Effect of plasma protein binding-on in vivo activity and brain penetration of glycine/NMDA receptor antagonists

Article abstract of DOI:10.1021/jm970417o

A major issue in designing drugs as antagonists at the glycine site of the NMDA receptor has been to achieve good in vivo activity. A series of 4- hydroxyquinolone glycine antagonists was found to be active in the DBA/2 mouse anticonvulsant assay, but improvements in in vitro affinity were not mirrored by corresponding increases in anticonvulsant activity. Here we show that binding of the compounds to plasma protein limits their brain penetration. Relative binding to the major plasma protein, albumin, was measured in two different ways: by a radioligand binding experiment or using an HPLC assay, for a wide structural range of glycine/NMDA site ligands. These measures of plasma protein binding correlate well (r = 0.84), and the HPLC assay has been used extensively to quantify plasma protein binding. For the 4-hydroxyquinolone series, binding to plasma protein correlates (r = 0.92) with log P (octanol/pH 7.4 buffer) over a range of log P values from 0 to 5. The anticonvulsant activity increases with in vitro affinity, but the slope of a plot of pED₅₀ versus pIC₅₀ is low (0.40); taking plasma protein binding into account in this plot increases the slope to 0.60. This shows that binding to albumin in plasma reduces the amount of compound free to diffuse across the blood-brain barrier. Further evidence comes from three other experiments: (a) Direct measurements of brain/blood ratios for three compounds (2, 16, 26) show the ratio decreases with increasing log P. (b) Warfarin, which compotes for albumin binding sites dose-dependently, decreased the ED₅₀ of 26 for protection against seizures induced by NMDLA. (c) Direct measurements of brain penetration using an in situ brain perfusion model in rat to measure the amount of drug crossing the blood-brain barrier showed that compounds 2, 26, and 32 penetrate the brain well in the absence of plasma protein, but this is greatly reduced when the drug is delivered in plasma. In the 4-hydroxyquinolones glycine site binding affinity increases with lipophilicity of the 3-substituent up to a maximum at a log P around 3, then does not improve further. When combined with increasing protein binding, this gives a parabolic relationship between predicted in vivo activity and log P, with a maximum log P value of 2.39. Finally, the plasma protein binding studies have been extended to other series of glycine site antagonists, and it is shown that for a given log P these have similar protein binding to the 4-hydroxyquinolones, except for compounds that are not acidic. The results have implications for the design of novel glycine site antagonists, and it is suggested that it is necessary to either keep log P low or pK(a) high to obtain good central nervous system activity.

Full text of DOI:10.1021/jm970417o

J . Med. Chem. 1997, 40, 4053-4068
4053
Effect of P la sm a P r otein Bin d in g on in Vivo Activity a n d Br a in P en etr a tion of
Glycin e/NMDA Recep tor An ta gon ists†
Michael Rowley,* J anusz J . Kulagowski, Alan P. Watt, Denise Rathbone, Graeme I. Stevenson,
Robert W. Carling, Raymond Baker, George R. Marshall, J ohn A. Kemp, Alan C. Foster, Sarah Grimwood,
Richard Hargreaves, Catherine Hurley, Kay L. Saywell, Mark D. Tricklebank, and Paul D. Leeson
Merck Sharp and Dohme Research Laboratories, Neuroscience Research Centre, Terlings Park, Eastwick Road,
Harlow, Essex CM20 2QR, United Kingdom
Received J une 25, 1997X
A major issue in designing drugs as antagonists at the glycine site of the NMDA receptor has
been to achieve good in vivo activity. A series of 4-hydroxyquinolone glycine antagonists was
found to be active in the DBA/2 mouse anticonvulsant assay, but improvements in in vitro
affinity were not mirrored by corresponding increases in anticonvulsant activity. Here we show
that binding of the compounds to plasma protein limits their brain penetration. Relative binding
to the major plasma protein, albumin, was measured in two different ways: by a radioligand
binding experiment or using an HPLC assay, for a wide structural range of glycine/NMDA
site ligands. These measures of plasma protein binding correlate well (r ) 0.84), and the HPLC
assay has been used extensively to quantify plasma protein binding. For the 4-hydroxyqui-
nolone series, binding to plasma protein correlates (r ) 0.92) with log P (octanol/pH 7.4 buffer)
over a range of log P values from 0 to 5. The anticonvulsant activity increases with in vitro
affinity, but the slope of a plot of pED50 versus pIC50 is low (0.40); taking plasma protein binding
into account in this plot increases the slope to 0.60. This shows that binding to albumin in
plasma reduces the amount of compound free to diffuse across the blood-brain barrier. Further
evidence comes from three other experiments: (a) Direct measurements of brain/blood ratios
for three compounds (2, 16, 26) show the ratio decreases with increasing log P. (b) Warfarin,
which competes for albumin binding sites dose-dependently, decreased the ED50 of 26 for
protection against seizures induced by NMDLA. (c) Direct measurements of brain penetration
using an in situ brain perfusion model in rat to measure the amount of drug crossing the blood-
brain barrier showed that compounds 2, 26, and 32 penetrate the brain well in the absence of
plasma protein, but this is greatly reduced when the drug is delivered in plasma. In the
4-hydroxyquinolones glycine site binding affinity increases with lipophilicity of the 3-substituent
up to a maximum at a log P around 3, then does not improve further. When combined with
increasing protein binding, this gives a parabolic relationship between predicted in vivo activity
and log P, with a maximum log P value of 2.39. Finally, the plasma protein binding studies
have been extended to other series of glycine site antagonists, and it is shown that for a given
log P these have similar protein binding to the 4-hydroxyquinolones, except for compounds
that are not acidic. The results have implications for the design of novel glycine site antagonists,
and it is suggested that it is necessary to either keep log P low or pKa high to obtain good
central nervous system activity.
In tr od u ction
min, is present in amounts around 40 g/L (or about 600
µM) in the blood.7 Albumin is responsible for the
binding of acidic molecules, while the globulin R1-acidic
glycoprotein is important for the binding of basic and
neutral compounds. There are believed to be a number
of different binding sites on albumin for different types
of compound, and association constants for natural
ligands vary from 103 L/mol for compounds such as
steroids to 108 L/mol for long-chain fatty acids. As-
sociation constants for drugs rarely exceed 106 L/mol.
The factors that affect the penetration of a drug from
the systemic circulation into the central nervous system
(CNS) are complicated. Several models for the physical
properties that affect a compound’s ability to cross the
blood-brain barrier have been proposed. The simplest,
and one which has wide currency in medicinal chemistry
folklore, is that there is an optimal octanol/water
partition (log P) for brain penetration. This is based
on observations1-3 that for neutral molecules a parabolic
relationship exists between central nervous system
activity and log P with a maximum at a log P value
around 2. Hansch1 discusses this optimal log P with
reference to increases in receptor binding with increased
lipophilicity and concomitant increases in metabolism
and simple, nonspecific binding by proteins.
Traditionally it has been felt that binding to plasma
protein can reduce brain penetration, because for com-
pounds that enter the central nervous system by a
process of passive diffusion it is only the free drug in
plasma that is important for the concentration gradient
across the blood-brain barrier. However, an examina-
tion8 of the relationship between percentage plasma
protein binding and the ratio of the concentrations in
CSF to that in plasma for a range of drugs showed that
the relationship is not simple. Generally speaking,
however, the compounds that were highly plasma
There is some disagreement in the literature on the
relevance of plasma protein binding to the clinical
efficacy of drugs.4-6 The major plasma protein, albu-
† Dedicated to Professor Y. Kishi on the occasion of his 60th birthday.
* Corresponding author, e-mail address: Michael_Rowley@Merck.com.
X Abstract published in Advance ACS Abstracts, November 15, 1997.
S0022-2623(97)00417-2 CCC: $14.00 © 1997 American Chemical Society

4054 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25
protein-bound did show poor CSF/plasma ratios.
Rowley et al.
Ch a r t 1. 4-Hydroxyquinolone Glycine/NMDA
A
Antagonists and Warfarin
study has been made of the effects of lipophilicity on
the brain uptake of benzodiazepines.9 An attempt was
made to fit the extraction of the drugs into the brain
with the concentration of free drug in plasma estimated
from in vitro measurements of binding to plasma
proteins; however, extraction into the brain was con-
siderably higher than that expected on the basis of the
in vitro measurements of the amount of free drug
available in the presence of plasma protein.
It has been suggested that an important property in
a compound for limiting its ability to cross the blood-
brain barrier is the ability to form hydrogen bonds.10
This can be calculated either by observation using
Stein’s11 figures for the number of hydrogen bonds a
particular functional group can form or by measurement
of the ∆log P parameter,12 defined as the difference
between the log P values for the octanol/water and
cyclohexane/water partitions. ∆log P is seen as a
measure of a compound’s ability to either donate or
accept hydrogen bonds and correlates negatively with
brain penetration. In an excellent study, Young et al.13
showed that for a structurally diverse group of com-
pounds designed as central H2 receptor antagonists
there is not a significant correlation between the
logarithm of the brain/blood concentration ratio and the
octanol/water partition coefficient, log Poct. However if
∆log P was used, then a significant correlation was
found with brain penetration. Young et al. make
passing comment at the end of their paper that the log
Poct term in ∆log P might account for protein binding
in peripheral blood, therefore limiting drug free to
penetrate the brain.
van der Waterbreemd14 has suggested that the blood/
brain concentration ratio for Young’s set of compounds
can be calculated from purely theoretical descriptors,
the molecular volume and surface area of the hydro-
philic portion of the molecule. Commenting on this
work, Ganellin15 found that when extrapolating either
Young’s or van der Waterbreemd’s correlations to his-
taminergic agonists, they both overestimated brain
penetration, with the latter behaving less well than the
former. The agonists used to test the correlations were
all smaller than the training set, and it is suggested
that part of the problem may lie in the lack of sufficient
experimental data relating to brain penetration, in
particular the extent of binding to plasma protein in
peripheral blood. Abraham16 has commented that
Young’s correlation is specific to his data set, and he
attempted to describe brain penetration more generally
in terms of a number of descriptor variables for com-
pounds, including excess molar refraction, polarizability,
volume, and hydrogen-bonding terms. While the ability
of a compound to bind to plasma protein in the blood
might well be associated with some of these variables,
this problem is not explicitly discussed in relation to
brain penetration.
(high CMC and C0), both of which do not penetrate the
blood-brain barrier, and an intermediate region where
they do. This model is predictive but is purely qualita-
tive and does little to help explain the factors underlying
brain penetration.
For some time we have been interested in antagonists
at the glycine site of the NMDA receptor for possible
use in stroke and schizophrenia.18 An early problem
was designing compounds that penetrated into the CNS.
While high binding affinity could be achieved in
kynurenic acid derivatives19 and tetrahydroquinolines,20
both these series contain carboxylic acids which, due to
the highly polar nature of the charged carboxylate, limit
their brain penetration. We set out to devise a carboxyl-
ic acid bioisostere that allowed brain penetration and
succeeded with the design of 4-hydroxy-2-quinolones,21,22
which for the first time gave consistent CNS activity in
a series of full antagonists at this receptor. The
prototype compound in this series, the methyl ester 2
(L-695,902; Chart 1), has moderate affinity (IC50 6.4 µM)
and activity in an anticonvulsant model in DBA/2 mouse
(ED50 52 µmol/kg). Manipulation of the substituent at
the 3-position of the quinolone, leading to 26 (L-701,
324; Chart 1), gave a 3000-fold increase in binding
affinity but only 20-fold improvement in in vivo activity
(IC50 2 nM, ED50 2.4 µmol/kg). In trying to understand
why this was the case, we considered the possibility that
these compounds bind to plasma protein in the systemic
circulation, therefore lowering the amount of drug free
in the aqueous phase of the blood to penetrate into the
brain. Compounds such as 26 resemble the structure
of warfarin (Chart 1), an anticoagulant that binds
tightly to a site on albumin defined as the warfarin
binding site.23 This paper discusses the results of
experiments performed to determine the extent of
plasma protein binding with 4-hydroxy-2-quinolones
and the effect of that binding on CNS-mediated in vivo
activity in an anticonvulsant screen. It then goes on to
investigate whether other structural classes of glycine
antagonists fit the trends found for the quinolone series.
Ch em istr y
A model to predict whether a compound can cross the
blood-brain barrier has been proposed based on the
measurement of its surface activity.17 In this model,
compounds are divided into those that are CNS-active
and those that are not. If one then plots a graph of the
critical micelle concentration (CMC) against the lowest
concentration for the onset of surface activity (C0), then
this divides into three regions: very hydrophobic com-
pounds (low CMC and C0), very hydrophilic compounds
Compounds 1-4, 6, 8, 9, 11, and 15 are described in
ref 21. 26, 31, 35, 39, and 48 are described in ref 22,
and full characterization is given in the Experimental
Section. Compounds 17, 29, 32, 33, and 59 were made
by the general route shown in Scheme 1. Thus, coupling
of 5-methylfuran-2-acetic acid24 to the methyl ester of
4-chloroanthranilic acid followed by cyclization with
potassium hexamethyldisilazide gave 33. Compounds

Effect of Plasma Protein Binding on in Vivo Activity
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25 4055
Sch em e 1a
the glycine site antagonist [3H]L-689,560 ([3H]-65). IC50
values (concentration of test compound required to
inhibit 50% of the specific binding) were evaluated via
construction of five-point inhibition curves. IC50 values
are given as the geometric mean of at least three
experiments. The maximum standard error calculated
from the pIC50 values was always less than 5% of the
mean. Functional antagonism was measured37 by the
ability of test compounds to cause a rightward shift in
the dose-response curve for NMDA-induced polariza-
tions in rat cortical slices (Kb values). Results are the
mean of at least three experiments, and errors are given
as standard deviations in parentheses after the result.
To test the effect of albumin on the functional antago-
nism, compounds were tested in the same way but
including 2% (w/v) bovine serum albumin (Sigma) in the
assay. The anticonvulsant activity of compounds was
measured by their ability to protect against audiogenic
seizure in DBA/2 mice when dosed intraperitoneally 30
min prior to seizure induction (either ED50 values or
quoted as number protected/number tested at a par-
ticular dose).38
a
Reagents: (i) methyl 4-chloroanthranilate, BOP-Cl, ClCH2-
CH2Cl, room temperature; (ii) KHMDS, THF, room temperature.
Sch em e 2a
P r otein Bin d in g Assa ys. The affinity of 26 for
human serum albumin (HSA) has been measured by the
equilibrium dialysis method39 and shown to have a Ka
of 5 × 105 L/mol, a high binding affinity by comparison
to many drug molecules. This means that when plasma
concentrations of 26 are well below the concentration
of albumin (i.e., ,600 µM), which is the case in all the
experiments described herein, then the drug is >99%
bound to plasma protein. However, to measure relative
binding to plasma protein for a large number of com-
pounds, more accessible screens were developed.
a
Reagents: (i) 2-methoxyphenylacetyl chloride, dichloroethane,
reflux, 50%; (ii) K2CO3, MeOH, 75%; (iii) PCC, CH2Cl2; (iv)
NaOMe, MeOH, room temperature, 68% (two steps); (v) BBr3,
CH2Cl2, 0 °C, 83%.
Sch em e 3a
Preliminary experiments were performed in the func-
tional assay, where apparent Kb values were measured
in the presence and absence of 2% (w/v) bovine serum
albumin. It was found that the Kb values for 2 and 26
were reduced from 23.0 ((1.6) and 0.028 ((0.006) µM,
respectively, when the assay was run under normal
conditions to 598 ((198) and 5.9 ((1.0) µM, respectively,
in the presence of bovine serum albumin. Therefore the
reduction in potency of 2 is 25-fold and that of 26 is 200-
fold when the plasma protein is present. This assay
provided the first indication that the response of recep-
tors to the 4-hydroxyquinolone glycine site antagonists
was affected by plasma protein, but it is not suitable
for screening large numbers of compounds due to the
time required for each experiment.
a
Reagents: (i) NH3, MeOH, 150 °C, 61%; (ii) (CF3CO)2O, Et3N,
THF, 0 °C; (iii) K2CO3, MeOH/H2O, 70 °C, 100%; (iv) phenylacetyl
chloride, dichloroethane, reflux, 44%; (v) NaH, DMF, 100 °C, 33%.
16,25 60 and 61,26 62,27 63 and 64,19 65,20 66,28 67 and
68,29 69,30 70,31 71 (MNQX),32 72 (ACEA 1021),33 73,34
and 7635 are described in the literature. Esters 5 and
7 and ketones 10 and 12-14 were made using general
procedures in ref 21.
The quinolone 74 was made from 2-amino-4-chlo-
robenzyl alcohol (77; Scheme 2). Coupling with the acid
chloride derived from 2-methoxyphenylacetic acid gave
the ester amide, in which the ester could be selectively
hydrolyzed with potassium carbonate in methanol to
give alcohol 78. Oxidation followed by cyclization to 79
and removal of the methyl group from the anisole gave
74. Conversion of the anthranilic ester 80 (Scheme 3)
to the nitrile 81 followed by coupling to phenylacetyl
chloride gave amide 82, which was cyclized with sodium
hydride to give the 4-aminoquinolone 75.
The binding affinity of a compound for the glycine site
of the NMDA receptor in the presence and absence of
0.2% (w/v) HSA was then investigated. Since com-
pounds under test bind to the NMDA receptor and
albumin, the affinity of a compound for the glycine site
of the NMDA receptor is reduced in the presence of HSA
in proportion to its binding to HSA. It is not possible
to use [3H]L-689,560 ([3H]-65) in such an assay, because
the radioligand itself binds to albumin (see Tables 1 and
2). However [3H]glycine does not bind to HSA, so its
affinity for the glycine site is not affected by HSA
(Figure 1a). It has been shown37 that affinities of
ligands for the glycine site are the same when measured
using either [3H]L-689,560 or [3H]glycine, and so the
latter radioligand can be used to measure the binding
of compounds in the presence and absence of HSA. The
results of one such experiment are illustrated in Figure
1b, for compound 26. As can be seen, the dose-response
Biology
Affinities of test compounds for the glycine site of the
NMDA receptor were determined36 by displacement of

4056 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25
Ta ble 1. 4-Hydroxyquinolones
Rowley et al.

Effect of Plasma Protein Binding on in Vivo Activity
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25 4057
Ta ble 1 (Continued)
a
b
Human serum albumin index, as measured by retention time on an HPLC column containing human albumin. Protein binding
index, the ratio of binding affinities for the displacement of [3H]glycine from rat cortical membranes in the presence and absence of 0.2%
human serum albumin. c ED50 for protection against audiogenic seizure in DBA/2 mouse or (number protected)/(number tested at dose).
ED50’s are quoted in mg/kg (plain text) and in µmol/kg (italics). IC50 for displacement of [3H]L-689,560 (65) from rat cortical membranes.
d
e log P either measured (in bold) or calculated47 from tables of π values. f Tested 10 min postdosing. Not tested.
g
curve is shifted to the right in the presence of HSA, with
the affinity of 26 reduced from 2 nM in the absence of
HSA to 194 nM in the presence of HSA. This allows
the calculation of a protein binding index (PBI) (eq 1),
defined as the ratio of the IC50 values of the test
compound for displacement of [3H]glycine in the pres-
ence and absence of 0.2% (w/v) human serum albumin.
sured for 18 compounds and are shown in Tables 1 and
2. While the binding assay to determine PBI values was
more efficient than measurement of an affinity constant
by equilibrium dialysis and quicker than measuring
functional potency both in the presence and in the
absence of albumin, it was still somewhat cumbersome
because it required the use of [3H]glycine to measure
affinities and therefore requires centrifugation rather
than filtration.
PBI ) IC50(with albumin)/IC50(without albumin)
The preferred assay that we developed used HPLC
to measure the relative binding of compounds to serum
albumin. This was performed by comparison of HPLC
(1)
The PBI for 26 is therefore 97. PBI values were mea-

4058 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25
Ta ble 2. Other Series of Glycine/NMDA Antagonists
Rowley et al.
g
a-e See corresponding footnotes in Table 1. f Not tested. This compound has activity at AMPA receptors which may contribute to its
h
i
j
anticonvulsant effect. One enantiomer. The other enantiomer has an HSAI of 23. Not tested due to insolubility. Not tested in this
work, but workers at Sumitomo26 report no activity at 100 mg/kg against NMDA-induced seizures. Reference 48.
k
HSAI(X) ) 100 × retention volume(X)/
retention volumes using a column with a stationary
retention volume(26)
phase of HSA immobilized on silica gel.40 Since the
major interaction governing retention on these phases
is with the protein and it is known that the immobilized
protein binds drugs in a fashion similar to that found
in free solution, the retention volume should reflect the
extent of binding to the protein.41 This method allows
a much higher throughput of compounds and was
therefore used more extensively than either the func-
tional or the binding assay. To allow a direct compari-
son between each compound, an HPLC human serum
albumin binding index (HSAI) was determined. This
represents the degree of drug-protein binding mea-
sured relative to an arbitrarily chosen standard com-
pound, 26. This figure is calculated as the ratio of the
retention volume of the compound to the retention
volume of 26, normalized to 100. The retention volume
is defined as
Hence 26 has a standardized HSAI of 100, and it follows
that compounds with HSAI > 100 are more highly
protein bound to albumin than 26 and conversely those
which are less bound will have HSAI < 100.
Cor r ela tion of P r otein Bin d in g Assa ys. Tables
1 and 2 show the compounds under discussion along
with their results in the binding and HPLC assays. The
compounds in Table 1 are 4-hydroxy-2-quinolones with
a carbon substituent at the 3-position. These com-
pounds are all acidic, with pKa values of around 5 or
below. Measured pKa values for 1, 26, and 32 are 4.4,
5.2, and 4.6, respectively. Table 2 shows a number of
other structural types. These include the N-linked
indole (59), tetramic acids (60, 61), 3-nitrodihydroqui-
nolone (62), kynurenic acid derivatives (63, 64), tet-
rahydroquinolines (65, 66), a new series of tyrosine-
based ligands (67, 68), indoles (69, 70), and quin-
oxalinediones (71, 72). The last four compounds in
Table 2 have considerably reduced acidity compared to
the other examples. These are a derivative (73) of the
quinoxalinediones from Sumitomo (pKa 8.6), phenol (74)
in which the acidic functionality of the 4-hydroxy-2-
retention volume ) (retention time -
system dead time) × flow rate
and the HSAI for compound X is

Effect of Plasma Protein Binding on in Vivo Activity
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25 4059
F igu r e 3. Correlation of plasma protein binding with log P
for the 4-hydroxyquinolones in Table 1.
pounds for which there are measured values for PBI.
Linear regression gives the fit:
log(PBI) ) 1.01((0.16) log(HSAI) + 0.06((0.22) (2)
n ) 18, r ) 0.84, F ) 40, s ) 0.39
Errors in parentheses are the standard deviations in
the coefficients.
F igu r e 1. Inhibition of [3H]glycine binding to rat cortex/
hippocampus membranes by (a) glycine and (b) L-701,324, in
the absence (O) and presence (b) of 0.2% human serum
albumin. Inhibition curves are from the same experiment
which was repeated several times with similar results.
These compounds cover the range of structural types,
including acidic and nonacidic molecules. It can be seen
that there is a highly significant correlation between
the prediction of binding to plasma protein from these
two measurements. Factors such as the different tem-
perature and slight solvent differences in the two assays
may explain some of the residual variance. A further
factor in the measurement of the PBI is solubility. For
compounds that have tight binding to plasma protein,
it is necessary to attain high concentrations to measure
their affinity in the presence of protein: achieving full
solubility may be a problem (the solubility of 26 in water
is less than 100 ng/mL), and therefore the PBI will, if
anything, be overestimated for these compounds. For
reasons of experimental ease and to overcome the
solubility problems, we have therefore used the HPLC
binding assay as a measure of a compound’s ability to
bind to plasma protein, in the knowledge that both the
measures are equivalent.
4-Hyd r oxyqu in olon es. It has been known for some
time7 that there is a correlation between affinity for
plasma protein and lipophilicity within a series of
compounds. This is also the case for 4-hydroxyquino-
lones: Figure 3 shows this correlation, as a plot of log-
(HSAI) versus log P (octanol/pH 7.4 buffer) for the
4-hydroxy-2-quinolones with carbon substituents in the
3-position (compounds 2-58, Table 1). There is a highly
significant correlation between log(HSAI) and log P,
with the line fitted by linear regression being
F igu r e 2. Correlation of the two measures of plasma protein
binding: protein binding index (PBI) and HPLC human serum
albumin index (HSAI).
quinolone has been moved to the pendent phenyl ring,
4-amino-2-quinolone (75) (pKa > 11) in which the
hydroxyl has been replaced with the less acidic amino
group, and finally the compound (76) in which the acidic
hydroxyl group of the 4-hydroxy-2-quinolone has been
blocked by cyclization, to produce a nonacidic compound.
Since the HPLC assay is performed under reverse-
phase conditions, where, for a nonpolar stationary phase
more lipophilic compounds have longer retention times,
it is important to establish that the HSAI is a measure
of a compound’s ability to bind to albumin, rather than
just a direct measure of its lipophilicity. Figure 2 shows
the correlation between the two measures of a com-
pound’s plasma protein binding, its PBI and HSAI.
Data are taken from both tables and include all com-
log(HSAI) ) 0.59((0.03) log P + 0.27((0.08) (3)
n ) 57, r ) 0.92, F ) 283, s ) 0.28
The obvious outlier in this correlation is 8 with higher
binding to plasma protein than would be predicted just
on the basis of log P. This compound has a phenolic
hydroxyl group in a region of the molecule where most
of the others have just lipophilic groupssa possible
interpretation is that this hydroxyl group picks up a
specific interaction with the protein.

4060 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25
Rowley et al.
F igu r e 6. Correlation of in vivo activity with binding affinity.
F igu r e 4. Nonacidic compounds 73-77 compared with Figure
3.
it would be expected that an increase in binding affinity
at the glycine site would translate directly to an increase
in in vivo activity in the anticonvulsant assay, i.e.,
pED50 ) pIC50 + k
(4)
where k is a constant. Examination of the data for all
the hydroxyquinolones that have measured ED50 values
(Figure 6) shows that although the correlation between
in vivo and in vitro activity is very good, there is clearly
not a one-to-one improvement in pED50 with pIC50.
pED50 ) 0.40((0.03)pIC50 + 2.02((0.24) (5)
n ) 39, r ) 0.88, F ) 130, s ) 0.28
(For the correlation the ED50 values have been con-
verted to mol/kg, which affects the intercept of this line
but not the gradient.) It can be seen that the trend
referred to for 2 and 26 holds out across the range of
compounds in Table 1: there is an increase in anticon-
vulsant activity as binding affinity increases, but the
slope of the graph is well below 1. Thus, there is some
factor that is stopping the compounds from realizing the
full effect of the increase in binding affinity at the
glycine site.
It has already been shown that, along with glycine
site binding affinity, binding to plasma protein increases
with log P. This binding to albumin has the conse-
quence of reducing the glycine site binding affinity in
the experiments performed in the presence of albumin,
as seen in the PBI. In these experiments the reduction
in affinity is a direct result of there being less drug
available in the aqueous component for binding to
receptors (or a competition between albumin and the
glycine site for drug). Substituting eq 2 [using log(PBI)
) log(HSAI)] into eq 1, one obtains
F igu r e 5. Effect of log P on glycine site binding affinity.
This correlation between log(HSAI) and log P does not
hold for all compounds. Figure 4 shows the four less
acidic compounds (73-76, Table 2) in comparison to eq
3. As can be seen, these compounds all show consider-
ably less protein binding in relation to their log P values
than the hydroxyquinolones, confirming that measure-
ment of protein binding by HPLC is not in fact just a
measurement of log P. Multiple regression analysis
(vide infra) shows that the correlation between log-
(HSAI) and log P for nonacidic compounds is signifi-
cantly different from that of the 4-hydroxyquinolones.
Since we are using pH 7.4 buffer as the aqueous phase
for the log P determinations, the acidic compounds are
very largely ionized during the experiment. It is pos-
sible to calculate the log P values for the neutral species
if the pKa is known, but since ultimately we are
interested in the fate of compounds in vivo where they
will be present in the ionized form, this does not seem
appropriate.
pIC50(with albumin) ) pIC50(without albumin) -
log(HSAI) (6)
It was found in the medicinal chemistry program that
increasing the lipophilicity in the region of the 3-sub-
stituent of the 4-hydroxyquinolones (2-58) increased
binding affinity. As a general trend it was found that
affinity increases up until a log P value around 3 and
then levels off and decreases with large lipophilic
substituents, indicating some steric inhibition of bind-
ing. This is illustrated graphically in Figure 5, which
plots the binding affinity at the glycine site of the
NMDA receptor against log P. The maximum affinity
found by variation of the 3-position substituent was
about 1 nM.22 If all other factors remained equal, then
In the in vivo experiments, if the drug is getting into
the brain by passive diffusion, then it is driven in by a
concentration gradient across the blood-brain barrier.
The reduction in the concentration in the aqueous
portion of blood (free for diffusion) caused by binding
to plasma protein will be proportional to the reduction
in the concentration in the aqueous phase of the binding
experiment caused by binding to plasma protein (as-
suming no other factors keep the drug from the aqueous
phase). It is therefore reasonable to use the term for
pIC50(with albumin) from eq 6 in eq 4 giving

Effect of Plasma Protein Binding on in Vivo Activity
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25 4061
pED50 ) [pIC50 - log(HSAI)] + k
(7)
as an expectation for the relationship of in vivo and in
vitro properties, if the binding to plasma protein is an
important factor in reducing brain penetration.
The data are plotted in this way in Figure 7 and fitted
by
pED50 ) 0.60((0.05)[pIC50 - log(HSAI)] +
1.50((0.25) (8)
n ) 39, r ) 0.89, F ) 143, s ) 0.27
It can be seen that the fit of the data is improved when
protein binding is taken into account, and the slope of
the graph is closer to unity. It is the slope of the graph
that is important, since it is a correlation between in
vitro and in vivo measures of activity that is being
sought in order to explain the weaker than expected
anticonvulsant activity for these compounds.
F igu r e 7. Correlation of in vivo activity with binding affinity,
taking plasma protein binding into account.
Ta ble 3. Brain/Plasma Ratios Measured Directly by HPLC
no.
HSAI
dose, mg/kg iv
brain/plasma ratio, %
2
16
26
3
10
100
20
10
3
8
4
1-2
The gradient is still not equal to unity, which is what
might be expected if all the lack of in vivo activity were
to be explained by protein binding. It must be remem-
bered however that, while the values obtained for the
HSAI and those for the PBI are essentially the same
(eq 2), the glycine binding experiments are performed
in the presence of only 0.2% (w/v) plasma protein.
Higher amounts cannot be used in the assay because
the binding affinity of [3H]glycine is then reduced, but
physiological concentrations are 20-fold higher than
this. If it were possible to estimate the effect on binding
in the presence of more physiologically appropriate
amounts of protein, then the gradient of such a graph
would be even higher. Of course there are probably
other factors affecting brain penetration and in vivo
activity in these compounds, including partitioning into
lipid both systemically and in the brain and possibly
drug metabolism, but eq 8 provides good evidence that
plasma protein binding plays a major role in restricting
in vivo activity in 4-hydroxyquinolones.
Further evidence for the role of plasma protein
binding in determining brain penetration and in vivo
activity has been obtained from a number of other
experiments. The brain/blood concentration ratios have
been measured directly by HPLC for three compounds
(2, 16, 26); 12 rats were dosed iv with the compound
under investigation and then sacrificed in groups of 3
at 5, 15, 30, and 60 min. The brains were removed, and
the amount of drug in plasma and brain was measured
by HPLC. Allowance was made for residual blood in
the brain (using a figure of 4.8 µL/g), and figures were
corrected for this. Calculations of areas under the curve
for both brain and plasma were made, and the brain/
blood ratio was estimated from these. Results are in
Table 3 and are expressed as percentages. It can be
seen that qualitatively the results are the same as those
found for the anticonvulsant assay, with the brain
penetration decreasing with increasing lipophilicity of
the compound, so that for 2 the figure is 8%, but this
falls to 4% for 16 and to around 1-2% for the highly
lipophilic 26. Since it is known that protein binding
increases with increasing lipophilicity, one suggestion
from this experiment is that increasing binding to
plasma protein reduces brain penetration.
Ta ble 4. Anticonvulsant Effecta of 26 in the Presence and
Absence of Warfarin
dose of warfarin,b
mg/kg iv
ED50 of 26,c
mg/kg iv
limits,d
mg/kg
0
60
100
3.6
1.9
0.64
2.1-5.4
1.3-2.5
0.48-0.77
a
To prevent seizure induced by NMDLA (500 mg/kg sc) in Swiss
Webster mice. Given 15 min prior to 26. c Given 15 min prior to
NMDLA. 95% confidence limits.
b
d
26 by testing its anticonvulsant profile in the presence
of warfarin, which is known to bind to albumin.23 The
ability of 26 to prevent seizures in Swiss Webster mice
induced by N-methyl-D,L-aspartate42 was tested on its
own and when warfarin was given 15 min before 26.
Warfarin (100 mg/kg iv) on its own does not protect
against seizure (0/8 animals protected). Results of these
experiments are presented in Table 4. It can be seen
that warfarin dose-dependently reduces the amount of
26 needed for protection against seizure in the mice,
with 100 mg/kg warfarin giving some 6-fold reduction
in the ED50. This indicates that when 26 is displaced
from plasma protein by a competitive binding drug,
more 26 is then free in plasma to penetrate into the
brain.
Takasato et al. have developed43 a method for directly
measuring cerebrovascular penetration in the rat, and
we have used this method, with modifications described
by Gratton,44 to investigate the brain penetration of
compounds 2, 26, and 32. In this experiment drug is
introduced directly into the carotid artery of rats in the
absence of blood flow from the heart and fluid flow
maintained at a constant rate and pressure using a
syringe pump. The concentration of drug is kept
constant at 100 µg/mL, and the experimental time is
kept short, so the flow across the blood-brain barrier
is essentially unidirectional. Drug in saline was per-
fused for 10 s and in plasma for 60 s, so brain
concentrations are low but measurable. After this
period of time the amount of drug in the brain can be
analyzed and a clearance from the perfusate calculated.
The advantage of this method is that either rat plasma
or phosphate-buffered saline can be used to dissolve the
It is possible to competitively displace the binding of
a drug to plasma protein with another compound that
binds to the same site. This has been investigated for

4062 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25
Rowley et al.
Ta ble 5. Clearances into Cortex from Plasma and Saline
Using in Situ Brain Perfusion
saline,a mL
plasma,b mL
compd
s-1 g-1 × 104
SEMc
s-1 g-1 × 104
SEMc
2
26
32
410
340
450
590
2
8
38
43
32
2.8
0
4.4
0.7
0
0.5
23
diazepam
thiourea
300
a
Clearance from phosphate-buffered saline into cortex. Perfu-
b
sion was over 10 s. Clearance from rat plasma into cortex.
Perfusion was over 60 s. c Standard error of the mean (n ) 4 for
each experiment).
drug, giving values for clearance into the brain in the
presence and absence of plasma protein. The results
of these experiments are shown in Table 5 along with
control values for diazepam, which is highly brain
penetrant, and thiourea, a compound that is considered
poorly penetrant. Values found for clearance into
frontal cortex for diazepam and thiourea are similar to
those found by Gratton.44
F igu r e 8. In vitro measure of expected in vivo activity plotted
against log P for 3-aryl-4-hydroxyquinolones 16-58, fitted with
a quadratic equation.
[pIC50 - log(HSAI)] )
2.39 log P - 0.50(log P)2 + 3.4
It can be seen that clearance for 2, 26, 32, and
diazepam from saline into cortex is high and similar for
all compounds. This shows that in the absence of
plasma protein the glycine antagonists cross the blood-
brain barrier as effectively as diazepam. The picture
is very different when rat plasma is used as the vehicle.
Clearance of 2 and 32 falls by 2 orders of magnitude,
and clearance of 26 is too low to be measured in this
experiment. The value for diazepam is reduced by only
a factor of 2. In an equilibrium dialysis experiment
diazepam was found45 to have an association constant
of about 5 × 104 L/mol, some 10-fold lower than that of
26. The results for 2, 32, and 26 are in qualitative
agreement with their binding to plasma protein as
measured by the HSAI, where the value for 26 is 50-
fold higher than that for 2 and 10-fold higher than that
of 32. This experiment provides direct evidence that
binding to plasma protein modulates the brain penetra-
tion of 2, 26, and 32.
n ) 43, r ) 0.73
which has a maximum at log P ) 2.39. Using statistical
methods to fit pIC50 - log(HSAI) to nth order polynomi-
als in log P, it was found that a significant improvement
was made by introducing a term in (log P)2 to the
equation, but higher order terms did not explain any
more of the variance in the data. The fit to a second-
order polynomial in log P is therefore statistically
significant. We have made a very large number of
compounds with log P values in the region of 2-4 to
explore structure-activity relationships, and the best
three compounds in the anticonvulsant assay, 48 (ED50
0.5 mg/kg), 39 (ED50 0.75 mg/kg), and 26 (ED50 0.88 mg/
kg), are noted on Figure 8. These lie at the top of the
curve, and we have reached the conclusion that these
represent the optimal in vivo activity that can be
achieved in this series by manipulation of a lipophilic
3-substituent. Compound 26 went into preclinical
development as a result of the structure-activity rela-
tionship studies on 4-hydroxyquinolones.22
The value of log P that gives a maximum predicted
in vivo activity (log P ) 2.39) is remarkably similar to
the log P that Hansch found1 for the best anaesthetic
activity in volatile ethers (log P0 ) 2.35). He has also
summarized a number of findings for CNS-acting de-
pressants2 (including nonspecific and specific com-
pounds such as benzodiazepines) and found that optimal
log P’s normally lie in a range around 2 ( 0.5. Hansch
comments that for nonspecific hypnotics, optimum
hydrophobicity is related primarily with movement
through the body so that log P0 says little about intrinsic
potency. In the case of our glycine site antagonists two
factors contribute to the existence of an optimal log P
for in vivo activity: First, binding affinity increases with
log P up to a certain point but then does not continue
to rise (Figure 5). Second, there is a concomitant
increase in factors that decrease the availability of drug
to receptors in the CNS, in which binding to plasma
protein plays a major role. The requirements of a
receptor for lipophilic binding will be different for
different receptors, and the relationship of log P to
protein binding will vary from series to series, so the
optimal log P found for 4-hydroxyquinolones might not
be applicable to other series or other receptors.
It has therefore been shown that binding to plasma
protein is implicated in reducing in vivo activity for 26
and reducing penetration of the blood-brain barrier for
2, 26, and 32. Also decreasing brain concentrations as
measured directly by HPLC for 2, 16, and 26 correlate
with increasing lipophilicity, which has been shown to
improve binding to serum albumin. When combined
with these results, the correlations of anticonvulsant
activity with in vitro binding experiments provide
overwhelming evidence that plasma protein plays a
major role in limiting the availability of the 4-hydroxy-
quinolone series of glycine site antagonists for activity
in the central nervous system.
These correlations have been used in guiding the
medicinal chemistry effort on this project. The com-
pounds of most interest are the 3-aryl-4-hydroxyquino-
lones (16-58), as these consistently have the best
affinities and anticonvulsant activities. It has been
shown in eq 8 that there is a linear relationship between
pIC50 - log(HSAI) and ED50. Because pIC50 - log-
(HSAI) can be calculated for all compounds from in vitro
measurements whereas the ED50 is often not measur-
able, we have explored the dependence of this in vitro
measure on physicochemical properties. The relation-
ship between pIC50 - log(HSAI) and log P is shown in
Figure 8 and can be fitted by a quadratic equation:

Effect of Plasma Protein Binding on in Vivo Activity
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25 4063
to these, the indole 70 suffers from the presence of a
carboxylic acid and a high HSAI. This means that
despite nanomolar affinity it does not have activity in
the DBA/2 anticonvulsant assay.46
Quinoxalinediones 72 (ED50 3.3 mg/kg, IC50 2.8 nM,
HSAI 28) and 73 (IC50 24 nM, HSAI 24) illustrate the
point that other factors need to be taken into account
as well as protein binding. The first of these has affinity
similar to that of 26 with lower protein binding. Despite
this, the anticonvulsant potency is lower than that of
26. Compound 73 was not active at 100 mg/kg against
NMDA-induced seizures despite good affinity and a low
HSAI. Compound 72 has two amide units and com-
pound 73 has three, with each having high hydrogen-
bonding ability, and they might be expected to have poor
brain penetration based on that fact.10
F igu r e 9. Plasma protein binding and log P values for
compounds in Table 2. Acidic compounds (9) and neutral
compounds ([) are separated.
Oth er Ser ies of Glycin e/NMDA An ta gon ists. To
examine the interaction of other types of glycine site
antagonists with plasma protein, representative ex-
amples of different structural series are shown in Table
2. The results of comparing plasma protein binding
with log P are shown in Figure 9, with the points divided
into two sets: those for acidic compounds and those for
nonacidic compounds (which have already been seen in
Figure 4). The line shown on the graph is that fitted
for the 4-hydroxyquinolones. Multiple linear regression
with dummy variables was used to see if the relation-
ship between log(HSAI) and log P differed for 4-hydroxy-
quinolones (2-58), acids (59-72), and neutral com-
pounds (73-76). As a first step the equation was
calculated using all three compound types. The result-
ing equation was highly significant with 71% of the
variation in log(HSAI) predicted from log P (r ) 0.85, P
< 0.0001). The contrast between 4-hydroxyquinolones
and acids was then entered into this equation but was
not significant (p ) 0.48). However, entry of the
4-hydroxyquinolone/neutral contrast at this stage would
have produced a highly significant (P < 0.001) improve-
ment in the equation with a 10% increase in variance
predicted. These results show that the relationship
between log(HSAI) and log P does not differ for 4-hy-
droxyquinolones and acids (59-72) but does differ
significantly between these and the neutral compounds.
This analysis has important consequences for in vivo
activity. The compounds in Table 2 with the best
activity in the DBA/2 mouse model are the quinoxa-
linedione 72, the 3-nitrodihydroquinolone 62 (ED50 13.5
mg/kg, IC50 414 nM, HSAI 7), and the aminoquinolone
75 (ED50 16.2 mg/kg, IC50 6750 nM, HSAI 7). Com-
pounds 62 and 75 maintain reasonable anticonvulsant
activity despite their relative poor binding affinity,
presumably due to the low plasma protein binding.
Compound 62 keeps plasma protein binding low with
the absence of lipophilic groups reducing log P, and 75
has low protein binding despite a relatively high log P
because of its high pKa. A further compound of interest
is 66 (ED50 30 mg/kg, IC50 13.5 nM, HSAI 2). This is
one of the few tetrahydroquinolines that had measur-
able activity when dosed systemically in the anticon-
vulsant assay.28 In comparison to 4-hydroxyquinolones
the in vivo activity is poor despite reasonable affinity,
but it is the only compound containing a carboxylic acid
that has measurable activity in this assay. For 66 the
HSAI is kept very low by reducing log P with the
presence of a polar group (CH2NH2) in an area of the
molecule tolerated by the glycine site. The low HSAI
presumably removes the obstacle provided by protein
binding that is present in other compounds. In contrast
Con clu sion
It has been shown by direct methods that selected
4-hydroxyquinolone glycine antagonists readily cross the
blood-brain barrier in the absence of plasma protein
but that binding to serum albumin reduces brain
penetration and in vivo activity for these compounds.
This conclusion is supported by correlations of in vitro
measurements of protein and glycine site binding with
in vivo activity with a larger group of 42 4-hydroxyqui-
nolones. The binding to plasma protein increases
directly with log P. Because of steric limitations on the
size of the 3-substituents in this series, there is an upper
limit on binding affinity. The anticonvulsant activity
therefore has a parabolic relationship with log P, and a
maximum is reached with a log P value around 2.4.
The relationship between binding to plasma protein
and log P is not significantly different in every other
series of acidic glycine antagonists that have been
investigated, and in these compounds good in vivo/in
vitro ratios are only achieved when plasma protein
binding is kept low. This can be achieved in two ways.
The first, and less appealing, is to introduce polar
substituents into the molecules in areas tolerated by the
receptor. This has the disadvantage that polar, or
hydrogen-bonding, groups are intrinsically poor for
brain penetration. The second is to reduce the acidity
of the compounds. In the absence of acidic functionality,
binding to the major plasma protein, albumin, is much
reduced. Until recently there were no examples of
nonacidic glycine site ligands, but compounds 75 and
73 show that this can be achieved. Although neither is
yet optimal for in vivo activity, the target of a high-
affinity nonacidic glycine site antagonist is a highly
desirable objective.
The compounds analyzed in this paper have all been
glycine/NMDA site antagonists. However, the problem
of brain penetration is one that affects all central
nervous system-based medicinal chemistry programs,
and the relevance of plasma protein binding to brain
penetration should be considered on a wider scale. None
of the physicochemical models for brain penetration that
have been proposed specifically take protein binding into
account, but it should be a factor when trying to design
drugs that are to act within the blood-brain barrier.
Exp er im en ta l Section
Melting points were taken on a Reichert Thermovar ap-
paratus and are uncorrected. Proton NMR were measured on
Bruker AM 360 or AC 250 spectrometers, and chemical shifts
are reported in parts per million (δ) downfield from tetra-

4064 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25
Rowley et al.
methylsilane as internal standard; coupling constants are in
hertz. Mass spectra were recorded on a VG 70/250 spectrom-
eter. Merck Kieselgel (230-400 mesh) was used for column
chromatography. For reactions, dry solvents were used as
bought from Aldrich. Organic solutions were dried with
anhydrous magnesium sulfate. Elemental analyses were done
by Elemental Analysis Ltd., Okehampton, Devon, U.K., and
Butterworth Laboratories Ltd, Teddington, Middlesex, U.K.
7-Ch lor o-3-(3-th ien yl)-4-h yd r oxy-2(1H)-qu in olon e (17):
white needles; mp 341 °C dec (from DMF); 1H NMR (360 MHz,
DMSO-d6) δ 7.22 (1 H, dd, J ) 8.5, 2, H-6), 7.31 (1 H, s, H-8),
7.43 (1 H, d, J ) 4.9, thiophene-H-5), 7.49-7.51 (1 H, m,
thiophene-H-4), 7.76 (1 H, s, thiophene-H-2), 7.98 (1 H, d, J )
8.5, H-5), 10.80 (1 H, br s, NH), 11.55 (1 H, s, OH); MS (EI)
m/z ) 277 (M+). Anal. (C13H8ClNO2S) C, H, N.
7-Ch lor o-4-h yd r oxy-3-(4-m et h oxyp h en yl)-2(1H )-q u i-
n olon e (18): mp >350 °C (from DMF/H2O); 1H NMR (250
MHz, DMSO-d6) δ 3.81 (3 H, s, Me), 6.96 (2 H, d, J ) 7.2,
H-2′), 7.21 (1 H, dd, J ) 8.7, 2.1, H-6), 7.29 (2 H, d, J ) 7.2,
H-3′), 7.33 (1 H, d, J ) 2.1, H-8), 7.96 (1 H, d, J ) 8.7, H-5),
11.49 (1 H, s, NH); MS (EI) m/z ) 301 (M+) (found, m/z )
301.0478; C16H12ClNO3 requires m/z ) 301.0507).
The general methods for the syntheses of 4-hydroxy-2(1H)-
quinolone esters and ketones are described in ref 21.
2-(3-Th ien yl)eth yl 7-ch lor o-4-h yd r oxy-2(1H)-oxo-qu in -
olin e-3-ca r boxyla te (5): white needles (from DMF/H2O); mp
1
293-295 °C; H NMR (360 MHz, DMSO-d6) δ 3.04 (2 H, t, J
) 6.7, CH2Ar), 4.46 (2 H, t, J ) 6.7, CO2CH2), 7.19 (1 H, dd, J
) 1.5, 4.8, 4-thienyl), 7.25 (1 H, dd, J ) 1.9, 8.7, H-6), 7.30 (1
H, d, J ) 1.9, H-8), 7.41 (1 H, dd, J ) 1.5, 3, 2-thienyl), 7.45
(1 H, dd, J ) 3, 4.8, 5-thienyl), 7.94 (1 H, d, J ) 8.7, H-5),
11.59 (1 H, s), 13.32 (1 H, br s); MS (CI+, NH3) m/z ) 350 (M+
+ H). Anal. (C16H12ClNO4S) C, H, N.
7-Ch lor o-4-h yd r oxy-3-(3-m et h ylp h en yl)-2(1H )-q u in -
olon e (19): mp >350 °C (from DMF/H2O); 1H NMR (250 MHz,
DMSO-d6) δ 2.35 (3 H, s, Me), 7.11-7.17 (4 H, m, ArH), 7.26-
7.31 (2 H, m, ArH), 7.91 (1 H, d, J ) 8.6, H-5), 11.49 (1 H, s,
NH); MS (EI) m/z ) 285 (M+). Anal. (C16H12ClNO2) C, H, N.
E t h yl 7-iod o-4-h yd r oxy-2(1H )-oxo-q u in olin e-3-ca r b -
oxyla te (7): white needles (from DMF); mp >300 °C; 1H NMR
(360 MHz, DMSO-d6) δ 1.30 (3 H, t, J ) 7.1, CH3), 4.33 (2 H,
q, J ) 7.1, CH2), 7.53 (1 H, dd, J ) 1.5, 8.5, H-6), 7.65 (1 H, d,
J ) 1.5, H-8), 7.67 (1 H, d, J ) 8.5, H-5), 11.45 (1 H, br s); MS
(EI) m/z ) 359 (M+). Anal. (C12H10INO4) C, H, N.
7-Ch lor o-4-h yd r oxy-3-(4-m et h ylp h en yl)-2(1H )-q u in -
olon e (20): mp >350 °C (from DMF); 1H NMR (360 MHz,
DMSO-d6) δ 2.34 (3 H, s, Me), 7.2-7.25 (5 H, m, ArH), 7.30 (1
H, d, J ) 2.0, H-8), 7.91 (1 H, d, J ) 8.6, H-5), 11.49 (1 H, s,
NH); MS (EI) m/z ) 285 (M+). Anal. (C16H12ClNO2‚0.25H2O)
C, H, N.
3-(2-Th ien ylca r b on yl)-7-ch lor o-4-h yd r oxy-2(1H )-q u i-
n olon e (10): yellow needles (from DMF); mp 314-315 °C; 1H
NMR (360 MHz, DMSO-d6) δ 7.18 (1 H, dd, J ) 4, 5, H-4′),
7.31 (1 H, dd, J ) 2, 8.5, H-6), 7.34 (1 H, d, J ) 2, H-8), 7.71
(1 H, 1 H, dd, J ) 1, 4, H-3′), 7.52 (1 H, d, J ) 8.5, H-5), 8.02
(1 H, dd, J ) 1, 5, H-5′), 11.6 (1 H, br s); MS (EI) m/z ) 305
(M+). Anal. (C14H8ClNO3S) C, H, N.
3-(Ben zo[b]t h ien -3-ylca r b on yl)-7-ch lor o-4-h yd r oxy-
2(1H)-qu in olon e (12): white needles (from DMF); mp 306-
307 °C; 1H NMR (360 MHz, DMSO-d6) δ 7.21 (1 H, dd, J ) 1,
11, H-6), 7.36 (1 H, d, J ) 1, H-8), 7.45-7.5 (2 H, m, H-5′,
H-6′), 7.98 (1 H, d, J ) 11, H-5), 8.03 (1 H, d, J ) 9, H-4′),
8.45 (1 H, s, H-2′), 8.47 (1 H, d, J ) 9, H-7′), 10.8 (1 H, br s);
MS (EI) m/z ) 355 (M+). Anal. (C18H10ClNO3S) C, H, N.
7-Ch lor o-3-(4-flu or op h e n yl)-4-h yd r oxy-2(1H )-q u in -
olon e (21): mp 335-337 °C (from DMF/H2O); H NMR (360
MHz, DMSO-d6) δ 7.13-7.18 (3 H, m, ArH), 7.32 (1 H, s, H-8),
7.38-7.42 (2 H, m, ArH), 7.92-7.94 (1 H, d, J ) 8.6, H-5),
10.19 (1 H, br s, OH), 11.54 (1 H, s, NH); MS (EI) m/z ) 289
(M+). Anal. (C15H9ClFNO2) C, H, N.
7-C h lo r o -4-h y d r o x y -3-(3-m e t h o x y p h e n y l)-2(1 H )-
qu in olon e (22): mp >320 °C dec; 1H NMR (360 MHz, DMSO-
d6) δ 3.76 (3 H, s, Me), 6.88-6.94 (3 H, m, ArH), 7.20 (1 H, dd,
J ) 8.7, 2.1, H-6), 7.28-7.32 (2 H, m, ArH), 7.91 (1 H, d, J )
8.7, H-5), 10.24 (1 H, br s, OH), 11.48 (1 H, s, NH); MS (EI)
m/z ) 301 (M+). Anal. (C16H12ClNO3) C, H, N.
1
4-Hyd r oxy-3-p h en yl-2(1H)-q u in olon e (23): mp 206 °C
1
(from DMF); H NMR (360 MHz, DMSO-d6) δ 7.17 (1 H, t, J
3-[(5-Ch lor ot h ien -2-yl)ca r b on yl]-7-ch lor o-4-h yd r oxy-
) 7.1, H-7), 7.2-7.4 (6 H, m, ArH), 7.50 (1 H, td, J ) 7.1, 1,
H-6), 7.94 (1 dd, J ) 7.1, 1, H-5), 10.09 (1 H, br s, OH), 11.43
(1 H, s, NH); MS (EI) m/z ) 237 (M+). Anal. (C15H11NO2) C,
H, N.
2(1H)-qu in olon e (13): white needles (from DMF); mp 304-
1
305 °C; H NMR (360 MHz, DMSO-d6) δ 7.21 (1 H, d, J ) 1,
H-4′), 7.24 (1 H, dd, J ) 1, 10, H-6), 7.32 (1 H, d, J ) 1, H-8),
7.65 (1 H, d, J ) 1, H-3′), 7.94 (1 H, d, J ) 10, H-5); MS (EI)
m/z ) 340 (M+). Anal. (C14H7Cl2NO3S) C, H, N.
7-Ch lor o-4-h ydr oxy-3-[3-(ph en yleth yn yl)ph en yl]-2(1H)-
1
qu in olon e (24): mp 297-300 °C (from DMF/H2O); H NMR
3-[(5-Br om ot h ien -2-yl)ca r b on yl]-7-ch lor o-4-h yd r oxy-
2(1H)-qu in olon e (14): white needles (from DMF); mp 309-
310 °C; 1H NMR (360 MHz, DMSO-d6) δ 7.25 (1 H, dd, J ) 1,
10, H-6), 7.32 (1 H, d, J ) 2, H-4′), 7.33 (1 H, d, J ) 1, H-8),
7.60 (1 H, d, J ) 2, H-3′), 7.93 (1 H, d, J ) 10, H-5); MS (EI)
m/z ) 384 (M+). Anal. (C14H7BrClNO3S) C, H, N.
(360 MHz, DMSO-d6) δ 7.23 (1 H, dd, J ) 8.6, 2.0, H-6), 7.33
(1 H, d, J ) 2.0, H-8), 7.41-7.51 (9 H, m, ArH), 7.95 (1 H, d,
J ) 8.6, H-5), 10.50 (1 H, br s, OH), 11.58 (1 H, s, NH); MS
(EI) m/z ) 371 (M+). Anal. (C23H14ClNO2) C, H, N.
7-C h lo r o -5-e t h y l-4-h y d r o x y -3-p h e n y l-2(1H )-q u in -
1
olon e (25): mp 284-288 °C (from DMF/H2O); H NMR (360
7-Ch lor o-4-h yd r oxy-3-(5-m et h yl-2-fu r yl)-2(1H )-q u in -
olon e (32). Bis(2-oxo-3-oxazolidinyl)phosphinic chloride (21.68
g, 113.6 mmol) was added to a stirred solution of methyl
4-chloroanthranilate (10.5 g, 56.8 mmol), 5-methyl furan-2-
acetic acid (7.95 g, 56.8 mmol), and triethylamine (15.8 mL
113.6 mmol) in dry 1,2-dichloroethane (100 mL) at room
temperature. The resulting mixture was stirred at room
temperature for 18 h. The reaction mixture was then washed
with aqueous 1 N HCl (2 × 50 mL), the organic layers were
separated and dried, and the solvent was removed under
reduced pressure to afford a red oil. The oil was dissolved in
dry THF (100 mL) and treated with a solution of potassium
hexamethyldisilazide (57.3 mL of 0.5 M solution in toluene,
28.4 mmol) at room temperature for 1 h. The reaction was
quenched by the addition of methanol (50 mL) followed by
aqueous 1 N HCl which resulted in the formation of a white
precipitate. The solid was filtered, dried, and recrystallized
MHz, DMSO-d6) δ 1.20 (3 H, t, J ) 7.3, Me), 3.12 (2 H, q, J )
7.3, CH2), 6.99 (1 H, d, J ) 2.1, H-6 or H-8), 7.20 (1 H, d, J )
2.1, H-6 or H-8), 7.30-7.43 (5 H, m, ArH), 9.83 (1 H, br s, OH),
11.49 (1 H, s, NH); MS (EI) m/z ) 299 (M+). Anal. (C17H14
-
ClNO2) C, H, N.
7-Ch lor o-4-h yd r oxy-3-(3-p h en oxyp h en yl)-2(1H )-q u i-
n olon e (26): mp 303-306 °C (from DMF/H2O); 1H NMR (360
MHz, DMSO-d6) δ 6.96 (1 H, dd, J ) 7.4, 1.8, H-4′), 7.01 (1 H,
s, H-2′), 7.08-7.14 (4 H, m, ArH), 7.20 (1 H, dd, J ) 8.8, 2.0,
H-6), 7.31 (1 H, d, J ) 2.0, H-8), 7.36-7.43 (3 H, m, ArH),
7.92 (1 H, d, J ) 8.8, H-5), 10.40 (1 H, br s, OH), 11.53 (1 H,
s, NH); MS (EI) m/z ) 363 (M+). Anal. (C21H14ClNO3) C, H,
N.
7-Ch lor o-4-h yd r oxy-3-(4-p h en oxyp h en yl)-2(1H )-q u i-
n olon e (27): mp 274-276 °C (from DMF/H2O); 1H NMR (360
MHz, DMSO-d6) δ 7.01-7.08 (4 H, m, ArH), 7.15 (1 H, t, J )
7.5, H-4′′), 7.21 (1 H, dd, J ) 8.7, 2.1, H-6), 7.32 (1 H, d, J )
2.1, H-8), 7.37-7.44 (4 H, m, ArH), 7.93 (1 H, d, J ) 8.7, H-5),
10.29 (1 H, br s, OH), 11.55 (1 H, s, NH); MS (EI) m/z ) 363
(M+). Anal. (C21H14ClNO3‚0.125H2O) C, H, N.
1
from DMF to give 32 as white needles: mp 280-282 °C; H
NMR (360 MHz, DMSO-d6) δ 2.39 (3 H, s, Me), 6.19 (1 H, d, J
) 1, furan H-3), 6.66 (1 H, d, J ) 1, furan H-4), 7.21 (1 H, dd,
J ) 10, 1, H-6), 7.31 (1 H, d, J ) 1, H-8), 7.91 (1 H, d, J ) 10,
H-5), 10.3 (1 H, br s, NH), 11.61 (1 H, s, OH); MS (EI) m/z )
275 (M+). Anal. (C14H10ClNO3) C, H, N.
3-[3-(Ben zyloxy)p h en yl]-7-ch lor o-4-h yd r oxy-2(1H)-qu i-
n olon e (28): mp 293-295 °C (from DMF/H2O); 1H NMR (360
MHz, DMSO-d6) δ 5.10 (2 H, s, CH2), 6.93-7.08 (3 H, m, ArH),
7.21 (1 H, dd, J ) 8.6, 2.0, H-6), 7.19-7.47 (7 H, m, ArH),
The following compounds were prepared in a similar fash-
ion.

Effect of Plasma Protein Binding on in Vivo Activity
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25 4065
7.92 (1 H, d, J ) 8.6, H-5), 10.26 (1 H, br s, OH), 11.52 (1 H,
s, NH); MS (EI) m/z ) 377 (M+). Anal. (C22H16ClNO3‚0.25H2O)
C, H, N.
NMR (360 MHz, DMSO-d6) δ 2.22 (3 H, s, Me), 6.84 (1 H, dd,
J ) 8.0, 1.7, ArH), 6.89 (1 H, s, ArH), 6.96 (1 H, d, J ) 8.0,
Ar-H), 7.05-7.10 (2 H, m, ArH), 7.18-7.22 (2 H, m, ArH),
7.30-7.36 (3 H, m, ArH), 7.92 (1 H, d, J ) 8.6, H-5), 11.49 (1
H, s, NH); MS (EI) m/z ) 377 (M+). Anal. (C22H16ClNO3‚
0.15H2O) C, H, N.
7-Ch lor o-4-h yd r oxy-3-(1-m eth yl-2-p yr r olyl)-2(1H)-qu i-
n olon e (29): white needles; mp >340 °C (from DMF); 1H NMR
(360 MHz, DMSO-d6) δ 3.41 (3 H, s, Me), 5.90 (1 H, dd, J ) 3,
1.5, pyrrole-H), 6.05 (1 H, dd, J ) 3, 1.5, pyrrole-H), 6.84 (1
H, d, J ) 1.5, pyrrole-H-5), 7.19 (1 H, dd, J ) 8, 1.5, H-6),
7.31 (1 H, d, J ) 1, H-8), 7.89 (1 H, d, J ) 8, H-5), 10.18 (1 H,
br s, NH), 11.50 (1 H, s, OH); MS (EI) m/z ) 274 (M+). Anal.
(C14H11ClN2O2‚0.4H2O) C, H, N.
3-[3-(N-Ben zylam in o)ph en yl]-7-ch lor o-4-h ydr oxy-2(1H)-
1
qu in olon e (41): mp 259-260 °C (from DMF/H2O); H NMR
(360 MHz, DMSO-d6) δ 4.27 (2 H, d, J ) 4.5, CH2), 6.16 (1 H,
t, J ) 5.9, NHCH2), 6.50 (2 H, dd, J ) 7.7, 1.7, ArH), 6.62 (1
H, s, H-2′), 7.06 (1 H, t, J ) 7.7, ArH), 7.17-7.24 (2 H, m,
ArH), 7.28-7.39 (5 H, m, ArH), 7.88 (1 H, d, J ) 8.7, ArH),
10.00 (1 H, br s, OH), 11.44 (1 H, br s, NH); MS (EI) m/z )
377 (M+). Anal. (C22H17ClN2O2) C, H, N.
7-Ch lor o-5-eth yl-4-h ydr oxy-3-(3-ph en oxyph en yl)-2(1H)-
qu in olon e (30): mp 222-224 °C (from ethyl acetate/hexane);
1H NMR (360 MHz, DMSO-d6) δ 1.21 (3 H, t, J ) 7.4, Me),
3.14 (2 H, q, J ) 7.4, CH2), 6.97-6.99 (3 H, m, Ar-H), 7.08-
7.13 (4H, m, Ar-H), 7.22 (1 H, d, J ) 2.2, Ar-H), 7.36-7.44 (3
H, m, Ar-H), 9.8 (1 H, br s, OH), 11.35 (1 H, br s, NH); MS
(EI) m/z ) 391 (M+). Anal. (C23H18ClNO3) C, H, N.
7-Ch lor o-4-h yd r oxy-3-(3-b e n zylp h e n yl)-2(1H )-q u in -
7-Ch lor o-4-h yd r oxy-3-[3-(2-m et h oxyb en zyl)p h en yl]-
2(1H)-qu in olon e (42): mp 265-270 °C (from DMF/acetone);
1H NMR (360 MHz, DMSO-d6) δ 3.79 (3 H, s, OMe), 3.92 (2
H, s, CH2), 6.86 (1 H, dt, J ) 7.5, 1, H-5′′), 6.97 (1 H, d, J )
7.9, H-3′′), 7.1-7.3 (8 H, m, ArH), 7.92 (1 H, d, J ) 8.7, H-5),
10.2 (1 H, br s, OH), 11.50 (1 H, s, NH); MS (EI) m/z ) 391
(M+). Anal. (C23H18ClNO3) C, H, N.
7-Ch lor o-4-h yd r oxy-3-[3-(3-fu r a n ylm e t h yl)p h e n yl]-
2(1H)-qu in olon e (43): mp >300 °C (from DMF/H2O); 1H NMR
(360 MHz, DMSO-d6) δ 3.76 (2 H, s, CH2), 6.38 (1 H, s, H-4′′),
7.15-7.25 (4 H, m, ArH), 7.30-7.35 (2 H, m, ArH), 7.50 (1 H,
s, H-2′′ or H-5′′), 7.56 (1 H, s, H-5′′ or H-2′′), 7.92 (1 H, d, J )
8.7, H-5), 10.2 (1 H, br s, OH), 11.50 (1 H, s, NH); MS (EI) m/z
) 351 (M+). Anal. (C20H14ClNO3) C, H, N.
7-Ch lor o-4-h yd r oxy-3-[3-(3-a cet yl-4-h yd r oxyb en zyl)-
p h en yl]-2(1H)-qu in olon e (44): mp 268-270 °C; 1H NMR
(360 MHz, DMSO-d6) δ 2.63 (3 H, s, Me), 3.95 (2 H, s, CH2),
6.89 (1 H, d, J ) 8.4, H-5′′), 7.18-7.22 (3 H, m, ArH), 7.26 (1
H, s, H-2′), 7.30-7.34 (2 H, m, ArH), 7.42 (1 H, dd, J ) 2.1,
8.4, H-6′′), 7.82 (1 H, d, J ) 2.1, H-2′′), 7.92 (1 H, d, J ) 8.6,
H-5), 10.2 (1 H, br s, OH), 11.52 (1 H, s, NH), 11.8 (1 H, br s,
OH); MS (CI-, NH3) m/z ) 419 (M-) (found, m/z ) 419.0897;
C24H18ClNO4 requires m/z ) 419.0924).
1
olon e (31): mp 304-306 °C (from DMF/H2O); H NMR (360
MHz, DMSO-d6) δ 3.96 (2 H, s, CH2), 7.1-7.4 (11 H, m, ArH),
7.92 (1 H, d, J ) 8.6, H-5), 10.2 (1 H, s, OH), 11.50 (1 H, s,
NH); MS (CI+, NH3) m/z ) 362 (M+ + H). Anal. (C22H16
ClNO2•0.5H2O) C, H, N.
-
7-Ch lor o-3-(2-fu r yl)-4-h yd r oxy-2(1H)-qu in olon e (33):
white needles; mp 246-250 °C (from DMF); 1H NMR (360
MHz, DMSO-d6) δ 6.62 (1 H, dd, J ) 2, 1, furan H-4), 6.98 (1
H, d, J ) 2, furan H-3), 7.21 (1 H, dd, J ) 10, 1.5, H-6), 7.30
(1 H, d, J ) 1.5, H-8), 7.79 (1 H, d, J ) 1, furan H-5), 8.01 (1
H, d, J ) 10, H-5), 11.0 (1 H, br s, NH), 11.63 (1 H, s, OH);
MS (EI) m/z ) 261 (M+). Anal. (C13H8ClNO3‚0.4H2O) C, H,
N.
7-Ch lor o-4-h yd r oxy-3-(3-b en zoylp h en yl)-2(1H )-q u in -
1
olon e (34): mp 294-295 °C (from DMF/H2O); H NMR (360
MHz, DMSO-d6) δ 7.31 (1 H, dd, J ) 8.7, 2.0, H-6), 7.32 (1 H,
d, J ) 2.0, H-8), 7.5-7.9 (9 H, m, ArH), 7.96 (1 H, d, J ) 8.7,
H-5), 10.6 (1 H, s, OH), 11.6 (1 H, s, NH); MS (EI) m/z ) 375
(M+). Anal. (C22H14ClNO3‚0.4H2O) C, H, N.
3-[3-[[2-[(Dim et h yla m in o)m et h yl]-3-t h ien yl]oxy]p h e-
7-Ch lor o-4-h yd r oxy-3-[3-(4-m et h oxyb en zyl)p h en yl]-
2(1H)-qu in olon e (35): mp 282-284 °C (from DMF); 1H NMR
(360 MHz, DMSO-d6) δ 3.71 (3 H, s, OMe), 3.90 (2 H, s, CH2),
6.83 (2 H, d, J ) 7, H-3′′), 7.1-7.3 (8 H, m, ArH), 7.91 (1 H, d,
J ) 8.6, H-5), 10.23 (1 H, br s, OH), 11.50 (1 H, s, NH); MS
(EI) m/z ) 391 (M+). Anal. (C23H18ClNO3) C, H, N.
n yl]-7-ch lor o-4-h yd r oxy-2(1H)-qu in olon e (45): hydrochlo-
i
1
ride salt, mp 216-219 °C (from PrOH); H NMR (360 MHz,
DMSO-d6) δ 2.75 (6 H, s, Me), 4.41 (2 H, s, CH2), 6.91 (1 H, d,
J ) 5.5, thiophene H-4), 7.0-7.05 (2 H, m, ArH), 7.18 (1 H, d,
J ) 8, ArH), 7.22 (1 H, dd, J ) 2.0, 8.7, H-6), 7.34 (1 H, d, J
) 2.0, H-8), 7.42 (1 H, t, J ) 8, phenyl H-5), 7.75 (1 H, d, J )
5.5, thiophene H-5), 7.98 (1 H, d, J ) 8.7, H-5), 10.4 (1 H, br
s, OH), 11.6 (1 H, s, NH); MS (EI) m/z ) 426 (M+). Anal.
(C22H20Cl2N2O3S) C, H, N.
7-C h lo r o -4-h y d r o x y -3-(3′-b ip h e n y ly l)-2(1H )-q u in -
1
olon e (36): mp >345 °C dec (from DMF/H2O); H NMR (360
MHz, DMSO-d6) δ 7.22 (1 H, dd, J ) 8.6, 1.9, H-6), 7.34-7.38
(3 H, m, ArH), 7.45-7.52 (3 H, m, ArH), 7.60-7.68 (4 H, m,
ArH), 7.96 (1 H, d, J ) 8.6, H-5), 10.39 (1 H, br s, OH), 11.57
7-Chloro-4-hydroxy-3-[3-[4-[(2-methoxyethoxy)methoxy]-
ben zyl]ph en yl]-2(1H)-qu in olon e (46): mp 239-241 °C (from
DMF/H2O); 1H NMR (360 MHz, DMSO-d6) δ 3.20 (2 H, s, Me),
3.42-3.45 (2 H, m, OCH2), 3.68-3.70 (2 H, m, OCH2), 3.90 (2
H, s, ArCH2), 5.19 (2 H, s, OCH2O), 6.94 (2 H, d, J ) 8.5, H-3′′),
7.1-7.2 (6 H, m, ArH), 7.28-7.32 (2 H, m, ArH), 7.91 (1 H, d,
J ) 8.6, H-5), 10.2 (1 H, br s, OH), 11.51 (1 H, s, NH); MS (EI)
m/z ) 465 (M+). Anal. (C26H24ClNO5) C, H, N.
7-Ch lor o-4-h yd r oxy-3-[3-[(2-m et h ylp r op -2-en yl)oxy]-
p h en yl]-2(1H)- qu in olon e (47): mp 297-299 °C (from DMF/
H2O); 1H NMR (360 MHz, DMSO-d6) δ 1.23 (3 H, s, Me), 4.40
(2 H, s, OCH2), 4.96 (1 H, s, CCH2), 5.08 (1 H, s, CCH2), 6.70-
6.94 (3 H, m, ArH), 7.20 (1 H, dd, J ) 8.6, 1.9, H-6), 7.27-
7.31 (2 H, m, ArH, H-8), 7.92 (1 H, d, J ) 8.6, H-5), 11.50 (1
H, s, NH); MS m/z ) 342 (M+). Anal. (C19H16ClNO3) C, H, N.
(1 H, s, NH); MS (EI) m/z ) 347 (M+). Anal. (C21H14
ClNO2‚0.25H2O) C, H, N.
-
7-Ch lor o-4-h yd r oxy-3-[3-(4-m et h ylp h en oxy)p h en yl]-
2(1H)-qu in olon e (37): mp 296-298 °C (from DMF/H2O); 1H
NMR (360 MHz, DMSO-d6) δ 2.28 (3 H, s, Me), 6.90-6.99 (4
H, m, ArH), 7.11 (1 H, d, J ) 6.6, H-6), 7.17-7.22 (3 H, m,
ArH), 7.31 (1 H, s, H-8), 7.38 (1 H, t, J ) 7.9, ArH), 7.92 (1 H,
d, J ) 8.6, H-5), 11.53 (1 H, br s, NH); MS (EI) m/z ) 376
(M+). Anal. (C22H16ClNO3•0.15H2O) C, H, N.
3-[3-(Cyclopr opylm eth oxy)ph en yl]-7-ch lor o-4-h ydr oxy-
2(1H)-qu in olon e (38): mp 320-322 °C (from DMF); 1H NMR
(360 MHz, DMSO-d6) δ 0.30-0.35 (2 H, m, cyclopropyl CH2),
0.55-0.60 (2 H, m, cyclopropyl CH2), 1.25-1.30 (1 H, m,
cyclopropyl CH), 3.81 (2 H, d, J ) 7, OCH2), 6.86-6.91 (3 H,
m, ArH), 7.21 (1 H, dd, J ) 2.0, 8.7, H-6), 7.26-7.31 (2 H, m,
ArH), 7.91 (1 H, d, J ) 8.7 H-5), 10.2 (1 H, br s, OH), 11.5 (1
H, s, NH); MS (EI) m/z ) 397 (M+). Anal. (C19H16ClNO3) C,
H, N.
7-Ch lor o-4-h yd r oxy-3-[3-[4-(m eth oxym eth oxy)ben zyl]-
p h en yl]-2(1H)-qu in olon e (48): mp 261-264 °C (from DMF/
1
acetone/H2O); H NMR (360 MHz, DMSO-d6) δ 3.35 (3 H, s,
Me), 3.90 (2 H, s, ArCH2Ar), 5.13 (2 H, s, O CH2O), 6.93 (2 H,
d, J ) 8.6, H-3′′), 7.13-7.22 (6 H, m, ArH), 7.28-7.32 (2 H,
m, ArH), 7.92 (1 H, d, J ) 8.6, H-5), 10.2 (1 H, br s, OH), 11.51
(1 H, s, NH); MS (EI) m/z ) 421 (M+). Anal. (C24H20ClNO4)
C, H, N.
7-Ch lor o-4-h yd r oxy-3-[3-(3-th ien yloxy)p h en yl]-2(1H)-
1
qu in olon e (39): mp >315 °C dec (from DMF/H2O); H NMR
(360 MHz, DMSO-d6) δ 6.93 (3 H, m, ArH), 7.05 (1 H, br s,
ArH), 7.14 (1 H, d, J ) 8, ArH), 7.21 (1 H, dd, J ) 8.6, 2, H-6),
7.31 (1 H, d, J ) 2, H-8), 7.37 (1 H, t, J ) 8, H-5′), 7.56 (1 H,
dd, J ) 5, 3, ArH), 7.93 (1 H, d, J ) 8.6, H-5), 10.37 (1 H, br
s, OH), 11.52 (1 H, s, NH); MS (EI) m/z ) 369 (M+). Anal.
(C19H12ClNO3S) C, H, N.
7-Ch lor o-4-h ydr oxy-3-[3-[4-(m eth ylth io)ben zyl]ph en yl]-
2(1H)-qu in olon e (49): mp 296-297 °C (from DMF/H2O); 1H
NMR (360 MHz, DMSO-d6) δ 2.43 (3 H, s, Me), 3.92 (2 H, s,
CH2), 7.1-7.3 (10 H, m, ArH), 7.92 (1 H, d, J ) 8.7, H-5), 10.1
(1 H, br s, OH), 11.51 (1 H, s, NH); MS (EI) m/z ) (EI) 407
(M+).
7-Ch lor o-4-h yd r oxy-3-[3-(2-m et h ylp h en oxy)p h en yl]-
2(1H)-qu in olon e (40): mp 288-290 °C (from DMF/H2O); 1H

4066 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25
Rowley et al.
7-Ch lor o-4-h ydr oxy-3-[3-[4-[(m eth ylam in o)m eth yl]ben -
zyl]p h en yl]-2(1H)-qu in olon e (50): hydrochloride; mp 254-
256 °C (from MeOH/H2O); 1H NMR (360 MHz, CD3OD) δ 2.68
(3 H, s, Me), 4.06 (2 H, s, CH2), 4.12 (2 H, s, CH2), 7.21-7.24
(4 H, m, ArH), 7.35-7.40 (6 H, m, ArH), 7.95 (1 H, d, J ) 8.5,
H-5); MS (EI) m/z ) 404 (M+). Anal. (C24H21ClN2O2‚1.7HCl)
C, H, N.
(1 H, m, indole H-4), 7.95 (1 H, d, J ) 8.6, H-5), 11.2 (1 H, br
s, OH), 11.3 (1 H, s, NH); MS (EI) m/z ) 324 (M+). Anal.
(C18H13ClN2O2‚0.1H2O) C, H, N.
3-(2-Hyd r oxyp h en yl)-7-ch lor oqu in olin -2(1H)-on e (74).
Oxalyl chloride (2.95 mL, 4.31 g, 34 mmol) was added to
2-methoxyphenylacetic acid (4.7 g, 28.3 mmol) in CH2Cl2 (80
mL) containing DMF (3 drops) at room temperature. After
90 min the mixture was evaporated, the residue was dissolved
in dichloroethane (80 mL), and 2-amino-4-chlorobenzyl alcohol
(77; 2 g, 12.9 mmol) was added. The mixture was refluxed
for 90 min, cooled, filtered, washed with NaHCO3, evaporated,
and purified by flash chromatography, eluting with 1:1 EtOAc-
hexane to give 2-(3-methoxyphenylacetamido)-4-chlorobenzyl
2-methoxyphenylacetate (2.9 g, 50%) as a white solid. This
was dissolved in MeOH (200 mL) and stirred with K2CO3 (5
g) at room temperature for 16 h. The mixture was partitioned
between EtOAc and water; then the organic phase was washed
with brine, dried, evaporated, and purified by flash chroma-
tography, eluting with 1:1 EtOAc-hexane to give 2-(3-meth-
oxyphenylacetamido)-4-chlorobenzyl alcohol (78; 1.47 g, 75%)
as a white solid; 850 mg (2.8 mmol) of this solid was dissolved
in CH2Cl2 (80 mL) with PCC (1.38 g, 6.4 mmol) and stirred
for 3 h, before being diluted with EtOAc (70 mL) and filtered
through silica gel. The solution was evaporated to give 870
mg of 2-(3-methoxyphenylacetamido)-4-chlorobenzaldehyde as
a brown waxy solid which was used without further purifica-
tion. The aldehyde was dissolved in MeOH (30 mL) with
NaOMe (230 mg, 4.3 mmol) and stirred for 18 h. The mixture
was poured into a citric acid solution and extracted three times
with EtOAc, giving a white suspension of the product in the
organic solvent. This suspension was washed with water and
evaporated to give a yellow solid which was recrystallized from
MeOH to give 3-(2-methoxyphenyl)-7-chloroquinolin-2(1H)-one
(79; 560 mg, 68%) as fine white needles, mp 235-236 °C; 300
mg (1.05 mmol) of this solid was suspended in CH2Cl2 (30 mL),
and BBr3 (1 M in CH2Cl2, 3.2 mL) was added at 0 °C. After
30 min the reaction mixture was poured into water and
extracted with EtOAc (2 × 70 mL). The combined organic
layers were washed with water and brine, dried, and evapo-
rated, and the residue was recrystallized from MeOH to give
74 (238 mg, 83%) as yellow crystals: mp 246-248 °C; 1H NMR
(360 MHz, DMSO-d6) δ 6.87 (1 H, t, J ) 8, H-5′), 6.90 (1 H, d,
J ) 8, H-3′), 7.22 (1 H, dt, J ) 1.7, 8, H-4′), 7.26 (1 H, dd, J )
2.1, 8.4, H-6), 7.31 (1 H, dd, J ) 1.7, 8, H-6′), 7.39 (1 H, d, J
) 2.1, H-8), 7.77 (1 H, d, J ) 8.4, H-5), 8.03 (1 H, s, H-4), 9.56
(1 H, br s, OH), 12.1 (1 H, br s, NH); MS (EI) m/z ) 271 (M+).
Anal. (C15H10ClNO2) C, H, N.
7-Ch lor o-4-h ydr oxy-3-[3-[4-(N-m or ph olin oeth yl)ben zyl]-
p h en yl]-2(1H)-qu in olon e (51): hydrochloride salt; mp 273-
1
276 °C (from MeOH); H NMR (360 MHz, DMSO-d6) δ 2.94-
3.10 (6 H, m, NCH2CH2O, NCH2CH2Ar), 3.10-3.30 (2 H, m,
NCH2CH2Ar), 3.84-3.94 (4 H, m, NCH2CH2O), 3.97 (2 H, s,
ArCH2Ar), 7.14-7.33 (9 H, m, ArCH2Ar, H-6), 7.38 (1 H, d, J
) 2.0, H-8), 7.95 (1 H, d, J ) 8.6, H-5), 11.52 (1 H, s, NH); MS
(CI+) m/z ) 475 (M+ + H). Anal. (C28H27ClN2O3‚HCl) C, H,
N.
3-(3-Ca r b oxyp h en yl)-7-ch lor o-4-h yd r oxy-2(1H )-q u in -
olon e (52): mp >340 °C (from DMF); 1H NMR (360 MHz,
DMSO-d6) δ 7.23 (1 H, dd, J ) 8.7, 2.2, H-6), 7.34 (1 H, d, J )
2, H-8), 7.52 (1 H, app t, J ) 7.9, H-5′), 7.63 (1 H, dt, J ) 7.8,
1.5, H-6′), 7.89 (1 H, dt, J ) 7.8, 1.5, H-4′), 7.95-7.97 (2.25 H,
m, H-5, H-2′, DMF), 10.48 (1 H, br s, OH), 11.61 (1 H, s, NH),
12.88 (1 H, br s, COOH); MS (CI+, NH3) m/z ) 333 (M + NH4+).
Anal. (C16H10ClNO4‚0.25DMF) C, H, N.
3-(Dib en zo[b,e][1,4]d ioxin -3-yl)-7-ch lor o-4-h yd r oxy-
2(1H)-qu in olon e (53): mp >390 °C dec; 1H NMR (360 MHz,
DMSO-d6) δ 6.9-7.0 (7 H, m, ArH), 7.21 (1 H, dd, J ) 2.0, 8.7
H-6), 7.31 (1 H, d, J ) 2.0, H-8), 7.93 (1 H, d, J ) 8.7, H-5),
10.4 (1 H, br s, OH), 11.55 (1 H, s, NH); MS (CI+, NH3) m/z )
397 (M + NH4+). Anal. (C21H12ClNO4) C, H, N.
7-Ch lor o-4-h yd r oxy-3-[3-(1-n a p h t h ylm et h yl)p h en yl]-
i
2(1H)-qu in olon e (54): sodium salt; mp >310 °C (from PrOH);
1H NMR (360 MHz, DMSO-d6) δ 4.35 (2 H, s, CH2), 6.60 (1 H,
d, J ) 7.6, ArH), 6.82 (1 H, dd, J ) 8.4, 2, ArH), 6.99 (1 H, t,
J ) 7.6, ArH), 7.02 (1 H, d, J ) 2, ArH), 7.4-7.6 (5 H, m,
ArH), 7.8-7.9 (3 H, m, ArH), 7.90-7.95 (1 H, m, ArH), 8.10-
8.15 (1 H, m, ArH), 9.75 (1 H, s, NH); MS (CI+, NH3) m/z )
429 (M + NH4+). Anal. (C26H17ClNNaO2) C, H, N.
7-Ch lor o-4-h yd r oxy-3-[3-[(N-m eth yl-N-p h en yla m in o)-
m eth yl]p h en yl]-2(1H)-qu in olon e (55): sodium salt; mp
i
1
>300 °C (from PrOH); H NMR (360 MHz, DMSO-d6) δ 3.00
(3 H, s, Me), 4.47 (2 H, s, CH2), 6.57 (1 H, t, J ) 7.2, H-4′′),
6.74 (2 H, t, J ) 8, H-2′′), 6.83 (1 H, dd, J ) 2, 8.4, H-6), 7.0-
7.2 (3 H, m, ArH), 7.61 (1 H, d, J ) 7.8, ArH), 7.71 (1 H, d, J
) 2, ARH), 7.86 (1 H, d, J ) 8.4, H-5), 9.76 (1 H, s, NH); MS
(CI+, NH3) m/z ) 391 (M+ + H). Anal. (C24H16ClNO2‚0.5H2O)
C, H, N.
7-C h lo r o -4-h y d r o x y -3-(3-in d o l-1-y lp h e n y l)-2(1H )-
1
qu in olon e (56): mp 290 °C dec (from AcOH); H NMR (360
4-Am in o-7-ch lor o-3-p h en ylq u in olin -2(1H )-on e (75).
Methyl 2-amino-4-chlorobenzoate (80; 20 g, 107 mmol) was
heated at 150 °C in MeOH (300 mL) saturated with NH3 in
an autoclave for 3 days, then cooled, and evaporated to a brown
solid. This was washed with ether, suspended in 1 M NaOH,
and filtered to give 2-amino-4-chlorobenzamide (11.2 g, 61%).
Triethylamine (34 mL, 245 mmol) was added to a solution of
the amide (9.5 g, 56 mmol) at 0 °C in THF (200 mL) followed
by trifluoroacetic anhydride (20.5 mL, 145 mmol) in THF (50
mL). After 30 min water (200 mL) was added, and the mixture
was extracted with Et2O (2 × 200 mL). The Et2O layer was
dried and evaporated, and the residue was dissolved in
MeOH-water (1:1, 200 mL) containing K2CO3 (15 g). The
mixture was heated at 70 °C for 24 h, cooled, and extracted
with EtOAc (2 × 200 mL), and the extracts were washed with
water and brine, dried, and evaporated to give 2-amino-4-
chlorobenzonitrile (81; 8.5 g, 100%): 1H NMR (360 MHz,
CDCl3) δ 4.47 (2 H, br s, NH2), 6.72 (1 H, dd, J ) 2.6, 11.9,
H-5), 6.75 (1 H, d, J ) 2.6, H-3), 7.31 (1 H, d, J ) 11.9, H-6);
MS (CI+, NH3) m/z ) 152 (M+ + H). The nitrile (0.9 g, 5.9
mmol) and phenylacetyl chloride (0.78 mL, 5.9 mmol) were
refluxed in dichloroethane (40 mL) for 16 h, cooled, and
evaporated, and the residue was recrystallized from MeOH to
give the amide 82 (0.7 g, 44%). This was heated in DMF (20
mL) with NaH (0.23 g, 80% in oil, 8 mmol) at 100 °C for 3 h,
cooled, and poured into water; the solid was collected and
recrystallized from DMF/water to give 75 (0.23 g, 33%) as
MHz, DMSO-d6) δ 6.71 (1Η, d, J ) 3.1, indolyl H-2), 7.13 (1
H, t, J ) 7.3, ArH), 7.18-7.25 (2 H, m, ArH), 7.42 (1 H, s,
H-2′), 7.43 (1 H, d, J ) 7.5, ArH), 7.52 (1 H, d, J ) 7.0, ArH),
7.58-7.67 (4 H, m, ArH), 7.72 (1 H, d, J ) 8.1, ArH), 7.98 (1
H, d, J ) 8.7, H-5), 10.58 (1 H, br s, OH), 11.60, (1 H, s, NH);
MS (EI-) m/z ) 386 (M•-). Anal. (C23H15ClN2O2‚0.25H2O) C,
H, N.
7-Ch lor o-4-h yd r oxy-3-(3-in d en ylp h en yl)-2(1H )-q u in -
1
olon e (57): red solid; mp 258-260 °C (from EtOH); H NMR
(360 MHz, DMSO-d6) δ 3.56 (2 H, s, CH2), 6.73 (1 H, s,
CHCH2), 7.1-7.7 (10 H, m, ArH), 7.97 (1 H, d, J ) 8.7, H-5),
10.5 (1 H, br s, OH), 11.6 (1 H, s, NH); MS (CI+, NH3) m/z )
388 (M+ + H). Anal. (C24H16ClNO2•0.7H2O) C, H, N.
7-Ch lor o-4-h yd r oxy-3-[3-[(3-th ien ylm eth yl)a m in o]p h e-
n yl]-2(1H)-qu in olon e (58): mp 254-256 °C (from DMF/H2O);
1H NMR (360 MHz, DMSO-d6) δ 4.26 (2 H, s, CH2), 6.04 (1 H,
br s, NHCHAr), 6.51 (1 H, d, J ) 7.5, H-6′), 6.58 (1 H, d, J )
8.1, H-4′), 6.64 (1 H, s, H-2′), 7.10 (1 H, t, J ) 7.8, H-5′), 7.12
(1 H, dd, J ) 6, 0.8, thienyl H-5), 7.20 (1 H, dd, J ) 8.6, 1.9,
H-6), 7.30 (1 H, d, J ) 1.9, H-8), 7.37 (1 H, s, thienyl H-2),
7.47 (1 H, dd, J ) 4.9, 2.8, thienyl H-4), 7.89 (1 H, d, J ) 8.6,
H-5), 10.00 (1 H, br s, OH), 1.48 (1 H, s, NH); MS (CI+, NH3)
m/z ) 383 (M+ + H). Anal. (C20H15ClN2O2S) C, H, N.
3-(3-Met h ylin d ol-1-yl)-7-ch lor o-4-h yd r oxy-2(1H )-q u i-
n olon e (59): mp 290 °C dec; 1H NMR (360 MHz, DMSO-d6) δ
2.31 (3 H, s, Me), 6.92-6.96 (1 H, m, indole H-7), 7.03 (1 H, d,
J ) 0.9, indole H-2), 7.04-7.09 (2 H, m, indole H-5,6), 7.27 (1
H, dd, J ) 1.9, 8.6 H-6), 7.36 (1 H, d, J ) 1.9, H-8), 7.51-7.55
1
white crystals: mp 337-338 °C; H NMR (360 MHz, DMSO-

Effect of Plasma Protein Binding on in Vivo Activity
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25 4067
d6) δ 5.96 (2 H, br s, NH2), 7.14-8.04 (8 H, m, ArH), 11.1 (1
H, br s, NH); MS (EI) m/z ) 270 (M+). Anal. (C15H11ClN2O)
C, H, N.
perfusate was 100 µg/mL. Drug was perfused in saline for 10
s or in plasma for 60 s during the experiment. The experiment
was ended by decapitation and the brain removed for analysis.
The procedure was repeated for groups of 4 female rats with
drugs dissolved in phosphate-buffered saline or rat plasma.
[3H]Glycin e Bin d in g. Assays were performed in polypro-
pylene tubes containing 150 µg of P2 membrane protein (rat
cortex/hippocampus36), 50 nM [3H]glycine (35.1-51.3 Ci/mmol;
DuPont/NEN), and 50 mM Tris-acetate buffer (pH 7.0). Drugs
were dissolved in dimethyl sulfoxide (final assay concentration
of 1%) and were added to the membranes on ice before
initiation of the incubation by addition of the radioligand.
Nonspecific binding was determined in the presence of 1 mM
glycine. Inhibition curves to glycine site ligands were con-
structed in the presence or absence of 0.2% human serum
albumin (HSA). Tubes were incubated at 4 °C for 30 min
before centrifugation at 50000g for 5 min, to separate free
radioactivity from bound. The supernatant was discarded, and
the pellet was superficially washed with 2 × 2 mL of assay
buffer. Pellets containing bound radioactivity were solubilized
overnight in 0.5 mL of 2% sodium dodecyl sulfate, and 0.4 mL
of the resulting solution was added to 10 mL of scintillation
fluid, for quantification of radioactivity using a â-counter.
Inhibition curves were analyzed for a one-site model by the
equation % I ) % Imax/1 + (IC50/[L])nH, using Research System
1 software (Bolt, Beranek and Newman, Inc., Cambridge, MA).
The degree of binding is expressed as the protein binding index
(PBI) which is the (mean IC50 in the presence of HSA)/(mean
IC50 in the absence of HSA). Binding experiments were
repeated three times, with the maximum standard error
calculated from the pIC50 values always less than 5% of the
mean. PBI values were then calculated from the mean binding
figures.
Sta tistics. Statistical analyses were performed using
BMDP New System for Windows, version 1.1, on a Compaq
Deskpro 386/20e computer.
Ack n ow led gm en t. We would like to thank Bill
Schofield for help with statistical analysis, Desmond
O’Connor and Natasha Daly for contributing to HSAI
determinations, and Richard Herbert and Steve Thomas
for NMR and MS data.
Su p p or tin g In for m a tion Ava ila ble: Microanalyses on
novel compounds and details of the π values used to calculate
log P’s in Tables 1 and 2 (15 pages). Ordering information is
given on any current masthead page.
Refer en ces
(1) Glave, W. R.; Hansch, C. Relationship between Lipophilic
Character and Anesthetic Activity. J . Pharm. Sci. 1972, 61, 589-
591.
(2) (a) Hansch, C.; Bjorkroth, J . P.; Leo, A. J . Hydrophobicity and
Central Nervous System Agents: On the Principle of Minimal
Hydrophobicity in Drug Design. J . Pharm. Sci. 1987, 76, 663-
687. (b) Hansch, C.; Steward, A. R.; Anderson, S. M.; Bentley,
D. The Parabolic Dependence of Drug Action upon Lipophilic
Character as Revealed by a Study of Hypnotics. J . Med. Chem.
1968, 11, 1-11.
(3) Gupta, S. P. QSAR Studies on Drugs Acting at the Central
Nervous System. Chem. Rev. (Washington, D.C.) 1989, 89, 1765-
1800.
(4) Curry, S. H. Binding of Psychotropic Drugs to Plasma Protein
and Its Influence on Drug Distribution. Clin. Pharmacol. Psy-
chiatr. 1981, 213-223.
(5) du Souich, P.; Verges, J .; Erill, S. Plasma Protein Binding and
Pharmacological Response. Clin. Pharmacokinet. 1993, 24, 435-
440.
(6) Rolan, P. E. Plasma Protein Binding Displacement Interac-
tionssWhy Are They Still Regarded as Clinically Important?
Br. J . Clin. Pharmacol. 1994, 37, 125-128.
(7) Lindup, W. E. Plasma Protein Binding of DrugssSome Basic
and Clinical Aspects. Prog. Drug Metab. 1987, 10, 141-185.
(8) J ezequel, S. G. Central Nervous System Penetration of Drugs:
Importance of Physicochemical Properties. Prog. Drug Metab.
1992, 13, 141-178.
(9) J ones, D. R.; Hall, S. D.; J ackson, E. K.; Branch, R. A.; Wilkinson,
G. R. Brain Uptake of Benzodiazepines: Effect of Lipophilicity
and Plasma Protein Binding. J . Pharmacol. Exp. Ther. 1988,
245, 816-822.
HP LC HSA Bin d in g In d ex. The relative binding of
compounds to albumin was performed by comparison of HPLC
retention volumes using an HPLC column with a stationary
phase of HSA immobilized on silica gel.40 Since the major
interaction governing retention on these phases is with the
protein and it is known that the immobilized protein binds
drugs in a fashion similar to that found in free solution, the
retention volume should reflect the extent of binding to the
protein.32 Briefly, samples were dissolved at a concentration
of 1 mg mL-1 in DMSO, and 5 µL injections were made on to
an HSA column (30 mm × 4.6 mm i.d., 5 µm) (Shandon HPLC,
Cheshire, U.K.) with a mobile phase of 50 mM K2HPO4, 50
mM KH2PO4 adjusted to pH 7.4 (1 M NaOH) with 10% propan-
1-ol. All analyses were performed in duplicate at a flow rate
of 2 mL min-1 and at a column temperature of 37 °C.
Reproducibility was excellent, with retention times of repeat
injections typically being within 0.1 min of the original,
irrespective of elution time.
Effect of Wa r fa r in on th e An ticon vu lsa n t Activity of
26. Experiments were performed on male Swiss Webster mice
(20-30 g; Bantin and Kingman, Hull, U.K). Animals were
housed in groups of 5 under a 12-h light/dark cycle (lights on
at 7 a.m.). Groups of 8 mice were injected iv with warfarin
(dissolved in 1 M NaOH, adjusted to pH 7 with HCl) or vehicle.
After 15 min animals were injected iv with 26 (dissolved in
25%, w/v, â-cyclodextrin in water at pH 7) and then 15 min
later with 500 mg/kg N-methyl-D,L-aspartic acid (NMDLA)
subcutaneously. The latency to onset of tonic seizures was
noted over the next 30 min, and animals not convulsing within
this period were considered protected. ED50 values were
determined by Probit analysis.
In Situ Br a in P er fu sion Stu d ies in th e An esth etized
Ra t. These experiments were performed using the method of
Takasato43 with modifications described by Gratton.44 Rats
were anesthetized, and a cannula was placed in the right
common carotid artery, with the external carotid artery ligated
just below the occipital branch. A programmable syringe
pump was used to control the flow of perfusate from the
syringe and maintained a pressure of 120 mmHg (100 mmHg
rat blood pressure and 20 mmHg tip pressure) during the
perfusion. This perfusion pressure is typical of cerebral
perfusion pressure and produces perfusion flow rates typical
of cerebral blood flow in intact anesthetized rats.
(10) Chikale, E. G.; Ng, K. Y.; Burton, P. S.; Borchardt, R. T.
Hydrogen Bonding Potential as a Determinant of the in Vitro
and in Situ Blood-Brain Barrier Permeability of Peptides.
Pharm. Res. 1994, 11, 412-419.
(11) Stein, W. D. The Movement of Molecules Across Cell Membranes;
Academic Press: New York, 1967; pp 65-91.
(12) Seiler, P. Interconversion of Lipophilicities from Hydrocarbon/
Water Systems into the Octanol/Water System. Eur. J . Med.
Chem. 1974, 9, 473-479.
(13) Young, R. C.; Mitchell, R. C.; Brown, T. H.; Ganellin, C. R.;
Griffiths, R.; J ones, M.; Rana, K. K.; Saunders, D.; Smith, I. R.;
Sore, N. E.; Wilks, T. J . Development of a New Physicochemical
Model for Brain Penetration and Its Application to the Design
of Centrally Acting H2 Receptor Histamine Antagonists. J . Med.
Chem. 1988, 31, 656-671.
(14) van der Waterbreemd, H.; Kansy, M. Hydrogen-Bonding Capac-
ity and Brain Penetration. Chimia 1992, 46, 299-303.
(15) Calder, J . A. D.; Ganellin, C. R. Predicting the Brain-Penetrating
Capability of Histaminergic Compounds. Drug Des. Discovery
1994, 11, 259-268.
(16) (a) Chadha, H. S.; Abraham, M. H.; Mitchell, R. C. Physico-
chemical Analysis of the Factors Governing Distribution of
Solutes between Blood and Brain. Bioorg. Med. Chem. Lett. 1994,
4, 2511-2516. (b) Abraham, M. H.; Chadha, H. S.; Mitchell, R.
C. Hydrogen Bonding. 33. Factors that Influence the Distribution
of Solutes between Blood and Brain. J . Pharm. Sci. 1994, 83,
1257-1267.
(17) Seelig, A.; Gottslich, R.; Devant, R. M. A Method to Determine
the Ability of Drugs to Diffuse Through the Blood-Brain Barrier.
Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 68-72.
(18) Leeson, P. D.; Iversen, L. L. The Glycine Site on the NMDA
Receptor: Structure-Activity Relationships and Therapeutic
Potential. J . Med. Chem. 1994, 37, 4053-4067.
At the beginning of the experiment the ventricles of the
heart were removed, thus allowing the free flow of perfusate,
and the pump was started. The drug concentration in the

4068 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 25
Rowley et al.
(19) Leeson, P. D.; Baker, R.; Carling, R. W.; Curtis, N. R.; Moore,
K. W.; Williams, B. J .; Foster, A. C.; Donald, A. E.; Kemp, J . A.;
Marshall, G. R. Kynurenic Acid Derivatives. Structure-Activity
Relationships for Excitatory Amino Acid Antagonism and Iden-
tification of Potent and Selective Antagonists at the Glycine Site
on the N-Methyl-D-aspartate Receptor. J . Med. Chem. 1991, 34,
1243-1252.
(20) Leeson, P. D.; Carling, R. W.; Moore, K. W.; Moseley, A. M.;
Smith, J . D.; Stevenson, G.; Chan, T.; Baker, R.; Foster, A. C.;
Grimwood, S.; Kemp, J . A.; Marshall, G. R.; Hoogsteen, K.
4-Amido-2-carboxytetrahydroquinolines. Structure-Activity Re-
lationships for Antagonism at the Glycine Site of the NMDA
Receptor. J . Med. Chem. 1992, 35, 1954-1968.
(21) Rowley, M.; Leeson, P. D.; Stevenson, G. I.; Moseley, A. M.;
Stansfield, I.; Sanderson, I.; Robinson, L.; Baker, R.; Kemp, J .
A.; Marshall, G. R.; Foster, A. C.; Grimwood, S.; Tricklebank,
M. D.; Saywell, K. L. 3-Acyl-4-hydroxyquinolin-2(1H)-ones.
Systemically Active Anticonvulsants Acting by Antagonism at
the Glycine Site of the N-Methyl-D-Aspartate Receptor Complex.
J . Med. Chem. 1993, 36, 3386-3396.
(22) Kulagowski, J . J .; Baker, R.; Curtis, N. R.; Leeson, P. D.; Mawer,
I. M.; Moseley, A. M.; Ridgill, M. P.; Rowley, M.; Stansfield, I.;
Foster, A. C.; Grimwood, S.; Hill, R. G.; Kemp, J . A.; Marshall,
G. R.; Saywell, K. L.; Tricklebank, M. D. 3’-(Arylmethyl)- and
3’-(Aryloxy)-3-phenyl-4-hydroxyquinoline-2(1H)-ones: Orally Ac-
tive Antagonists of the Glycine Site on the NMDA Receptor. J .
Med. Chem. 1994, 37, 1402-1405.
(23) The association constant of warfarin for human serum albumin
is around 105 L/mol. Kragh-Hansen, U. Relations between High-
Affinity Binding Sites of Markers for Binding Regions on Human
Serum Albumin. Biochem. J . 1985, 225, 629-638.
(24) Breen, L.; Eastwood, F. W.; Ockman, T.; Rae, J . D.; Redwood,
A. M. Formation of Aryl Acetic Acids, Arylacetamides and
Aryldihydropyrroles from Aromatic Aldehydes. Aust. J . Chem.
1973, 26, 2221-2227.
(25) Kappe, T.; Karem, A. S.; Stadlbauer, W. Synthesis of Benzo-
halogenated-4-hydroxy-2(1H)-quinolones. J . Heterocycl. Chem.
1988, 25, 857-862.
(26) Mawer, I. M.; Kulagowski, J . J .; Leeson, P. D.; Grimwood, S.
Tetramic Acids as Novel Glycine Site Antagonists. Bioorg. Med.
Chem. Lett. 1995, 5, 2643-2648.
(27) Carling, R. W.; Leeson, P. D.; Moore, K. W.; Smith, J . D.; Moyes,
C. R.; Mawer, I. M.; Thomas, S.; Chan, T.; Baker, R.; Foster, A.
C.; Grimwood, S.; Kemp, J . A.; Marshall, G. R.; Tricklebank, M.
D.; Saywell, K. L. 3-Nitro-3,4-dihydro-2(1H)-quinolones. Exci-
tatory Amino Acid Antagonists Acting at the Glycine-Site NMDA
and (RS)-R-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic Acid
Receptors. J . Med. Chem. 1993, 36, 3397-3408.
(28) Carling, R. W.; Leeson, P. D.; Moseley, A. M.; Smith, J . D.;
Saywell, K.; Tricklebank, M. D.; Kemp, J . A.; Marshall, G. R.;
Foster, A. C.; Grimwood, S. Anticonvulsant Activity of Glycine-
Site NMDA Antagonists. 2. Trans 2-Carboxy-4-substituted tet-
rahydroquinolines. Bioorg. Med. Chem. Lett. 1993, 3, 65-70.
(29) Curtis, N. R.; Kulagowski, J . J .; Leeson, P. D.; Mawer, I. M.;
Ridgill, M. P.; Rowley, M.; Grimwood, S.; Marshall, G. R.
Synthesis and SAR of Diiodotyrosine-Derived Glycine-Site N-
Methyl-D-Aspartate Receptor Ligands. Bioorg. Med. Chem. Lett.
1996, 6, 1145-1150.
(30) (a) Salituro, F. G.; Harrison, B. L.; Baron, B. M.; Nyce, P. L.;
Stewart, K. T.; Kehne, J . H.; White, H. S.; McDonald, I. A. 3-(2-
Carboxyindol-3-yl)propionic Acid-Based Antagonists of the N-
Methyl-D-Aspartic Acid Receptor Associated Glycine Binding
Site. J . Med. Chem. 1992, 35, 1791-1799. (b) Rowley, M.;
Leeson, P. D.; Grimwood, S.; Foster, A.; Saywell, K. 2-Carboxy
Indolines and Indoles as Potential Glycine/NMDA Antagonists:
Effect of Five-Membered Ring Conformation on Affinity. Bioorg.
Med. Chem. Lett. 1992, 2, 1627-1630.
(31) Di Fabio, R.; Capelli, A. M.; Conti, N.; Cugola, A.; Donati, D.;
Feriani, A.; Gastaldi, P.; Gavirighi, G.; Hewkin, C. T.; Micheli,
F.; Missio, A.; Mugnaini, M.; Pecunioso, A.; Quaglia, A. M.; Ratti,
E.; Rossi, L.; Tedesco, G.; Trist, D. G.; Reggiani, A. Substituted
Indole-2-carboxylates as in Vivo Potent Antagonists Acting at
the Strychnine-Insensitive Glycine Binding Site. J . Med. Chem.
1997, 40, 841-850.
(32) (a) Horner, L.; Schwenk, U.; J unghanns, E. Derivate des
Chinoxalins als Isostere der Pteridine. J ustus Liebigs Ann.
Chem. 1953, 579, 212-234. (b) Randle, J . C. R.; Guet, T.;
Bobichon, C.; Moreau, C.; Curutchet, P.; Lambolez, B.; de
Carvalho, L. P.; Cordi, A.; Lepagnol, J . M. Quinoxaline Deriva-
tives: Structure-Activity Relationships and Physiological Im-
plications of Inhibition of N-Methyl-D-aspartate and Non-N-
Methyl-D-aspartate Receptor-Mediated Currents and Synaptic
Potentials. Mol. Pharmacol. 1992, 41, 337-345.
(33) Keana, J . F. W.; Kher, S. M.; Cai, S. X.; Dinsmore, C. M.; Glenn,
A. G.; Guastella, J .; Huang, J .-C.; Ilyin, V.; Lu, Y.; Mouser, P.
L.; Woodward, R. M.; Weber, E. Synthesis and Structure-
Activity Relationships of Substituted 1,4-Dihydroquinoxaline-
2,3-diones: Antagonists of N-Methyl-D-aspartate (NMDA) Re-
ceptor Glycine Sites and Non-NMDA Glutamate Receptors. J .
Med. Chem. 1995, 38, 4367-4379.
(34) Nagata, R.; Tanno, N.; Kodo, T.; Ae, N.; Yamaguchi, H.;
Nishimura, T.; Antoku, F.; Tatsuno, T.; Kato, T.; Tanaka, Y.;
Nakamura, M. Tricyclic Quinoxalinediones: 5,6-Dihydro-1H-
pyrrolo[1,2,3-de]quinoxaline-2,3-diones and 6,7-1H,5H-Pyrido-
[1,2,3-de]quinoxaline-2,3-diones as Potent Antagonists for the
Glycine Binding Site of the NMDA Receptor. J . Med. Chem.
1994, 37, 3956-3968.
(35) Kappe, T.; Brandner, A.; Stadlbauer, W. Synthesis of Fused
Benzofurans by Dehydration of Cyclic Phenyl-â-dicarbonyl
Compounds. Monatsch. Chem. 1987, 118, 1177-1184.
(36) Grimwood, S.; Moseley, A. M.; Carling, R. W.; Leeson, P. D.;
Foster, A. C. Characterisation of the binding of [3H]-L-689,560,
an antagonist for the glycine site of the N-methyl-D-aspartate
receptor, to rat brain membranes. Mol. Pharmacol. 1992, 41,
923-930.
(37) Kemp, J . A.; Foster, A. C.; Leeson, P. D.; Priestley, T.; Tridgett,
R.; Iversen, L. L.; Woodruff, G. N. 7-Chlorokynurenic Acid is a
Selective Antagonist at the Glycine Modulatory Site of the
N-Methyl-D-Aspartate Receptor Complex. Proc. Natl. Acad. Sci.
U.S.A. 1988, 85, 6547-6550.
(38) Tricklebank, M. D.; Singh, L.; Oles, R. J .; Preston, C.; Iversen,
S. D. The Behavioural Effects of MK-801: A Comparison with
Antagonists Acting Noncompetitively and Competitively at the
NMDA Receptor. Eur. J . Pharmacol. 1989, 167, 127-135.
(39) Bos, O. J . M.; Remijn, J . P. M.; Fischer, M. J . E.; Wilting, J .;
J anssen, L. H. M. Location and Characterisation of the Warfarin
Binding Site of Human Serum Albumin. A Comparative Study
of Two Large Fragments. Biochem. Pharmacol. 1988, 37, 3905-
3909. (b) Colangelo, P.; Chandler, M.; Blouin, R.; McNamara,
P. Stereoselective Binding of Propranolol in the Elderly. Br. J .
Clin. Pharmacol. 1989, 27, 519-522.
(40) (a) Domenici, E.; Bertucci, C.; Salvadori, P.; Felix, G.; Cahagne,
I.; Motellier, S.; Wainer, I. W. Synthesis and Chromatographic
Properties of an HPLC Chiral Stationery Phase Based upon
Human Serum Albumin. Chromatographia 1990, 29, 170-176.
(b) Domenici, E.; Bertucci, C.; Salvadori, P.; Motellier, S.;
Wainer, I. W. Immobilized Serum Albumin: Rapid HPLC Probe
of Stereoselective Protein-Binding Interactions. Chirality 1990,
2, 263-268. (c) Noctor, T. G.; Felix, G.; Wainer, I. W. Stereo-
chemical Resolution of Enantiomeric 2-Aryl Propionic Acid Non-
Steroidal Antiinflammatory Drugs on a Human Serum Albumin
Based High-Performance Liquid Chromatographic Chiral Sta-
tionary Phase. Chromatographia 1991, 31, 55-59. (d) Noctor,
T. G.; Wainer, I. W.; Hage, D. S. Allosteric and Competitive
Displacement of Drugs from Human Serum Albumin by Octanoic
Acid, as Revealed by High-Performance Liquid Chromatography,
on a Human Serum Albumin-Based Stationary Phase. J . Chro-
matogr. 1992, 577, 305-315.
(41) Noctor, T. G.; Diaz-Perez, M. J .; Wainer, I. W. Use of a Human
Serum Albumin-Based Stationary Phase for High-Performance
Liquid Chromatography as a Tool for the Rapid Determination
of Drug-Plasma Protein Binding. J . Pharm. Sci. 1993, 82, 675-
676.
(42) Tricklebank, M. D.; Bristow, L. J .; Hutson, P. H.; Leeson, P. D.;
Rowley, M.; Saywell, K.; Singh, L.; Tattersall, F. D.; Thorn, L.;
Williams, B. J . The Anticonvulsant and Behavioural Profile of
L-687,414, a Partial Agonist Acting at the Glycine Modulatory
Site on the N-Methyl-D-aspartate (NMDA) Receptor Complex.
Br. J . Pharmacol. 1994, 113, 729-736.
(43) Takasato, Y.; Rapoport, S. I.; Smith, Q. R. An in Situ Brain
Perfusion Technique To Study Cerebrovascular Transport in the
Rat. Am. J . Physiol. 1984, 247, H484-H493.
(44) Gratton, J . A.; Lightman, S. L.; Bradbury, M. W. Transport into
Retina Measured by Short Vascular Perfusion in the Rat. J .
Physiol. (London) 1993, 470, 651-663.
(45) Measured as part of this work. See also ref 9.
(46) In their paper31 Di Fabio et al. report that 70 has an ED50 of 6
mg/kg when dosed po against NMDA-induced convulsions in the
mouse. In our hands the compound is not active when dosed ip
against audiogenic seizure in DBA/2 mouse at 10 mg/kg ip.
(47) (a) Hansch, C.; Leo, A.; Unger, S. H.; Kim, K. H.; Nikaitani, D.;
Lien, E. J . “Aromatic” Substituent Constants for Structure-
Activity Correlations. J . Med. Chem. 1973, 16, 1207-1216. (b)
Rekker, R. F.; Mannhold, R. Calculaton of Drug Lipophilicity.
The Hydrophobic Fragmental Constant Approach; VCH: Wein-
heim, Gemany, 1992.
(48) Cai, S. X.; Kher, S. M.; Zhou, Z.-L.; Ilyin, V.; Espitia, S. A.; Tran,
M.; Hawkinson, J . E.; Woodward, R. M.; Weber, E.; Keana, J .
F. W. Structure-Activity Relationships of Alkyl- and Alkoxy-
Substituted 1,4-Dihydroxyquinoxaline-2,3-diones: Potent and
Systemically Active Antagonists for the Glycine Site of the
NMDA Receptor. J . Med. Chem. 1997, 40, 730-738.
J M970417O