Discovery of a series of nonpeptide small molecules that inhibit the binding of insulin-like growth factor (IGF) to IGF-binding proteins

Article abstract of DOI:10.1021/jm010304b

Insulin-like growth factors (IGF-I and II) play an important role in metabolic and mitogenic activities through stimulation of the IGF-I receptor on the cell surface. Although the concentration of IGF in blood and cerebrospinal fluid is quite high (>100 nM) this large pool of IGF is biologically inactive because of its association with six distinct binding proteins which form high-affinity complexes with IGF. Thus, inhibitors of IGF-binding proteins (IGFBPs), especially IGFBP-3, could potentially alter the distribution between the "free" and "bound" forms of IGF and thereby elevate biologically active IGF-I to exert a beneficial effect on those patients with diseases that respond to the application of exogenous IGF-I. Whereas IGF-I peptide variants which bind to IGFBPs but not the IGF-I receptor have been shown to be potent IGF/IGFBP inhibitors small molecule nonpeptide IGF/IGFBP inhibitors have the potential advantages of oral bioavailability and flexible dosing regimen. Here we report the discovery of several isoquinoline analogues, exemplified by 1 and 2, which bind IGFBP-3 as well as other IGFBPs at low nanomolar concentrations. More importantly, both compounds were shown to be able to release biologically active IGF-I from the IGF-I/IGFBP-3 complex. These results point to the feasibility of developing orally active therapeutics to treat IGF-responsive diseases by optimization of the lead molecules 1 and 2.

Full text of DOI:10.1021/jm010304b

J . Med. Chem. 2001, 44, 4001-4010
4001
Discover y of a Ser ies of Non p ep tid e Sm a ll Molecu les Th a t In h ibit th e Bin d in g
of In su lin -lik e Gr ow th F a ctor (IGF ) to IGF -Bin d in g P r otein s
Chen Chen,* Yun-Fei Zhu, Xin-J un Liu, Zi-Xian Lu, Qiu Xie, and Nicholas Ling
Neurocrine Biosciences, Inc., 10555 Science Center Drive, San Diego, California 92121
Received J uly 3, 2001
Insulin-like growth factors (IGF-I and II) play an important role in metabolic and mitogenic
activities through stimulation of the IGF-I receptor on the cell surface. Although the
concentration of IGF in blood and cerebrospinal fluid is quite high (>100 nM), this large pool
of IGF is biologically inactive because of its association with six distinct binding proteins, which
form high-affinity complexes with IGF. Thus, inhibitors of IGF-binding proteins (IGFBPs),
especially IGFBP-3, could potentially alter the distribution between the “free” and “bound”
forms of IGF and thereby elevate biologically active IGF-I to exert a beneficial effect on those
patients with diseases that respond to the application of exogenous IGF-I. Whereas IGF-I
peptide variants, which bind to IGFBPs but not the IGF-I receptor, have been shown to be
potent IGF/IGFBP inhibitors, small molecule nonpeptide IGF/IGFBP inhibitors have the
potential advantages of oral bioavailability and flexible dosing regimen. Here we report the
discovery of several isoquinoline analogues, exemplified by 1 and 2, which bind IGFBP-3 as
well as other IGFBPs at low nanomolar concentrations. More importantly, both compounds
were shown to be able to release biologically active IGF-I from the IGF-I/IGFBP-3 complex.
These results point to the feasibility of developing orally active therapeutics to treat IGF-
responsive diseases by optimization of the lead molecules 1 and 2.
In tr od u ction
nant IGF-I, once or twice daily, throughout the trial
period. Thus, orally active compounds that can stimu-
late the IGF-I receptor would have a distinct advantage
over IGF-I. Despite the publication of a preliminary
report on the identification of a small molecule insulin-
receptor activator from screening of a natural product
library,8 discovery of small molecule agonists for the
IGF-I receptor could be difficult, as the agonist has to
bind both â subunits of the receptor simultaneously to
induce the activation. A further complication in the
design of an orally active IGF-I receptor agonist is that
IGF-I, upon binding to its receptor, undergoes a con-
formational change prior to activation of the receptor9,
and, thus, a small molecule agonist would presumably
have to mimic this change. An alternative approach to
the development of an orally active compound that can
potentiate the action of endogenous IGF-I could be the
use of a small molecule ligand that displaces the bound
IGF-I from the large pool of IGF/IGFBP complexes in
the body fluids.
The insulin-like growth factors (IGF-I and II) are
small polypeptide hormones regulating cell proliferation,
cell differentiation, cell death, and cell metabolic activi-
ties.1 The mitogenic and metabolic actions of the IGFs
are mediated by their binding and activation of the cell
surface IGF-I receptor, a R2â2 heterotetramer closely
related to the insulin receptor.2 The affinity of the IGF-I
receptor for IGF-II is much lower than it is for IGF-I.1
Upon binding the IGF ligand, the IGF-I receptor changes
its conformation, resulting in autophosphorylation of the
intracellular â-subunits and activation of the receptor’s
intrinsic tyrosine kinase activity to propagate the signal
to the nucleus.3 In blood and interstitial fluids, including
the cerebrospinal fluid, the concentration of freely
circulating IGF is exceedingly low because most of the
IGFs are bound to one or more of six high-affinity IGF-
binding proteins (IGFBPs), which inhibit their interac-
tion with the IGF-I receptor.4 The major binding protein
in circulation is IGFBP-3, which together with an acid
labile subunit forms a trimolecular complex with IGF
to sequester most of the IGFs in blood.5 In addition,
IGFBP-5 can also form a similar trimolecular complex,6
whereas the remaining four binding proteins form
dimolecular complexes with IGF.
IGF-I could be liberated from its trimolecular or
dimolecular complexes with the IGFBPs, either in the
blood or in tissue-specific interstitial fluids in the body,
and the “free” biologically active IGF-I could then exert
its effect on the responding cells. Indeed, Loddick et al.
have recently demonstrated an IGF-I-like neuroprotec-
tive effect in rat brain by using a peptide IGF/IGFBP
inhibitor derived from IGF-I that binds to the IGFBPs
but is inactive on the IGF-I receptor.10 Lowman et al.,
using a similar peptide IGF/IGFBP inhibitor, have also
shown that the inhibitor mimicked the IGF-I action on
animal models of diabetes and growth failure.11 More-
over, the same group also discovered a small peptide
(18 amino acid), which binds to human IGFBP-1 and
The multiple activities of IGF-I, from regulation of
glucose homeostasis to whole body growth, have prom-
pted its use in clinical trials to treat insulin resistant
diabetes and growth failure, as well as other IGF-
responsive diseases.7 Because of the intrinsic nature of
a peptide therapeutic, patients participating in these
trials had to be injected with various doses of recombi-
* Corresponding author. Tel: 1-858-658-7600. Fax: 1-858-658-7619.
E-mail: cchen@neurocrine.com.
10.1021/jm010304b CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/03/2001

4002 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23
Chen et al.
2-4
Ta ble 1.
J
Correlations from HMBC Experiments for
C-H
Sch em e 1
Compounds 1 and 2 (in CD3OD)
long-range correlations to carbon number
compound 1 compound 2
C-5, C-9 C-5, C-9
proton number
H-4
H-5
H-8
H-13
H-16
H-17
C-4, C-6, C-7, C-9
C-1, C-6, C-7, C-10
C-11, C-14, C-15, C-17
C-12, C-14, C-15
C-4, C-7, C-9
C-1, C-6, C-7, C-9
C-11, C-15, C-17
C-12, C-14, C-15
C-11, C-13, C-15
C-11, C-13, C-15
Ta ble 2. 1H NMR and 13C NMR Data for Compounds 1 and 2
compound 1 compound 2
proton proton
no.
carbon
carbon
1
3
4
5
6
7
8
9
- - -
- - -
6.95 (s)
7.02 (s)
- - -
- - -
6.96 (s)
- - -
149.9 - - -
157.5 - - -
103.6 8.52 (s)
107.5 7.41 (s)
155.7 - - -
148.7 - - -
107.6 7.21 (s)
118.5 - - -
154.5
138
displaces IGF-I from the IGF-I/IGFBP-1 complex.11 This
peptide, identified from a phage-displayed peptide
library, has a well-defined turn-helix structure in aque-
ous solution, suggesting the possibility of designing
small molecule mimics. However, it remains a formi-
dable challenge to develop a small molecule IGF/IGFBP
inhibitor based on this peptide. As a result, we took the
direct approach to identify small molecule IGF/IGFBP
inhibitors from screening of our in-house chemical
library collection by using the [125I]-IGF-I/IGFBP-3
binding assay. Here we report the discovery of a series
of nonpeptide IGFBP inhibitors, the structures of which
are amenable to optimization for development of orally
active compounds that could displace bound IGF-I from
its binding protein complexes.
122.4
110.1
151.9
150.4
107.1
122.6
132.3
193.2
127.7
10 - - -
11 - - -
12 - - -
141.9 - - -
192.5 - - -
129.5 - - -
13 7.37 (d, J ) 2.0 Hz) 117.6 7.37 (d, J ) 2.0 Hz) 116.5
14 - - -
15 - - -
16 6.84 (d, J ) 8.3 Hz) 116.4 6.83 (d, J ) 8.3 Hz) 115.4
17 7.18 (dd, J ) 2.0,
8.3 Hz)
147.1 - - -
154.2 - - -
145.3
151.8
126.4 7.18 (dd, J ) 1.9,
124.1
166.5
8.3 Hz)
- - -
18 - - -
Resu lts a n d Discu ssion
spectrometry). High-resolution mass spectrometric analy-
sis and elemental analysis confirmed the molecular
formula of 1 as C16H11NO6. An intense absorption band
at 3300 cm-1 in the FT-IR spectrum indicated the
presence of hydroxyl groups in 1. Also, absorption at
1640 cm-1 in conjunction with a 13C NMR signal at δ
192.5 were in agreement with the presence of a conju-
gated ketone carbonyl group.
High throughput screening of the in-house chemical
libraries resulted in the identification of an initial hit
compound, which was designated as L-(3,4-dihydroxy-
phenyl)-alanine (L-DOPA) in the library collection.
However, this result could not be confirmed by using a
freshly prepared sample of L-(3,4-dihydroxyphenyl)-
alanine powder dissolved in DMSO, but the same
sample solution did show activity after being stirred at
room temperature for 3 days in DMSO with the con-
tainer opened to air. HPLC analysis of the stirred
sample revealed several small peaks distinct from L-(3,4-
dihydroxyphenyl)alanine, which exhibited IGF/IGFBP-3
binding inhibitory activity. Preparative HPLC separa-
tion of these peaks (see Experimental Section for details)
resulted in the isolation of several fractions, two of
which showed low nanomolar inhibitory activity on the
IGFBP-3 binding assay. These two components were
then fully characterized by chemical analysis, and the
structures of these two compounds (1 and 2 in Scheme
1) were elucidated on the basis of NMR (proton, carbon,
HMQC, and HMBC), IR, MS, and elemental analysis
methods as illustrated below.
Str u ctu r a l Elu cid a tion of th e Active Com p o-
n en ts. Compound 1 was obtained as a yellow solid. Ion
spray mass spectrometric analysis of the compound
showed a parent peak at m/e 314 (M+H)+. In negative
charge ionization mode, a parent peak at m/e 312 (M-
H)- was detected. These results indicated that the
molecular weight of this compound was 313 and that
the molecule might have both acidic and basic functional
groups. Moreover, the odd number molecular weight
implied it contained an odd number of nitrogen atoms
(which could be the basic center for protonation in mass
1
The H NMR spectrum of 1 in CD3OD showed only
six aromatic protons that were not exchangeable with
deuteriomethanol, and the 13C NMR spectrum of 1
(Table 2) confirmed the presence of 16 carbons. These
NMR results indicated a high degree of unsaturation,
as all carbon resonances were downfield of 100 ppm. A
benzene ring bearing 1,3,4-tri-substituents was identi-
1
fied by H NMR analysis: H-13 (δ 7.37, doublet, J meta
) 2.0 Hz), H-16 (δ 6.84, doublet, J ortho ) 8.3 Hz), and
H-17 (δ 7.18, doublet of doublet, J ortho ) 8.3 Hz, J meta
)
2.0 Hz). All carbons attached to a proton were assigned
on the basis of a 1H-detected heteronuclear one bond
1H-13C correlation experiment (HMQC). Thus, C-13 (δ
117.6), C-16 (δ 116.4), and C-17 (δ 126.4) for the 1,3,4-
tri-substitued phenyl group were identified from the 13
C
NMR spectrum. The remaining protons in 1H NMR
spectrum, three singlets at δ 6.95 (H-4), 7.02 (H-5), and
6.96 (H-8) were correlated to carbons at δ 103.6 (C-4),
107.5 (C-5), and 107.6 (C-8) on the basis of the HMQC
experiment.
Because most carbons in compound 1 were not bonded
to a proton, further information regarding the skeletal
framework was sought from multiple bond proton-
1
carbon couplings, which were identified by a H-detected
1
heteronuclear multiple bond H-13C correlation experi-
ment (HMBC) (Table 1). The presence of a 1,3,4-

Nonpeptide Small Molecules Inhibit Binding of IGF
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23 4003
Sch em e 2a
a
Reagents and conditions. (a) PPA/110 °C. (b) NH4OAc/AcOH/reflux. (c) SeO2/dioxane/reflux, 1 h. (d) 48% HBr/reflux, o/n.
trisubstituted phenyl ring (C-12, C-13, C-14, C-15, C-16,
and C-17) was further confirmed by cross-peaks of H-13
to C-14, C-15, and C-17; H-16 to C-12, C-14, and C-15;
suggested the presence of a carboxylic acid group along
with hydroxyl moieties. An IR band at 1640 cm-1 and
two 13C NMR signals at δ 194.6 and 167.9 indicated the
presence of a conjugated ketone carbonyl group and a
carboxylic acid carbonyl group in 2. The 1H NMR
spectrum of 2 (Table 2) possessed many similarities to
that of 1, including three protons on a 1,3,4-trisubsti-
tuted benzene ring and two singlets on another aromatic
moiety. The significant difference in chemical shift was
H-4, which appeared at δ 8.52 (compared with δ 6.95
in 1). This finding implied the presence of an electron-
withdrawing carboxylic acid group on the pyridine ring.
The 13C NMR spectrum of 2 (Table 2) revealed 17 carbon
signals, which is one more carbon than that of 1. The
extra carbonyl carbon signal at δ 167.9 matched with a
carboxylic acid group. HMQC and HMBC NMR experi-
ments of 2 showed results very similar to those of
compound 1. Therefore, the chemical structure of 2 was
elucidated as 1-(3,4-dihydroxybenzoyl)-3-hydroxy-
carbonyl-6,7-dihydroxy-isoquinoline (Scheme 1).
3
and H-17 to C-13 and C-15. The J C-H coupling of H-13
and H-17 to the carbonyl carbon (δ 192.5) indicated that
this carbonyl group was attached to the phenyl ring with
the two protons located at its ortho positions. The
relative downfield chemical shifts of C-14 and C-15 (δ
147.1 and 154.2) implied that they were bonded to a
hydroxy group. Thus, a 3,4-dihydroxybenzoyl group was
identified.
The presence of a second benzene ring (C-5, C-6, C-7,
3
C-8, C-9, and C-10) was identified from J C-H couplings
of H-5 to C-7 and C-9, and H-8 to C-6 and C-10, and
2
also J C-H couplings of H-5 to C-6 and H-8 to C-7. The
downfield chemical shifts of C-6 and C-7 (δ 155.7 and
148.7) indicated a hydroxy substituent was attached to
each of these two carbons. The cross-peaks of H-5 to C-4
and H-8 to C-1 revealed that C-1 and C-4 were con-
nected to this benzene ring. This finding indicated the
presence of a 1,2,4,5-tetra-substituted benzene ring and
explained why H-5 and H-8 appeared as singlets (no
para coupling observed) in the proton NMR spectrum.
This substitution pattern was confirmed further by the
3J C-H coupling of H-4 to C-5 and C-9. The downfield
chemical shifts of C-1 (δ 149.9) and C-3 (δ 157.5)
suggested that they were bonded to an aromatic nitro-
gen. Thus, a benzene ring fused to a pyridine ring (C-1,
N-2, C-3, C-4, C-10, and C-9) moiety was established
Ch em ica l Syn th esis of th e Active Com p on en ts
a n d Th eir An a logu es. The structures of compounds
1 and 2 were further confirmed by chemical synthesis.
The preparation of 1-(3,4-dihydroxybenzoyl)-3,6,7-tri-
hydroxyisoquinoline 1 is outlined in Scheme 2.
2-[(3,4,-Dim et h oxyph en yl)a cet yl]-4,5-dim et h oxy-
phenylacetic acid 4, synthesized from self-condensation
of 3,4-dimethoxyphenylacetic acid 3 in polyphosphoric
acid (PPA), was subjected to a cyclization with am-
monium acetate in refluxing acetic acid to give 1-(3,4-
dimethoxybenzyl)-3-hydroxy-6,7-dimethoxyisoquino-
line 5 according to a reported protocol.12 Oxidation of
compound 5 with selenium dioxide in dioxane or acetic
acid gave the corresponding ketone 6, but in low yield
(16%). We also tried a direct oxidation of papaverine
with vanadium pentoxide as reported13 but no desired
product was isolated. Demethylation of 6 was ac-
complished in 48% HBr after reflux overnight to give
the designed compound 1. Chemical characterization of
this compound verified that it matched the active
component 1 isolated from the L-(3,4-dihydroxyphenyl)-
alanine/DMSO solution. Similarly, demethylation of
intermediate 5 afforded 1-(3,4-dihydroxybenzyl)-3,6,7-
trihydroxyisoquinoline 7 in good yield.
3
on the basis of intensive J C-H coupling of H-4 to C-9.
By comparing the NMR spectrum of isoquinoline with
that of 3-hydroxyisoquinoline, the chemical shift of C-3
(δ 157.5) in 1 implied that this carbon was bonded to a
hydroxy group like 3-hydroxyisoquinoline. As a result,
the 3,4-dihydroxybenzoyl moiety was logically placed at
C-1 and not at C-3 because no long-range 3J C-H coupling
in the HMBC experiment was observed between H-4
and the C-11 carbonyl carbon. Therefore, the chemical
structure of 1 was elucidated as 1-(3,4-dihydroxybenz-
oyl)-3,6,7-trihydroxy-isoquinoline (Scheme 1).
Compound 2 was also obtained as a yellow solid. Its
molecular formula, C17H11NO7, was established by ion
spray mass spectrometry (m/e 342 in positive charge
ionization mode, m/e 340 in negative charge ionization
mode) and confirmed by high-resolution EIMS and
microanalysis. A very strong IR absorption band at 3400
Scheme 3 depicts the synthesis of 1-(3,4-dihydroxy-
benzoyl)-3-hydroxycarbonyl-6,7-dihydroxyisoquinoline 2.

4004 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23
Chen et al.
Sch em e 3a
a
Reagents and conditions. (a) Mel/NaOH/EtOH. (b) TFA/r.t., then 3,4-dimethoxyphenylacetyl chloride/NaHCO3/H2O/EtOAc. (c) POCl3/
MeCN/reflux. (d) S/160 °C. (e) SeO2/dioxane/reflux, 1 h. (f) 48% HBr/reflux, o/n.
Sch em e 4
1-(3,4-Dimethoxybenzyl)-6,7-dimethoxy-3-ethoxycarb-
onylisoquinoline 8 was synthesized from D,L-(3,4-dihy-
droxyphenyl)-alanine according to a literature proce-
dure.14 Compound 8 was then oxidized with selenium
dioxide in refluxing dioxane to give the corresponding
ketone 9 in excellent yield. Deprotection of compound
9 in 48% hydrobromic acid under reflux overnight
afforded the designed compound 2, which matched the
HPLC retention time, as well as the NMR spectra of,
the active component 2 isolated from the L-(3,4-dihy-
droxyphenyl)alanine/DMSO solution. Deprotection of
compound 8 in refluxing hydrobromic acid gave a
related compound 10 in good yield.
To explore the importance of the hydroxyl groups in
the two catechol moieties for binding activity, compound
9 was partially hydrolyzed in aqueous HBr (reflux, 6 h)
to afford several partially demethylated compounds
(2a -e), which were isolated by HPLC on a reversed
phase C-18 column (Scheme 4). Their structures were
confirmed by HMQC and HMBC NMR experiments.
Oxidation of papaverine 11 with selenium dioxide in
dioxane gave the corresponding 1-(3,4-dimethoxyben-
zoyl)-6,7-dimethoxyisoquinoline 12, which was de-
methylated in refluxing hydrobromic acid to afford the
corresponding 1-(3,4-dihydroxybenzoyl)-6,7-dihydroxy-
isoquinoline 13 (Scheme 5).
Isoquinoline analogues 14a -d were also synthesized
from D,L-(3,4-dihydroxyphenyl)alanine and 3,4-methyl-
enedioxyphenyl-, 2,4-dichlorophenyl-, 1-naphthyl-, or
2-naphthylacetic acid, respectively, according to a pro-
cedure similar to that described in Scheme 3.
In vitr o Bin d in g Activity of Com p ou n d s 1 a n d 2
a n d Th eir An a logu es. The synthesized compounds 1,
2, 2a -e, 6, 7, 9, 10, 13, and 14a -d were evaluated for
their ability to inhibit the binding of radiolabled IGF-I
to IGFBP-3 in a binding assay as described in the
Experimental Section. The data obtained were used to
calculate the Ki value for each compound using Prism
Software, and are summarized in Table 3. Compound
1 showed very good potency (Ki ) 28 nM), but its

Nonpeptide Small Molecules Inhibit Binding of IGF
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23 4005
F igu r e 1. Dose-dependent inhibition of [125I]hIGF-I binding to IGFBP-1, BP-2, BP-3, BP-4, BP-5, and BP-6 by hIGF-I (9), compound
1 (2), and compound 2 (3). The radioligand-binding assay was carried out in duplicate in glass test tubes with 0.5 pmol binding
protein, labeled hIGF-I (30,000 cpm), plus increasing concentrations of hIGF-I, 1, or 2. After incubation at room temperature for
2 h, the IGF/IGFBP complex was precipitated by the addition of 20% BSA and 20% PEG-800 solutions, and the radioactivity in
the pellet was counted on a γ-counter. The data obtained were analyzed by the Prism software to generate the dose-response
curves. The standard errors are within the boundaries of the symbols for those points on the curves that do not have error bars.
Sch em e 5a
hydroxyl groups on compound 2. Whereas methylation
of both 6- and 7-hydroxyl groups of compound 2 abol-
ished its binding affinity (compound 2c and 2e, Ki >10,-
000 nM), demethylation at position-7 of compound 2e
resulted in an active analogue 2a (Ki ) 320 nM).
Introduction of a bromine atom at position-8 and dem-
ethylation at position-15 of 2a increased the activity of
the resulting compound (compound 2b, Ki ) 140 nM)
about 2-fold. A closely related analogue of 2b with a
5-bromo-6-hydroxy-7-methoxy-isoquinoline moiety (com-
pound 2d , Ki ) 70 nM) also showed relatively high
activity.
The catechol functional group was crucial for impart-
ing high affinity to the isoquinoline nucleus. By contrast,
the catechol moiety in the 1-benzoyl group of compounds
1 and 2 was not critical for activity. This conclusion was
based on the SAR data of compounds 14a -d , which only
lost 3- to 12-fold activity in comparison with the parent
compound 2. Compound 14d was only 3-fold less active
than compound 2, and more importantly it was much
more lipophilic than compound 2. These compounds
could serve as leads for the design of orally bioavailable
small molecules that disrupt the binding of IGFs to their
binding proteins. The isoquinoline derivatives 1 and 2
also showed high binding activity on other IGFBPs in
addition to IGFBP-3 as shown in Figure 1 and their
respective affinity constants for the six IGFBPs are
presented in Table 4.
a
Reagents and conditions. (a) SeO2/dioxane/reflux, 1 h. (b) 48%
HBr/reflux, o/n.
precursor 6 was completely inactive, indicating that one
or more of the catechol hydroxy groups are essential for
activity. Replacement of the 3,4-dihydroxybenzoyl moi-
ety of 1 at position-1 with a 3,4-dihydroxybenzyl group
resulted in a compound with the same binding affinity
(compound 7, Ki ) 29 nM). Although removal of the
hydroxyl group on C-3 of the isoquinoline ring (com-
pound 13, Ki ) 58 nM) caused a 2-fold decrease in
activity compared to 1, replacement of the same hy-
droxyl group with a more acidic carboxylic moiety
yielded a molecule with 5-fold increase in affinity
(compound 2, Ki ) 5.6 nM). The corresponding benzylic
analogue of 2 (compound 10, Ki ) 7.5 nM) showed only
slightly decreased activity, whereas the precursor of 2,
compound 9, displayed no activity. These data suggested
that the ketone functional group in compounds 1 and 2
had minimum effect on IGFBP-3 binding affinity, posi-
tion-3 of the isoquinoline ring required an acidic group
for high potency, and one or both of the two catechol
moieties were crucial for IGFBP-3 activity.
In vitr o Bioa ctivity of Com p ou n d s 1 a n d 2. To
demonstrate the ability of compounds 1 and 2 to release
bioactive IGF-I from its bound form with IGFBP-3, we
employed a cell proliferation assay using 3T3 fibro-
blasts,15 which expressed IGF-I receptors and responded
to IGF-I stimulation through cell proliferation. The cell
proliferation effect was monitored by quantitation of
[3H]-thymidine incorporation into the cells. As seen in
Figure 2, human IGF-I (2 nM) induced a robust prolif-
eration of 3T3 fibroblasts as reflected by a >2-fold
Next, we examined several partially methylated
compounds to determine the importance of the catechol

4006 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23
Chen et al.
Ta ble 3. IGFBP-3 Inhibitory Activity of Compounds 1 and 2 and Their Analogues
* Single test.
increase of [3H]-thymidine incorporation above basal
level into the cells and the hIGF-I-induced proliferation
was substantially blocked by addition of hIGFBP-3 (4
nM). Addition of compounds 1 (Figure 2A) and 2 (Figure
2B) dose-dependently reversed the neutralization effect
of hIGFBP-3 on the activity of hIGF-I, demonstrating
the ability of 1 and 2 to displace bioactive IGF-I bound
to IGFBP-3. Moreover, maximal reversal of the IGFBP-3
inhibitory was reached at 1 µM concentration for
compound 2 but at 10 µM concentration for compound
1. These results are in agreement with the binding
assay data (Table 3), showing that 2 is 5-fold more
potent than 1. Compounds 1 or 2 (10 µM) alone had no
effect on [3H]-thymidine incorporation in these cells.
These in vitro data clearly demonstrate that isoquino-
lines 1 and 2 are capable of interacting with the binding
protein in a specific manner to displace and release
“free” bioactive IGF-I.
Con clu sion
A series of highly active, nonpeptide small molecule
inhibitors of IGF/IGFBP-3 binding was identified from
high throughput screening, followed by structural elu-
cidation of the active components isolated from the

Nonpeptide Small Molecules Inhibit Binding of IGF
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23 4007
Ta ble 4. Binding Affinities of Compounds 1 and 2 and IGF-I
on IGFBP-1 to IGFBP-6
were performed on a Varian FT-500 NMR spectrometer.
Electrospray ionization mass spectra (EI-MS) were recorded
on a PE Sciex API LC/MS System. FT-IR spectra were
recorded on a Perkin-Elmer 1605 IR Spectrophotometer (KBr).
HPLC analyses were carried out on a Beckman 322 gradient
HPLC system, whereas preparative HPLC separations were
performed on a Biotage Kiloprep 100G HPLC system. All
reagents and solvents were of ACS reagent grade or higher
and were used as provided.
Sep a r a tion of th e Active Com p on en ts fr om L-3,4-
(Dih yd r oxyp h en yl)-a la n in e. One gram of l-3,4-dihyroxy-
phenyl-alanine (Sigma) was stirred in 200 mL of dimethyl
sulfoxide at room temperature in an open flask for 3 days. The
resulting brown solution was purified on a Kiloprep 100G
preparative HPLC system equipped with a C18 cartridge and
in an acetonitrile gradient in aqueous 0.1% trifluoroacetic acid.
The collected fractions were monitored by the IGFBP-3 radio-
ligand-binding assay after removal of the solvent. The major
active fractions from four repetitive runs were pooled, and the
solvent was evaporated; the residue was characterized by MS,
IR, and NMR (including NOE, HMQC, and HMBC) analyses.
Ki (nM)
IGFBP
hIGF-I
1
2
BP-1
BP-2
BP-3
BP-4
BP-5
BP-6
0.17 ( 0.03
0.07 ( 0.02
0.21 ( 0.05
0.06 ( 0.02
0.20 ( 0.04
3.10 ( 0.20
34.23 ( 11.41
17.13 ( 5.65
28.23 ( 8.06
6.23 ( 0.62
20.95 ( 0.05
84.65 ( 12.35
5.07 ( 0.74
2.07 ( 0.03
5.63 ( 0.46
1.04 ( 0.19
5.20 ( 0.30
23.65 ( 3.35
1-(3,4-Dim eth oxyben zyl)-3-h yd r oxy-6,7-d im eth oxyiso-
qu in olin e (5).12 3,4-Dimethoxyphenyl-acetic acid (5 g) was
mixed with PPA (100 g), and the mixture was heated on a
water bath for 0.5 h and then stirred at room-temperature
overnight. The dark brown syrup was poured into ice-water
(800 mL) and allowed to stand for 3 h. The yellow solid was
collected by filtration. The collected solid was recrystallized
from ethanol to afford a yellowish solid (2 g), which was mixed
with ammonium acetate (4 g) in acetic acid (5 mL) and heated
at reflux for 1 h and then allowed to cool. The cold solution
was diluted with methylene chloride and washed with water.
The organic phase was dried over sodium carbonate and
concentrated in vacuo to give the desired compound as a yellow
solid (1.8 g). 1H NMR (CDCl3/TMS): 3.79 (s, 3H), 3.80 (s, 3H),
3.87 (s, 3H), 3.95 (s, 3H), 4.37 (s, 2H), 6.65 (s, 1H), 6.68 (s,
1H), 6.75 (d, J ) 8.4 Hz, 1H), 6.80 (dd, J ) 1.0, 8.4 Hz, 1H),
6.95 (d, J ) 1.0 Hz, 1H), 6.96 (s, 1H). 13C NMR (CDCl3/TMS):
98.0, 98.4, 100.9, 106.9, 107.5, 109.3, 116.0, 124.9, 137.9, 143.3,
143.5, 143.7, 144.7, 150.7, 156.4. EI-MS m/e 355 (M+H)+.
1-(3,4-Dim et h oxyb en zoyl)-3-h yd r oxy-6,7-d im et h oxy-
isoq u in olin e (6). 1-(3,4-Dimethoxybenzyl)-3-hydroxy-6,7-
dimethoxyisoquinoline (355 mg) was mixed with selenium
dioxide (1.1 equiv) in dioxane (5 mL) and heated at reflux for
2 h. The resultant mixture was purified on silica gel with 5%
MeOH in methylene chloride as the eluent to give the product
1
as a yellow oil (60 mg). H NMR: (CDCl3/TMS): 3.93 (s, 3H),
3.95 (s, 3H), 4.00 (s, 3H), 4.09 (s, 3H), 6.82 (s, 1H), 6.89 (d, J
) 1.8 Hz, 1H), 6.95 (d, J ) 8.4 Hz, 1H), 7.02 (dd, J ) 1.8, 8.4
Hz, 1H), 7.30 (s, 1H), 7.62 (s, 1H), 8.88 (brs, 1H). EI-MS m/e
370 (M+H)+. HR-MS Anal. for C20H20NO6: calcd. m/e, 370.1291;
found m/e, 370.1280.
F igu r e 2. Reversal of IGFBP-3 inhibition of IGF-I-stimulated
Balb/c 3T3 fibroblast proliferation by compounds 1 (panel A)
and 2 (panel B). Human IGF-I (2 nM) induced a >2-fold
increase in proliferation of the fibroblast compared to that of
the control as measured by the incorporation of [3H]thymidine.
Human IGFBP-3 (4 nM) completely inhibited the proliferative
effect of 2 nM IGF-I. Addition of compound 1 (panel A) or
compound 2 (panel B) dose-dependently reversed this inhibi-
tion. Both 1 and 2 alone have no effect on fibroblast prolifera-
tion. Data represent the mean ( SEM.
1-(3,4-Dih yd r oxyben zoyl)-3-h yd r oxy-6,7-d ih yd r oxyiso-
qu in olin e (1). 1-(3,4-Dimethoxybenzoyl)-3-hydroxy-6,7-dimeth-
oxyisoquinoline (15 mg) was suspended in 48% HBr (3 mL)
and heated at reflux for 4 h. The crude product was purified
by reversed-phase HPLC on a C-18 column using a gradient
of acetonitrile in 0.1% TFA/H2O to give the product as a
1
yellowish powder (2 mg). H NMR (CD3OD/TMS): 6.83 (d, J
) 8.3 Hz, 1H), 6.94 (s, 1H), 6.95 (s, 1H), 7.01 (s, 1H), 7.17 (dd,
J ) 2.0, 8.4 Hz, 1H), 7.36 (d, J ) 2.0 Hz, 1H). 13C NMR: 103.9,
107.6, 116.4, 117.6, 118.6, 126.6, 129.5, 141.9, 147.2, 148.8,
150.0, 154.2, 155.9, 192.3. IR (KBr): 3300 (s), 1629 (s), 1590
(m), 1465 (m). EI-MS m/e 314 (M+H)+; (negative charge
mode): m/e 312 (M-H)-. HR-MS Anal. for C16H12NO6: calcd.
m/e, 314.0665; found m/e, 314.0660. Anal. (C16H11NO6 ‚ 0.5
H2O) C, H, N.
screening hit solution. These compounds inhibited IGF-I
binding to IGFBP-3 and thus were able to release “free”
IGF-I from its binding protein complex at low nano-
molar concentrations. Furthermore, the IGF-I released
by compounds 1 and 2 was shown to be biologically
active in an in vitro bioassay. These results demon-
strated that nonpeptide small molecule IGF/IGFBP
binding inhibitors could be developed and possibly used
as therapeutics to treat IGF-responsive diseases.
1-(3,4-Dih yd r oxyben zyl)-3-h yd r oxy-6,7-d ih yd r oxyiso-
qu in olin e (7). 1-(3,4-Dimethoxybenzyl)-3-hydroxy-6,7-dimeth-
oxyisoquinoline (15 mg) was suspended in 48% HBr (3 mL)
and heated at reflux for 4 h. The crude product was purified
with HPLC on a reversed-phase column to give the product
Exp er im en ta l Section
1H NMR and 13C NMR were recorded on a Varian Mercury
FT-300 NMR spectrometer, HMQC and HMBC experiments
1
as a yellowish powder (2 mg). H NMR (CDCl3/TMS): 4.43 (s,
2H), 6.53 (dd, J ) 2.1, 8.4 Hz, 1H), 6.60 (d, J ) 2.1 Hz, 1H),

4008 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23
Chen et al.
6.70 (d, J ) 8.4 Hz, 1H), 7.00 (s, 1H), 7.08 (s, 1H), 7.43 (s,
1H). MS (EI) m/e 300 (M+H)+.
and the refluxing continued for 0.5 h. After the mixtured cooled
to room temperature, methylene chloride was added, and the
solution was filtered through a short silica column to remove
a black solid. The filtrate was concentrated to give a tan solid,
which was washed with ether. Filtration isolated the desired
product as a white solid (19.5 g). 1H NMR (CDCl3/TMS): 1.44
(t, J ) 7.5 Hz, 3H), 3.96 (s, 3H), 3.97 (s, 3H), 4.01 (s, 3H), 4.08
(s, 3H), 4.49 (q, 2H), 6.87 (d, J ) 8.4 Hz, 1H), 7.28 (s, 1H),
7.52 (dd, J ) 1.8, 8.4 Hz, 1H), 7.57 (s, 1H), 7.81 (d, J ) 1.8
1-(3,4-Dim eth oxyben zyl)-3-eth oxyca r bon yl-6,7-d im eth -
oxyisoqu in olin e (8).14 N-Boc-D,L-3,4-dimethoxyphenylalanine
ethyl ester (82.0 g, 0.232 mol) was stirred in trifluoroacetic
acid (70 mL, 0.58 mol) until bubbling ceased (approximately
30 min). The trifluoroacetic acid was removed in vacuo, and
the residue was coevaporated with toluene twice to give D,L-
3,4-dimethoxyphenylalanine ethyl ester trifluoroacetic acid salt
as an oil. 1H NMR (CDCl3/TMS): 1.31 (t, J ) 7.0 Hz, 3H), 3.13
(dd, J ) 7.8, 14.4 Hz, 1H), 3.30 (dd, J ) 4.8, 14.4 Hz, 1H),
3.83 (s, 3H), 3.84 (s, 3H), 4.27 (m, 3H), 6.60 (brs, 3H), 6.73 (d,
J ) 2.1 Hz, 1H), 6.74 (dd, J ) 2.1, 9.0 Hz, 1H), 6.82 (d, J )
9.0 Hz, 1H).
The deprotected amino acid ethyl ester was dissolved in 600
mL of H2O containing sodium bicarbonate (60 g, 0.510 mol).
To this mixture was added slowly a solution of 3,4-dimethoxy-
phenylacetyl chloride, synthesized freshly from 3,4-dimeth-
oxyphenylacetic acid (47.8 g, 0.244 mol) and oxalyl chloride
(40.5 mL, 0.464 mol) in 400 mL of ethyl acetate. The mixture
was stirred for 16 h at room temperature. The ethyl acetate
and H2O layers were separated, and the organic layer was
washed with 2 × 300 mL of saturated sodium bicarbonate,
dried over MgSO4, filtered, and concentrated in vacuo to give
N-(3,4-dimethoxyphenylacetyl)-D,L-3,4-dimethoxyphenyl-
alanine ethyl ester as a tan solid (81 g). 1H NMR (CDCl3/
TMS): 1.24 (t, J ) 7.2 Hz, 3H), 2.98 (m, 2H), 3.50 (s, 2H),
3.78 (s, 3H), 3.82 (s, 3H), 3.86 (s, 3H), 3.89 (s, 3H), 4.15 (q, J
) 7.2 Hz, 2H), 4.78 (m, 1H), 5.90 (d, J ) 7.5 Hz, 1H), 6.40 (dd,
J ) 2.1, 8.1 Hz, 1H), 6.51 (d, J ) 2.1 Hz, 1H), 6.65 (d, J ) 8.1
Hz, 1H), 6.70 (d, J ) 1.8 Hz, 1H), 6.71 (dd, J ) 7.8, 1.8 Hz,
1H), 6.80 (d, J ) 7.8 Hz, 1H).
Hz, 1H), 8.55 (s, 1H). EI-MS m/e 426 (M+H)+. Anal. (C23H23
-
NO7) C, H, N.
1-(3,4-Dih yd r oxyben zoyl)-3-h yd r oxyca r bon yl-6,7-d ih y-
d r oxyisoqu in olin e (2). 1-(3,4-dimethoxybenzoyl)-3-ethoxy-
carbonyl-6,7-dimethoxyisoquinoline (9, 2 g) was refluxed in
48% HBr (15 mL) for 40 h. A yellow solid that precipitated
upon cooling was collected by filtration, washed with ethyl
acetate, and dried under vacuum with P2O5 to give a yellow
solid (1.4 g). A 1-g portion of the compound was dissolved in
refluxing MeOH (20 mL), and boiling water (about 40 mL) was
added until the solution was cloudy. This mixture was allowed
to sit overnight, and the precipitated light yellow solid was
collected by filtration to give 0.78 g of the pure product. 1H
NMR (DMSO-D6/TMS): 6.82 (d, J ) 8.1 Hz, 1H), 7.09 (dd, J
) 1.5, 8.1 Hz, 1H), 7.11 (s, 1H), 7.24 (d, J ) 1.5 Hz, 1H), 7.43
(s, 1H), 8.43 (s, 1H), 9.20 (brs, 1H), 10.06 (brs, 1H), 10.50 (brs,
2H). 13C NMR: 109.1, 111.3, 116.3, 117.7, 124.6, 124.8, 126.5,
129.8, 135.2, 138.1, 147.0, 152.3, 153.9, 154.3, 156.0, 167.9,
194.6; IR (KBr): 3408, 1650, 1588, 1516, 1409, 1286,
1197. EI-MS m/e 342 (M+H)+, (negative charge mode): m/e
340 (M-H)-. HR-MS for C17H12NO7 (M+H)+: calcd. m/e,
342.0614; found m/e, 342.0618. Anal. (C17H11NO7 ‚ 2.7H2O) C,
H, N.
1-(3,4-Meth ylen edioxyben zoyl)-3-h ydr oxycar bon yl-6,7-
d ih yd r oxyisoqu in olin e (14a ). This compound was prepared
in a manner similar to the procedure described for 2 from
methylenedioxyphenylacetic acid. White powder. 1H NMR
(DMSO-D6/TMS): 6.30 (s, 2H), 7.14 (d, J ) 8. Hz, 1H), 7.31
(s, 1H), 7.40 (d, J ) 8.1 Hz, 1H), 7.55 (s, 1H), 7.57 (s, 1H),
8.58 (s, 1H), 10.70 (brs, 2H). EI-MS m/e 354 (M+H)+. Anal.
(C18H11NO7‚0.67 H2O‚0.33 HBr) C, H, N.
1-(2,4-Dich lor ob en zoyl)-3-h yd r oxyca r b on yl-6,7-d ih y-
d r oxyisoqu in olin e (14b). This compound was prepared in
a manner similar to the procedure described for 2 from 2,4-
dichlorophenylacetic acid. Yellowish powder. 1H NMR (DMSO-
D6/TMS): 7.40 (s, 1H), 7.51 (dd, J ) 1.8, 8.1 Hz, 1H), 7.60 (d,
J ) 8.1 Hz, 1H), 7.67 (d, J ) 2.1 Hz, 1H), 8.01 (s, 1H), 8.45 (s,
1H), 10.60 (brs, 1H), 10.79 (brs, 1H). EI-MS m/e 378 (M+H)+.
Anal. (C17H9NCl2O5‚0.5 H2O‚0.5 MeOH) C, H, N.
N-(3,4-dimethoxyphenylacetyl)-D,L-3,4-dimethoxyphenyl-
alanine ethyl ester (81 g, 0.188 mol) was dissolved in 300 mL
of acetonitrile. Phosphorus oxychloride (88 mL, 0.94 mol) was
added and the reaction mixture was heated to reflux for 2 h.
After the mixture was cooled to room temperature, the solvent
and excess phosphorus oxychloride were removed under
vacuum using a KOH trap. The residual dark oil was then
diluted with 100 mL of methylene chloride and 400 mL of ice-
water, and neutralized with Na2CO3 while stirring vigorously
on an ice-water bath. The mixture was then diluted with 200
mL of ethyl acetate, the organic and aqueous layers were
separated, and the aqueous layer was extracted with 2 × 300
mL of ethyl acetate. The combined organic layers were washed
with 2 × 300 mL of saturated sodium bicarbonate, dried over
MgSO4, and concentrated to give 1-(3,4-dimethoxybenzyl)-3-
ethoxycarbonyl-3,4-dihydro-6,7-dimethoxyisoquinoline as an
orange solid, which was carried forward without purification.
1H NMR (CDCl3/TMS): 1.31 (t, J ) 7.2 Hz, 3H), 2.95 (d, J )
9.0 Hz, 2H), 3.72 (s, 3H), 3.83 (s, 6H), 3.88 (s, 3H), 4.03 (d, J
) 14.8 Hz, 1H), 4.10 (d, J ) 14.8 Hz, 1H), 4.27 (q, J ) 7.2 Hz,
2H), 4.39 (t, J ) 9.0 Hz, 1H), 6.66 (s, 1H), 6.78 (d, J ) 8.1 Hz,
1H), 6.88 (dd, J ) 8.1, 1.8 Hz, 1H), 6.94 (d, J ) 1.8 Hz, 1H),
7.00 (s, 1H), 7.27 (s, 1H). EI-MS m/e 414 (M+H)+.
1-(3,4-Dimethoxybenzyl)-3-ethoxycarbonyl-3,4-dihydro-6,7-
dimethoxyisoquinoline (77.7 g, 0.188 mol) was dissolved in 30
mL of methylene chloride. Sulfur (9.04 g, 0.282 mol) was
added, and the mixture was heated to 160 °C until bubbling
ceased in approximately 30 min. The mixture was cooled to
approximately 50 °C, diluted with ethyl acetate, and stirred
while being cooled to room temperature, followed by filtration
through a short silica gel column with ethyl acetate. The
product precipitated from the filtrate was collected to give a
yellow solid (48 g). 1H NMR (CDCl3/TMS): 1.48 (t, J ) 7.0
Hz, 3H), 3.76 (s, 3H), 3.82 (s, 3H), 3.87 (s, 3H), 4.02 (s, 3H),
4.53 (q, J ) 7.0 Hz, 2H), 4.65 (s, 2H), 6.75 (d, J ) 8.1 Hz, 1H),
6.80 (dd, J ) 8.1, 1.8 Hz, 1H), 6.83 (d, J ) 1.8 Hz, 1H), 7.17
(s, 1H), 7.36 (s, 1H). EI-MS m/e 412 (M+H)+.
1-(1-Na p h t h yl)-3-h yd r oxyca r b on yl-6,7-d ih yd r oxyiso-
qu in olin e (14c). This compound was prepared in a manner
similar to the procedure described for 2 from 1-naphthylacetic
1
acid. Yellowish powde. H NMR (CD3OD/TMS): 7.29 (s, 1H),
7.32 (d, J ) 7.2 Hz, 1H), 7.36 (d, J ) 1.8 Hz, 1H), 7.37 (t, J )
7.8 Hz, 1H), 7.48 (dt, J ) 1.5, 6.6 Hz, 1H), 7.55 (dt, J ) 1.2,
6.9 Hz, 1H), 7.92 (d, J ) 7.8 Hz, 1H), 8.05 (dd, J ) 2.1, 7.2 Hz,
1H), 8.31 (s, 1H), 8.62 (d, J ) 8.4 Hz, 1H), 10.45 (brs, 1H).
EI-MS m/e 360 (M+H)+. Anal. (C21H13NO5 ‚2 H2O) C, H, N.
1-(2-Na p h t h yl)-3-h yd r oxyca r b on yl-6,7-d ih yd r oxyiso-
qu in olin e (14d ). This compound was prepared in a manner
similar to the procedure described for 2 from 2-naphthylacetic
acid. Yellowish powder. 1H NMR (CD3OD/TMS): 7.33 (s, 1H),
7.41 (s, 1H), 7.53 (t, J ) 7.2 Hz, 1H), 7.64 (d, J ) 7.2 Hz, 1H),
7.89 (d, J ) 7.5 Hz, 1H), 7.94 (d, J ) 8.4 Hz, 1H), 8.00 (d, J )
9.0 Hz, 1H), 8.09 (dd, J ) 1.8, 9.0 Hz, 1H), 8.26 (s, 1H), 8.55
(s, 1H). EI-MS m/e 360 (M+H)+. Anal. (C21H13NO5 ‚1.3 H2O)
C, H, N.
P a r t ia l H yd r olysis of 1-(3,4-Dim et h oxyb en zoyl)-3-
et h oxyca r b on yl-6,7-d im et h oxyisoq u in olin e (9). 1-(3,4-
dimethoxybenzoyl)-3-ethoxycarbonyl-6,7-dimethoxyisoquino-
line (200 mg) was refluxed in 48% HBr (5 mL) for 8 h. The
following compounds, in addition of compound 2, were sepa-
rated by reversed-phase HPLC on a C-18 column. The struc-
tures of these compounds were assigned on the basis of NMR
studies including HMBC NMR experiments.
1-(3,4-Dim e t h oxyb e n zoyl)-3-e t h oxyca r b on yl-6,7-d i-
m eth oxyisoqu in olin e (9). 1-(3,4-Dimethoxybenzyl)-3-ethoxy-
carbonyl-6,7-dimethoxyisoquinoline (8, 20.55 g, 0.05 mol) was
stirred in 500 mL of dioxane and heated to reflux to give a
clear solution. Selenium dioxide (5.55 g, 0.05 mol) was added

Nonpeptide Small Molecules Inhibit Binding of IGF
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23 4009
1-(3-Hyd r oxy-4-m eth oxyben zoyl)-3-h yd r oxyca r bon yl-
6-m et h oxy-7-h yd r oxyisoq u in olin e (2a ). White powder.
HPLC: >96% purity, tR ) 12.4 min. 1H NMR (DMSO-D6/TMS,
500 MHz): 3.84 (s, 3H), 3.97 (s, 3H), 7.03 (d, J ) 8.5 Hz, 1H),
7.16 (s, 1H), 7.18 (dd, J ) 1.6, 8.5 Hz, 1H), 7.30 (d, J ) 1.6
Hz, 1H), 7.70 (s, 1H), 8.56 (s, 1H), 9.55 (brs, 2H). 13C NMR
(DMSO-D6/TMS, 500 MHz): 55.8, 56.0, 106.7, 107.3, 111.5,
115.7, 123.1, 123.3, 124.1, 128.9, 132.2, 138.3, 146.6, 150.8,
153.1, 153.2, 154.0, 166.4, 193.3. EI-MS m/e 370 (M+H)+.
1-(3,4-Dih ydr oxyben zoyl)-3-h ydr oxycar bon yl-8-br om o-
7-h yd r oxy-6-m et h oxyisoqu in olin e (2b ). White powder.
HPLC: 92% purity, tR ) 11.95 min. 1H NMR (DMSO-D6/TMS,
500 MHz): 4.07 (s, 3H), 6.81 (d, J ) 8.5 Hz, 1H), 7.00 (d, J )
8.1 Hz, 1H), 7.20 (s, 1H), 7.85 (s, 1H), 8.60 (s, 1H), 9.40 (brs,
1H), 9.98 (brs, 1H), 10.92 (brs, 1H). 13C NMR (DMSO-D6/TMS,
500 MHz): 56.8, 102.3, 107.3, 115.3, 116.3, 122.5, 123.2, 123.6,
129.3, 133.6, 138.8, 145.2, 148.4, 151.1, 152.0, 155.5, 166.1,
192.8. EI-MS m/e 436 (M+H)+.
buffer, 100 µL of hIGF-I (Sigma, St. Louis, MO) in buffer
solution, followed by 100 µL of [125I]hIGF-I (30,000 cpm,
specific activity ∼2,000 Ci/mmol; New England Nuclear,
Boston, MA) to a 12 × 75-mm glass test tube. Then 200 µL of
a 2.5 nM hIGFBP-3 in buffer solution (0.5 pmol) was added,
and the mixture was incubated for 2 h. After incubation, 100
µL of ice-cold 20% BSA and 500 µL of ice-cold 20% PEG-8000
in PBS buffer were added and the mixture was vortexed and
then centrifuged for 30 min at 3000 rpm. The supernatant was
carefully removed by suction, and the pellet was counted on a
γ-counter. One-half pmol of hIGFBP-3 bound ∼24% of [125I]-
hIGF-I, and the nonspecific binding was ∼2%. Binding assays
for the other five human IGFBPs were set up in a similar
manner. The data obtained were used to calculate the Ki values
by the Prism Software (GraphPad Software, Inc., San Diego,
CA).
High -Th r ou gh p u t Scr een in g. To identify molecules that
could displace IGF from its binding protein, an in-house library
of ∼85,000 compounds was screened by an adaptation of the
hIGFBP-3 radioligand-binding assay on a Biomek 2000 Labo-
ratory Automation Workstation (Beckman-Coulter, Fullerton,
CA). Compounds predissolved in dimethyl sulfoxide were
screened at a final concentration of 10 µM in a 96-well format,
and the criterion for a positive hit was set at >50% inhibition
of [125I]hIGF-I binding to hIGFBP-3 at this concentration.
F ibr obla st P r olifer a tion Assa y. Balb/c 3T3 fibroblast
cells (ATCC, Rockville, MD) were cultured in DMEM medium
(Fisher Scientific, Springfield, NJ ). The cells grown up in the
culture flask were trypsinized and diluted to 50,000 cells/mL
in 10% FBS (HyClone, Logan, UT) in DMEM. The cells were
aliquoted to 96-well tissue culture plates (Corning, Corning,
NY) at 180 µL/well. After 48 h of incubation at 37 °C and 5%
CO2, the plates were washed twice with serum-free DMEM
and incubated with 180 µL of serum-free DMEM for an
additional 2 h. To each well, 20 µL of sample was then added,
and the cells were incubated for 4 h. (Compounds 1 and 2 were
predissolved in dimethyl sulfoxide at 10 mM concentration and
diluted 1:10 in serum-free DMEM before being added to the
cells.) Then 1 µCi [3H]thymidine (New England Nuclear) was
added to each well and the plates were incubated for another
20 h. Following the incubation, the cells in each well were
transferred to a 96-well GF/B filter plate (Packard, Meriden,
CT) and dried at 65 °C for 1 h. Scintillation fluid (50 µL/well)
was added, and the plates were counted on a microplate
scintillation counter (Packard Instruments, Meriden, CT).
1-(3,4-D ih y d r o x y b e n zo y l)-3-h y d r o x y c a r b o n y l-6,7-
d im eth oxyisoqu in olin e (2c). White powder. HPLC: >99%,
1
tR ) 13.3 min. H NMR (CD3OD/TMS): 3.93 (s, 3H), 4.08 (s,
3H), 6.87 (d, J ) 8.1 Hz, 1H), 7.28 (s, 1H), 7.34 (dd, J ) 8.1
Hz, 1H), 7.50 (d, J ) 2.4 Hz, 1H), 7.52 (s, 1H), 8.56 (s, 1H).
EI-MS m/e 370 (M+H)+.
1-(3,4-Dih ydr oxyben zoyl)-3-h ydr oxycar bon yl-5-br om o-
6-h yd r oxy-7-m et h oxyisoq u in olin e (2d ). White powder.
HPLC: 95% purity, tR ) 13.9 min. 1H NMR (CD3OD/TMS, 500
MHz): 3.93 (s, 3H), 6.78 (d, J ) 8.5 Hz, 1H), 7.04 (d, J ) 2.1,
8.5 Hz, 1H), 7.09 (d, J ) 2.1 Hz, 1H), 7.13 (s, 1H), 8.28 (s,
1H). 13C NMR (CD3OD/TMS, 500 MHz): 57.0, 107.0, 113.2,
116.6, 118.0, 124.3, 124.6, 126.4, 129.5, 137.9, 139.4, 146.7,
152.7, 153.7, 156.8, 157.1, 167.6, 195.2. EI-MS m/e 436
(M+H)+.
1-(3-Hyd r oxy-4-m eth oxyben zoyl)-3-h yd r oxyca r bon yl-
6,7-dim eth oxyisoqu in olin e (2e).White powder.HPLC: >99%
purity, tR ) 17.65 min. 1H NMR (CD3OD/TMS, 500 MHz): 3.97
(s, 3H), 3.98 (s, 3H), 4.09 (s, 3H), 6.90 (d, J ) 8.1 Hz, 1H),
7.32 (s, 1H), 7.42 (dd, J ) 2.4, 8.1 Hz, 1H), 7.48 (s, 1H), 7.57
(d, J ) 2.4 Hz, 1H), 8.60 (s, 1H). EI-MS m/e 384 (M+H)+.
1-(3,4-Dih yd r oxyben zyl)-3-h yd r oxyca r bon yl-6,7-d ih y-
d r oxyisoqu in olin e (10). 1-(3,4-Dimethoxybenzyl)-3-ethoxy-
carbonyl-6,7-dimethoxyisoquinoline (100 mg) was refluxed in
48% HBr (3 mL) for 20 h. After cooling, the solid was collected
to give the product as a white powder. 1H NMR (CD3OD/
TMS): 5.80 (s, 2H), 6.57 (d, J ) 8.0 Hz, 1H), 6.65 (s, 1H), 6.72
(d, 1H), 7.56 (s, 1H), 7.78 (s, 1H), 8.62 (s, 1H). EI-MS m/e 328
(M+H)+. Anal. (C17H13NO6 ‚ 0.67 H2O) C, H, N.
Ack n ow led gm en t. This work was partially sup-
ported by a SBIR grant R43-DK52243 from NIH. We
thank Mr. Greg Reinhart for performing the IGFBP
binding assays, Lee Carter for high-throughput screen-
ing, Ms. Mila Lagman for HPLC purification of some
compounds, and Dr. Changlu Liu and Ms. Yan Gao for
IGFBP expression and purification.
1-(3,4-Dih yd r oxyb en zoyl)-6,7-d ih yd r oxyisoq u in olin e
(13). Papaverine (1 g) was mixed with selenium dioxide (1.1
equiv) in dioxane (20 mL) and heated at reflux for 2 h. The
resultant mixture was purified on silica gel with 5% MeOH
in methylene chloride to give 1-(3,4-dimethoxybenzoyl)-6,7-
dimethoxyisoquinoline as a white solid. EI-MS 354 (M+H)+.
This compound (200 mg) was suspended in 48% HBr (3 mL)
and heated at reflux for 4 h. The mixture was cooled to room
temperature and a gray solid precipitated. The solid was
collected by vacuum filtration to give the product as a brownish
Refer en ces
(1) For a recent review, see J ones, J . I.; Clemmons, D. R. Insulin-
like growth factors and their binding proteins: Biological actions.
Endocr. Rev. 1995, 16, 3-34.
(2) Ullrich, A.; Gray, A.; Tam, A. W.; Yang-Fang, T.; Tsubokawa,
M.; Collins, C.; Henzel, W.; Le Bon, T.; Kathuria, S.; Chen, E.;
J acobs, S.; Francke, U.; Ramachandran, J .; Fujita-Yamaguchi,
1
solid (130 mg). H NMR (CD3OD/TMS, 500 MHz): 6.89 (d, J
) 8.1 Hz, 1H), 7.12 (dd, J ) 2.1, 8.1 Hz, 1H), 7.27 (s, 1H), 7.41
(d, J ) 2.1 Hz, 1H), 7.52 (s, 1H), 8.20 (d, J ) 6.6 Hz, 1H), 8.28
(d, J ) 6.6 Hz, 1H). EI-MS m/e 298 (M+H)+. Anal. (C16H11
NO5) C, H, N.
-
Y. Insulin-like growth factor
I receptor primary structure:
IGF BP Ra d ioliga n d -Bin d in g Assa y. Human IGFBP-1,
BP-2, BP-4, BP-5, and BP-6 cDNA’s were used to express the
five binding proteins in the BaculoGold Expression System
(PharMingen, San Diego, CA) in Sf9 insect cells according to
the manufacturer’s instruction, and the proteins were purified
by affinity chromatography on a hIGF-II-coupled Affi-Gel 15
column, followed by reversed-phase HPLC as described.4b
IGFBP-3 was isolated from out-dated human plasma as
reported previously.4b The radioligand-binding assay was
performed at room temperature in duplicate in 0.02% NP-40/
PBS buffer, pH 7.2. Briefly, for the human IGFBP-3 binding
assay, the reaction was started by the addition of 100 µL of
Comparison with insulin receptor suggests structural determi-
nants that define functional specificity. EMBO J . 1986, 5, 2503-
2512.
(3) McInnes, C.; Sykes, B. Growth factor receptors: Structure,
mechanism, and drug discovery. Biopolymers 1997, 43, 339-
366.
(4) (a) Shimasaki, S.; Ling, N. Identification and molecular char-
acterization of insulin-like growth factor binding proteins (IG-
FBP-1, -2, -3, -4, -5, and -6). Prog. Growth Factor Res. 1992, 3,
243-266. (b) Shimasaki, S.; Shimonaka, M.; Zhang, H.; Ling,
N. Identification of five different insulin-like growth factor
binding proteins (IGFBPs) from adult rat serum and molecular
cloning of a novel IGFBP-5 in rat and human. J . Biol. Chem.
1991, 266, 10646-10653.

4010 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 23
Chen et al.
(5) Leong, S. R.; Baxter, R. C.; Camerato, T.; Dai, J .; Wood, W. I.
Structure and functional expression of the acid-labile subunit
of the insulin-like growth factor-binding protein complex. Mol.
Endocrinol. 1992, 6, 870-876.
(6) Twigg, S. M.; Baxter, R. C. Insulin-like growth factor (IGF)-
binding protein 5 forms an alternative ternary complex with
IGFs and the acid-labile subunit. J . Biol. Chem. 1998, 273,
6074-6079.
(7) Bondy, C. A.; Underwood, L. E.; Clemmons, D. R.; Guler H.-P.;
Bach, M. A.; Skarulis, M. Clinical uses of insulin-like growth
factor I. Ann. Intern. Med. 1994, 120, 593-601.
(8) Zhang, B.; Salituro, G.; Szalkowski, D.; Li, Z.; Zhang, Y.; Royo,
I.; Vilella, D.; Diez, M. T.; Pelaez, F.; Ruby, C.; Kendall, R. L.;
Mao, X.; Griffin, P.; Calaycay, J .; Zierath, J . R.; Heck, J . V.;
Smith, R. G.; Moller, D. E. Discovery of a small molecule insulin
mimetic with antidiabetic activity in mice. Science 1999, 284,
974-977.
(9) J ansson, M.; Uhlen, M.; Nilsson, B. Biochemistry 1997, 36,
4108-4117.
(10) Loddick, S. A.; Liu, X.-J .; Lu, Z.-X.; Liu, C.; Behan, D. P.;
Chalmers, D. C.; Foster, A. C.; Vale, W. W.; Ling, N.; De Souza,
E. B. Displacement of insulin-like growth factors from their
binding proteins as a potential treatment for stroke. Proc. Natl.
Acad. Sci. U.S.A. 1998, 95, 1894-1898.
(11) Lowman, H. B.; Chen, Y. M.; Skelton, N. J .; Mortensen, D. L.;
Tomlinson, E. E.; Sadick, M. D.; Robinson, I. C. A. F.; Clark, R.
Molecular mimics of insulin-like growth factor 1 (IGF-1) for
inhibiting IGF-1: IGF-binding protein interactions. Biochemistry
1998, 37, 8870-8878.
(12) Venkov, A. P.; Ivanov, I. I. Synthesis of 3-oxo-2,3-dihydroiso-
quinolines from ethyl 2-acylphenylacetates and formamides.
Synth. Commun. 1994, 24, 1145-1150.
(13) Postaire, E.; Viel, C.; Martinez, D.; Likforman, J .; Hamon, M.
Identification de Nouvelles Structures Benzoyl-1 Isoquinolein-
iques Obtenues par Oxidation de la Papaverine. Chem. Pharm.
Bull. 1987, 35, 4064-4067.
(14) (a) Riedel-de Haen German Patent, DE 674400, 1937. (b) Redel,
J .; Bouteville, A. Sur les esters des papaverine et norpapaverine
carboxylees avec le diethylaminoethanol. Bull. Soc. Chim. Fr.
1949, 443-445.
(15) Maciag, T.; Cerundolo, J .; Ilesley, S.; Kelly, P. D.; Forand, R.
An endothelial cell growth factor from bovine hypothalamus:
Identification and partial characterization. Proc. Natl. Acad. Sci.
U.S.A. 1979, 76, 5674-5678
J M010304B