Hughes, P.; Sherwood, R. F.; Lowe, C. R. A single report for the
use of triazine dye, Procion red MX-8B, for the use of affinity
labeling of CPG2 at the active site has been published. Studies
on the nature of transition-metal-ion-mediated binding of tri-
azine dyes to enzymes. The interaction of procion red MX-8B
with carboxypeptidase G2. Eur. J . Biochem. 1984, 144, 135-
142.
(7) Lloyd, L. F.; Collyer, C. A.; Sherwood, R. F. Crystallization and
preliminary crystallographic analysis of carboxypeptidase G2
from Pseudomonas sp. strain RS-16. J . Mol. Biol. 1991, 220, 17-
18.
(8) Goldman, P.; Levy, C. C. Carboxypeptidase G: purification and
properties. Proc. Nat. Acad. Sci. U.S.A. 1967, 58, 1299.
(9) Blakey, D. C.; Valaccia, B. E.; East, S.; Wright, A. F.; Boyle, F.
T.; Springer, C. J .; Burke, P. J .; Melton, R. G.; Bagshawe, K. D.
Antitumor effects of an antibody-carboxypeptidase G2 conjugate
in combination with benzoic acid mustard prodrug. Cell Biophys.
1993, 22(1-3), 1-8.
(10) Blakey, D. C.; Davies, D. H.; Dowell, R. I.; East, S. I.; Burke, P.
J .; Sharma, S. K.; Springer, C. J .; Mauger, A. B.; Melton, R. G.
Anti-tumour effects of an antibody-carboxypeptidase G2 conju-
gate in combination with phenol mustard prodrugs. Br. J . Cancer
1995, 72(5), 1083-1088.
(11) Springer, C. J .; Dowel, R.; Burke, P. J .; Hadley, E.; Davis, D.
H.; Blakey, D. C.; Melton, R. G.; Niculescu-Duvaz, I. Optimisa-
tion of alkylating agent prodrugs derived from phenol and
aniline mustards: a new clinical candidate prodrug (ZD2767)
for antibody-directed enzyme prodrug therapy (ADEPT). J . Med.
Chem. 1995, 38, 5051-5065.
(12) DeAngelis, L. M.; Tong, W. P.; Lin, S.; Fleisher, M.; Bertino, J .
R. Carboxypeptidase G2 rescue after high-dose methotrexate. J .
Clin. Oncol. 1996, 14 (7), 2145-2149.
(13) Dowell, R. I.; Springer, C. J .; Davies, D. H.; Hadley, E. M.; Burke,
P. J .; Boyle, T. F.; Melton, R. G.; Connors, T. A.; Blakey, D. C.;
Mauger, A. B. New Mustard prodrugs for antibody-directed
enzyme prodrug therapy: Alternatives to the amide link. J . Med.
Chem, 1996, 39, 1100-1105.
(14) McCullough, J . L.; Chabner, B. A.; Bertino, J . R. Purification
and properties of carboxypeptidase G1. J . Biol. Chem. 1971, 246
(23), 7207-7213.
(15) Ruf, M.; Weis, K.; Brasack, I.; Vahrenkamp, H. Modelling
transition state analogues and enzyme-inhibitor complexes of
zinc-containing class II aldolases and metalloproteases. Inorg.
Chim. Acta 1996, 259, 271-281.
(16) Roesell, S.; Pauptit, R. A.; Tucker, A. D.; Melton, R. G.; Blow,
D. M.; Brick, P. Crystal structure of carboxypeptidase G2, a
bacterial enzyme with application in cancer therapy. Structure
1997, 5(3), 337-347.
(17) Antoniw, P.; Springer, C. J .; Bagshawe, K. D.; Searle, F.; Melton,
R. G.; Rogers, G. T.; Burke, P. J .; Sherwood, R. F. Disposition of
the prodrug 4-(bis(2-chloroethyl)amino)benzoyl-L-glutamic acid
and its active parent drug in mice. Br. J . Cancer 1990, 62, 909-
914.
(18) Baish, J . W.; Gazit, Y.; Berk, D. A.; Nozue, M.; Baxter, L. T.;
J ain, R. K.Role of tumor vascular architecture in nutrient and
drug delivery: an invasion percolation-based network model.
Microvasc. Res. 1996, 51(3), 327-346.
(19) A time course HPLC analysis was carried out by Prof. R. J . Knox
on a Waters HPLC system using C18 reversed-phase silica gel
analytical column, UV detector, and methanol as eluent, to
evaluate t1/2 for the most useful inhibitor.
(20) Dixon, M. The determination of Km and Ki. Biochem. J . 1972,
129(1), 197-202.
(21) Sherwood, R. F.; Melton, R. G.; Alwan, S. M.; Hughes, P.
Purification and properties of carboxypeptidase G2 from Pseudo-
monas sp'Å. Strain RS-16. Use of a novel triazine dye affinity
method. Eur. J . Biochem. 1985, 148(3), 447-453.
(22) Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.; McMahon, J .;
Vistica, D.; Warren, J . T.; Bokesch, H.; Kenney, S.; Boyed, M.
R. New colorimetric cytotoxicity assay for anti-cancer drug
screening. J . Natl. Cancer Inst. 1990, 82, 1107-1112.
J M990004I