Synthesis and biological evaluation of a new class of geldanamycin derivatives as potent inhibitors of Hsp90

Article abstract of DOI:10.1021/jm0306125

The heat shock protein Hsp90 has increasingly become an important therapeutic target especially for treatment of cancers. Inhibition of the ATPase activity of Hsp90 by natural products (e.g., 17-allylaminogeldanamycin or radicicol) leads to the ubiquitina

Full text of DOI:10.1021/jm0306125

J . Med. Chem. 2004, 47, 3865-3873
3865
Syn th esis a n d Biologica l Eva lu a tion of a New Cla ss of Geld a n a m ycin
Der iva tives a s P oten t In h ibitor s of Hsp 90
J ean-Yves Le Brazidec,*,§ Adeela Kamal,† David Busch,† Lia Thao,† Lin Zhang,§ Gregg Timony,‡ Roy Grecko,‡
Katy Trent,‡ Rachel Lough,‡ Tim Salazar,‡ Samina Khan,‡ Francis Burrows,† and Marcus F. Boehm§
Conforma Therapeutics Corporation, Suite 240, San Diego, California 92121
Received December 12, 2003
The heat shock protein Hsp90 has increasingly become an important therapeutic target
especially for treatment of cancers. Inhibition of the ATPase activity of Hsp90 by natural
products (e.g., 17-allylaminogeldanamycin or radicicol) leads to the ubiquitination of oncogenic
client proteins such as Her-2, Raf-1, and p-Akt followed by their proteasomal degradation.
Hsp90 inhibitors simultaneously target multiple oncogenic proteins and provide an advantage
for cancer therapy due to the potential for increased efficacy and overcoming drug resistance.
In an effort to convert geldanamycin into a druglike compound with better pharmacokinetic
properties and efficacy in human tumor xenograft models, geldanamycin was derivatized on
the 17-position to prepare new analogues such as 17-geldanamycin amides, carbamates, and
ureas and 17-arylgeldanamycins. All the compounds were first evaluated ex vivo using a cell-
based Her-2 degradation assay and in vitro using biochemical assays that measure recombinant
Hsp90 (rHsp90) competitive binding and changes in rHsp90 conformation. In addition, we
confirmed the selectivity of geldanamycin analogues for Hsp90 derived from tumor cells using
a novel cell lysate binding assay.
Sch em e 1
In tr od u ction
Deregulation of signaling proteins in cells by mutation
or overexpression results in the disruption of normal
cellular function including growth, differentiation, pro-
liferation, and ultimately survival. When key oncogenic
proteins are affected, one of the results is the formation
of cancer cells and tumor growth. Consequently, selec-
tive modulators of signaling proteins such as kinases
and nuclear receptors have become important molecular
targets for identifying new anticancer agents. This
approach has been rewarded by the discovery of several
effective modulators of cellular regulatory pathways
including Tamoxifen,1 Casodex,2 Herceptin,3 Iressa,4
and Gleevec.5
Hsp90 is an ATP-dependent molecular chaperone that
consists of two cytosolic isoforms in human: Hsp90R
and Hsp90â. In cells, Hsp90 is complexed to an array
of cochaperones (e.g., Hop, aha1, cdc37, p23) to form a
superchaperone complex that interacts with client on-
cogenic proteins involved in signal transduction, cell
cycle regulation, and apoptosis.6
Of great therapeutic interest, selective modulation of
Hsp90 is achievable either by inhibiting its ATPase
activity by natural products such as geldanamycin and
radicicol7 (Scheme 1) or, as recently described, by
inducing its direct degradation via hypericin treatment.8
Furthermore, inhibition of Hsp90 leads to degradation
of multiple key proteins that depend on the interaction
with Hsp90 for maintaining their conformation, their
stability, and their function. Simultaneously targeting
multiple oncogenic proteins provides an advantage for
cancer therapy due to the potential for increased efficacy
and overcoming drug resistance, which occurs in many
cancers including certain forms of leukemia,9 prostate,10
and breast cancers.11 For example, many early stage
chronic myelogenous leukemia (CML) patients respond
to the new kinase inhibitor Gleevec, but then they
become resistant over time. This is often due to muta-
tions in the Bcr-Abl kinase that prevent the drug from
binding. Similarly, early stage prostate cancer patients
often respond to anti-androgen therapy but develop
resistance due to mutations in the androgen receptor.
Hsp90 inhibitors selectively cause the degradation of
mutant kinases and other mutant client proteins and
thus are likely to be effective where conventional
inhibitor therapy fails. In the case of the signal trans-
duction protein Her-2/neu, a key driver of breast cancer
growth, Hsp90 inhibition leads to its ubiquitination and
* To whom correspondence should be addressed. Phone: 858-794-
8844. Fax: 858-657-0343. E-mail: jlebrazidec@conformacorp.com.
§ Department of Chemistry.
† Department of Biology.
‡ Department of Pharmacology.
10.1021/jm0306125 CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/17/2004

3866 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 15
Le Brazidec et al.
Sch em e 2a
degradation by the proteasome.12 Similarly, inhibition
of Hsp90 induces the degradation of the mutant andro-
gen receptor,13 a gene transcription factor critical to the
growth of both hormone-dependent and hormone-
independent prostate cancer. At the same time, Hsp90
inhibitors down-regulate the PI3K/Akt pathway leading
to the deactivation14 or degradation of Akt and subse-
quently reducing the expression of D-cyclins.15 This
process of deactivation also applies to other client
proteins such as tyrosine kinase FLT316 involved in
signal transduction in leukemia cells. Consequently, the
ability of Hsp90 inhibitors to degrade both normal and
mutant client proteins provides the opportunity to treat
currently incurable late-stage resistant tumors.
In 1994, Whitesell et al.17 discovered that geldana-
mycin (Scheme 1), a naturally occurring benzoquinone
ansamycin isolated from the culture broth of Strepto-
myces hygroscopicus, bound very tightly to the N-
terminal ATPase domain of rHsp90 (Kd ) 1.2 µM).18
Because geldanamycin is too chemically and metaboli-
cally unstable to become a drug, it has been derivatized,
mainly on the 17-position of the quinone moiety, leading
to a plethora of 17-alkylaminogeldanamycins including
17-allylamino-17-demethoxygeldanamycin (17-AAG). This
latter is a potent inhibitor of Hsp90, and largely as a
result of its excellent in vitro potency, the National
Cancer Institute has initiated phase I clinical trials for
advanced cancer patients. Although 17-AAG is a very
potent Hsp90 inhibitor, it also suffers from pharmaceu-
tic deficiencies including difficult formulation chal-
lenges. As a result, we investigated methods to prepare
novel analogues of geldanamycin with improved chemi-
cal and metabolic stability and increased formulation
options such as water solubility while retaining high
potency in functional cell-based assays and in vivo
models.
a
(a)Na2S2O4 10%, AcOEt’ (b) RCOCl, THF, 4 Å molecular sieves,
0 °C to room temp; (c) CuSO4, MeOH, EtN(i-Pr)2; (d) Me2NH, HCl
salt, EtN(i-Pr)2, CH2Cl2, reflux; (e) R1R2NH, THF, Nal, reflux.
to note that without the presence of base, the oxidation
was much slower. This three-step sequence gave rise
to the geldanamycin amides 1a -k with moderate to
good yields. Treatment of the geldanamycin analogue
1d with a series of secondary amines under standard
conditions gave rise to geldanamycin amides 2a -j
(Scheme 2). Replacing the acyl chloride by either an
alkyl chloroformate or an alkyl isocyanate in the second
step of the sequence gave access to the geldanamycin
carbamates 4a -d and the geldanamycin urea 5 with
good yields (Scheme 3). For the synthesis of the 17-
arylgeldanamycins, we first hydrolyzed geldanamycin
and then, using standard conditions, converted the
resulting 17-hydroxy-17-demethoxygeldanamycin into
the triflate 6, which is a stable yellow solid (Scheme 4).
Several conditions were attempted for the Suzuki
coupling, and we obtained the best results using Neel’s
conditions,23 which afforded the 17-arylgeldanamycins
7a ,b in moderate yields.
Ch em istr y
Geldanamycin is a member of the ansamycin family
of natural products and displays two key structural
features, the quinone moiety and the ansa ring, which
are responsible, via an induced fit,19 for its binding to
the N-terminal ATPase domain of Hsp90. The goal of
our synthetic program was to improve the pharmaceuti-
cal properties20 and chemical stability of geldanamycin
by modifying the quinone moiety. Recent studies of the
redox potential of mitomycins21 show that it is possible
to tune the electronic properties of the quinone by
modifying substituents in various positions of the ring.
Because synthetic geldanamycin analogues having
amides, carbamates, or carbon-carbon bonds attached
directly to the 17-position had never been described, we
directed our effort toward their synthesis. Following a
reported procedure,22 geldanamycin was treated with
ammonia in methanol to afford 17-amino-17-demethoxy-
geldanamycin (17-AG). Because of the deactivating
effect of the quinone, this latter compound was unre-
active toward N-acylation using standard procedures.
We were able to circumvent this problem by using a
three-step route: (1) reduction of the quinone moiety
of 17-AG with 10% sodium thionite, (2) addition of the
desired acyl chloride in the presence of molecular sieves,
and (3) air oxidation of the dihydroquinone mediated
by CuSO4 and Hunig’s base (Scheme 2). It is of interest
Biologica l Eva lu a tion
The in vitro biological activity of the new geldana-
mycin analogues was evaluated using a human epider-
mal growth factor 2 (Her-2) degradation assay, a
competitive rHsp90 binding assay, a cell lysate binding
assay, a bis-ANS conformational assay, and a cytotoxic
assay (MTS). The Her-2 degradation assay measures the
degradation of Her-2 in MCF7 tumor cells resulting
from inhibition of the ATPase activity of Hsp90. The
competitive binding assays utilized the biotinylated

Geldanamycin Derivatives
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 15 3867
Ta ble 1. Activity of Geldanamycin Amides
Sch em e 3a
b-ANS
Her-2
rHsp90
rHsp90
compd
R
IC50 (nM)a,b IC50 (µM)a,c IC50 (µM)a,d
1a
1b
1c
1d
phenyl
2-furyl
2-thienyl
p-chloromethyl-
phenyl
m-chloromethyl-
phenyl
200
500
400
550
3
4
3
2
4
4
2
4
1e
1000
10
NDe
a
1f
1g
1h
1i
1j
1k
acetyl
o-anisyl
m-anisyl
p-anisyl
p-nitrophenyl
p-fluorophenyl
1700
180
180
400
300
250
20
1
0.9
4
2
15
1.1
1.3
5
3.5
1.5
(a) Na2S2O4, 10%, AcOEt; (b) ROCOCl, THF, 4 Å molecular
sieves, 0 °C to room temp; (c) CuSO4, MeOH, EtN(i-Pr)2; (d) allyl
isocyanate, THF, 0 °C to room temp.
Sch em e 4a
1.5
a
b
Values correspond to n ) 2. Standard errors in the range of
15-20%. c Standard errors in the range of 15-20%. Standard
d
errors in the range of 35-40%. e ND ) not determined. For assay
conditions, see Experimental Section.
changes of Grp94,24 an endoplasmic reticulum homo-
logue of cytosolic Hsp90. Finally, selected analogues
were evaluated in mouse xenograft models of human
tumors for pharmacokinetic and efficacy studies.
Resu lts a n d Discu ssion
The primary assay for evaluating the structure-
activity relationships of new geldanamycin derivatives
was the Her-2 degradation assay in intact MCF7 cells
using a flow cytometric assay based on fluorescence-
activated cell sorter technology (FACS). By use of 17-
AAG as a reference (IC50 ) 15 ( 4 nM)25 in this
functional assay, the screening of the first series of
geldanamycin amides revealed a decrease in activities
ranging from 180 to 1700 nM (Table 1). Among those
amides, aromatic derivatives such as 1a , 1d , 1e, 1j, and
1k exhibited better potencies than their aliphatic coun-
terpart 1f; furthermore, changing the size and the
polarity of the aromatic ring from phenyl to furyl and
thienyl (1b, 1c) led to a 2-fold drop in activity. A study
of the substituent effects showed that an electron-
donating group on the ortho or meta position (1g, 1h )
was better than on the para position (1i). For analogues
1d and 1e, the chloromethyl group reverses the position
effect with the para isomer being twice as active as the
meta isomer. On the other hand, electron-withdrawing
groups on the para position (1j, 1k ) maintain the
activity at 250 nM. To improve the formulation proper-
ties of the geldanamycin amides, we prepared a second
series of geldanamycin analogues by adding a dialkyl-
aminomethyl group on the para position of the amide
1a . The addition of an ionizable amino group provided
water solubility via the salt form and was more readily
formulatable for intravenous administration (free base
solubility range, 10-20 µg/mL; salt solubility range,
5-40 mg/mL). As exemplified by compound 2a (Table
2), addition of a dimethylamino methyl group led to a
a
(a) Ba(OH)2, THF/H2O, 65 °C; (b) Tf2O, (i-Pr)2NEt, CH2Cl2;
(c) ArB(OH)2, Pd2dba3, CsBr, CsF, dioxane.
Sch em e 5a
a
(a) 5-(Biotinamido)pentylamine, THF, room temp.
geldanamycin derivative 8, which was prepared in one
step from 5-(biotinamido)pentylamine and geldanamy-
cin (Scheme 5). The 1,1′-bis(4-anilino-5-naphthalene-
sulfonic acid) (bis-ANS) is an environmentally sensitive
fluorophore that becomes fluorescent when it binds and
induces a conformational change in rHsp90. This assay
has been previously used to look at the conformational

3868 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 15
Ta ble 2. Activity of Water-Soluble Geldanamycin Amides
Le Brazidec et al.
Her-2
rHsp90
b-ANS rHsp90
compd
R1R2N
dimethylamino
1-(4-(4-chlorophenyl))piperazinyl
1-(4-(4-methoxyphenyl))piperazinyl
1-(4-(4-nitrophenyl))piperazinyl
1-(4-methyl)piperazinyl
1-(4-cyclohexyl)piperazinyl
piperidinyl
IC50 (nM)a
IC50 (µM)a
IC50 (µM)a
2a
2b
2c
2d
2e
2f
2g
2h
2i
1700
250
220
200
800
3.0
1.0
1.5
1.3
8.0
5.8
2.1
1.2
1.1
1.5
17.0
1.0
3.0
1.0
5.0
5.5
2.5
3.0
1.0
1.0
600
>10000
dipropylamino
benzylethylamino
500
140
200
2j
1,2,3,4-tetrahydroisoquinolinyl
a
Values correspond to n ) 2. For assay conditions, see Experimental Section.
Ta ble 3. Activity of Geldanamycin Carbamates and Ureas
Her-2
rHsp90
b-ANS rHsp90
compd
R
X
IC50 (nM)a
IC50 (µM)a
IC50 (µM)a
4a
4b
4c
4d
5
ethyl
p-acetoxybenzyl
isobutyl
4-(N-morpholino)butyl
allyl
O
O
O
O
260
50
120
3.0
2.5
4.0
1.0
NDb
3.0
1.0
2.0
1.0
35.0
500
NH
>10000
a
b
Values correspond to n ) 2. ND ) not determined. For assay conditions, see Experimental Section.
Ta ble 4. Activity of 17-Arylgeldanamycins
9-fold decrease in activity. The screening of various
tertiary amines showed that the compounds with ben-
zylalkylamino groups (2i) were three times more active
than the analogues with dialkylamino groups (2h ).
Moreover, in the cyclic amine series, the 4-arylpipera-
zine derivatives 2b-d were much more potent than the
piperidine derivative 2g. In a second SAR study, we
examined geldanamycin carbamate derivatives and
found their activities to be within the same range of
potencies as the amides. Carbamate 4b (Table 3) was
the most active derivative due to release of 17-AG in
the cell via hydrolysis followed by 1,6-elimination of the
quinone methide moiety.26 Compound 4d was 4-fold less
active than 4c, showing a detrimental effect of both the
extension of the alkyl chain and the addition of a high
pKa tertiary amine such as a morpholino group. In
comparison to the amide series for which 17-alkylamido
groups were only weakly active, the alkyl carbamates
were very potent. For the aryl carbamates, no data were
produced because of their chemical instability that
prevented them from being successfully isolated. For
reasons that are still unclear, the 17-urea-geldanamycin
derivatives were completely inactive in this assay. 17-
Arylgeldanamycins 7a ,b are approximatively 2-fold less
active than the best amide 2i in the Her-2 degradation
Her-2
rHsp90
b-ANS rHsp90
compd
Ar
IC50 (nM)a
IC50 (µM)a
IC50 (µM)a
7a
7b
phenyl
2-thienyl
280
250
3
1.7
16.0
10.0
a
Values correspond to n ) 2. For assay conditions, see Experi-
mental Section.
assay, indicating that replacing the amide by an aryl
group on the 17-postion of geldanamycin affects the
activity only moderately (Table 4).
Since the Her-2 degradation assay provides an indi-
rect readout of Hsp90 inhibition, namely, the degrada-
tion of Her-2, and indicates that multiple factors are
associated with a cell-based assay such as the amount
of cell penetration, it was important to confirm the
compound bioactivity with biochemical assays that

Geldanamycin Derivatives
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 15 3869
Ta ble 5. Cell Lysate Hsp90 Binding of 17-AAG,
17-Arylgeldanamycin, and Geldanamycin Amides and
Carbamates
Ta ble 6. MTS Assay of Geldanamycin Amides and
Carbamates
compd
MCF7 cells IC50 (µM)a
cell lysate Hsp90
IC50 (nM)a
(rHsp90 IC50)/
(cell lysate Hsp90 IC50)
2a
2b
2f
2h
2i
>30
0.15
>30
0.5
compd
17-AAG
1c
2a
2b
2e
2f
2i
4c
4d
20
40
15
9
120
300
12
3
40
75
200
111
67
19
92
1333
100
43
0.2
4b
4c
0.04
0.18
a
Values correspond to n ) 2. Standard errors are in the range
of 40-45%. For assay conditions, see Experimental Section.
10
70
55
7a
7b
31
a
Values correspond to n ) 2. Standard errors are in the range
of 20-25%. For assay conditions, see Experimental Section.
directly measure their interaction with Hsp90. For this
purpose, recombinant Hsp90 (rHsp90) competitive bind-
ing and bis-ANS conformational assays were used to
further screen our compounds and allowed us to cluster
them by relative binding activity. Very active com-
pounds such as 1c,g,h ,k and 2c,d ,i,j bind tightly to
rHsp90 and efficiently inhibit the rHsp90 conforma-
tional change with IC50 values between 1 and 3 µM. In
contrast, compounds that are less active or inactive in
the Her-2 degradation assay, such as 1b,e and 2a ,e,
have IC50 values between 4 and 20 µM. Nevertheless,
the 17-arylgeldanamycins seem to behave differently by
performing very poorly in the bis-ANS assay while
maintaining good IC50 values in the recombinant Hsp90
binding assay. Although the binding and conformational
activities tracked with the results of the Her-2 degrada-
tion assay, the potency gap between those assays as
exemplified for 2i (Table 2; Her-2, IC50 ) 140 nM;
rHsp90 binding, IC50 ) 1100 nM; bis-ANS, IC50 ) 1000
nM) remained to be explained. Also observed for 17-AAG
(rHsp90 binding, IC50 ) 800 nM; bis-ANS, IC50 ) 700
nM), this discrepancy was recently clarified by showing
that the Hsp90 from cell lysates of cancer cells such as
MCF7 have 100-fold higher affinity for 17-AAG than
does recombinant Hsp90.27 This increased Hsp90 bind-
ing affinity was also observed for our compounds in
MCF7 cell lysates (Table 5) with IC50 values ranging
from 9 to 300 nM for the 17-amide-geldanamycin series,
from 3 to 10 nM for the 17-carbamate-geldanamycin
series, and from 55 to 70 nM for the 17-arylgeldana-
mycin series. Before trying to rationalize this affinity
increase, we first want to emphasize that the high-
affinity Hsp90 in the cell lysate is fully complexed to
its cochaperones and that the 100-fold affinity increase
occurs only for geldanamycin derivatives and does not
occur for other Hsp90 inhibitors such as radicicol (Kd )
19 nM18). As shown in a recent study,28 geldanamycin
requires major conformational changes such as isomer-
ization of the amide and rotation of the quinone ring in
order to adopt the optimal C-shape conformation.29 The
presence of cochaperones complexed to Hsp90 may
facilitate the conformational changes of geldanamycin
from the extended to the C-shape conformation by
optimizing the interactions between Hsp90 and geldana-
mycin at each step of this enthropically disfavored
process.
F igu r e 1. Serum concentrations of 2i following iv and oral
administration to balb/c mice.
(Table 6). For the most potent derivatives such as 2b,
2h , 2i, 4b, and 4c, there is a good correlation between
their activity in the Her-2 degradation assay (Tables 2
and 3) and in the MTS assay (Table 6). In contrast, the
weaker compounds such as 2a and 2f do not induce
sufficient degradation of key clients such as Her-2 to
be cytotoxic in MCF7 cells.
Because one of our goals was to identify geldanamycin
derivatives with improved pharmaceutic properties, we
evaluated several analogues in vivo. The geldanamycin
analogue 2i was one of the most potent in the in vitro
assays, and its pharmacokinetic properties were deter-
mined in mice. Following iv administration (Figure 1),
2i exhibited a high total clearance (79 (mL/min)/kg), a
large steady-state volume of distribution (1878 mL/kg),
and a terminal half-life of 37 min. These pharmacoki-
netic parameters are similar to those reported for 17-
AAG (microdispersion) administered intravenously to
mice.30 After oral administration, 2i was detectable in
two of three animals at the first sample time (0.5 h) and
was below the limit of quantitation in all animals
thereafter. We anticipated that the relatively high iv
clearance rate of 2i was likely due to metabolism, and
thus, low bioavailability was anticipated for this com-
pound. However, significant metabolite levels of 2i were
not observed in mouse serum following either iv or oral
routes of administration. Compounds 2i and 2b were
subsequently tested for antitumor activity in several
human tumor xenograft models in mice using ip ad-
ministration. Compound 2b was tested against two lung
cancers, the small-cell H69 xenograft (Figure 2a) and
the non-small-cell A549 xenograft (Figure 2c). Com-
pound 2i was tested on the glioblastoma U87 xenograft
(Figure 2b) and in the A549 xenograft. Each of the
tumors used is dependent on one or more Hsp90 client
proteins including Raf-1 (A549), IGF-1 receptor (H69),
or Akt (U87). As shown in Figure 2, both compounds
displayed potent antitumor activity in each model
tested. Compounds 2b and 2i were well tolerated when
given at low doses daily (parts a and b of Figure 2), at
intermediate doses three times per week (Figure 2c),
Compounds were also screened in the MTS assay in
order to determine their cytotoxicity to cancer cells

3870 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 15
Le Brazidec et al.
This study also demonstrates that converting 17-ami-
nogeldanamycin into the corresponding amide such as
2i leads to a 10-fold decrease in activity in the Her-2
assay without affecting the cell lysate Hsp90 binding
(Table 5), which correlates well with the efficacy data
in animal models.
The high selectivity of geldanamycins for tumor cell
Hsp90 could confer to this class of compounds a unique
status among Hsp90 inhibitors used in oncology. In
addition, for other diseases such as viral diseases,
autoimmune disorders, and inflammation, for which the
cells are highly dependent on Hsp90 for survival, similar
selectivity could be observed.
Exp er im en ta l Section
Ch em istr y. All reactions were carried out with continuous
stirring under an atmosphere of nitrogen. Commercial re-
agents and solvents were used as received without further
purification or drying. 1H NMR were obtained using a Bruker-
400 spectrometer. Chemical shifts are reported in parts per
million relative to tetramethylsilane as internal standard.
Thin-layer chromatography (TLC) was performed on 250 µM
silica gel plates (Whatman 4861-820). Flash chromatography
was run using EM Science silica gel (230-400 mesh). Low-
resolution mass spectra were run using a ThermoFinnigan
LCQ spectrometer. High-resolution mass spectra were run at
the mass spectrometry facility of UC Riverside using a PE
BIOSYSTEMS DE-STR MALDI TOF spectrometer. Melting
points were measured using an electrothermal MEL-TEMP
apparatus and are reported uncorrected. Compound purity was
determined by HPLC analysis using a C18 reverse-phase
column with an Agilent 1100 series system attached to a
Hewlett-Packard chromatograph manager.
Gen er a l P r oced u r e for th e Syn th esis of Geld a n a m ycin
Am id e Is Exem p lified for 17-Ben zoyla m in o-17-d em eth -
oxygeld a n a m ycin 1a . A solution of 17-amino-17-demethoxy-
geldanamycin (3.1 g, 5.62 mmol) in EtOAc (400 mL) was
treated with Na2S2O4 (10%, 60 mL) at room temperature. After
2 h, the aqueous layer was extracted with EtOAc (100 mL)
and the combined organic layers were dried over Na2SO4 and
concentrated under reduced pressure to give 18,21-dihydro-
17-amino17-demethoxygeldanamycin as a yellow solid. This
solid was dissolved in anhydrous THF (70 mL) and transferred
via cannula to a mixture of benzoyl chloride (0.65 mL, 5.62
mmol) and 4 Å molecular sieves (4.5 g) in THF (20 mL) at 0
°C. After 12 h at room temperature, EtN(i-Pr)2 (1.1 mL), CuSO4
(0.44 g, 2.81 mmol), and MeOH (8 mL) were further added to
the reaction mixture. After 3 h, the reaction mixture was
filtered and concentrated under reduced pressure. The crude
material was dissolved in EtOAc (300 mL) and washed with 2
N HCl (50 mL) and saturated NaHCO3 (50 mL), dried over
Na2SO4, and concentrated under reduced pressure to give the
crude material, which was purified by recrystallization from
EtOH to give 17-benzoyl-amino-17-demethoxygeldanamycin as
a yellow solid (3.0 g, 82%). Rf ) 0.50 in 80:15:5 CH2Cl2/EtOAc/
MeOH. Mp ) 218-220 °C. HPLC purity: 99.9%. 1H NMR
CDCl3 δ 0.92-0.94 (m, 6H, C10-Me, C14-Me), 1.70 (br s, 2H,
2xH13), 1.79 (br s, 4H, H14 and C8-Me), 2.03 (s, 3H, C2-Me),
2.56 (dd, 1H, J ) 8.7, 13.9 Hz, H15), 2.64 (dd, 1H, J ) 4.1,
14.1 Hz, H15), 2.76-2.79 (m, 1H, H10), 3.33 (br s, 7H, 2 ×
OMe, H11), 3.44-3.46 (m, 1H, H12), 4.325 (d, 1H, J ) 9.0 Hz,
H6), 5.16 (s, 1H, H7), 5.77 (d, 1H, J ) 9.3 Hz, H9), 5.91 (t, 1H,
J ) 9.9 Hz, H5), 6.57 (t, 1H, J ) 11.2 Hz, H4), 6.94 (d, 1H, J
) 11.4 Hz, H3), 7.48 (s, 1H, H19), 7.52 (t, 2H, J ) 7.4 Hz, Ph),
7.62 (t, 1H, J ) 7.2 Hz, Ph), 7.91 (d, 2H, J ) 7.3 Hz, Ph), 8.47
(s, 1H, NHCO), 8.77 (s, 1H, N22-H). MS m/z 672.5 (M+ + Na).
HRMS calculated for C35H43N3O9Na (M+ + Na): 672.2897.
Found 672.2903.
F igu r e 2. (a) Efficacy of 2b in H69 small-cell lung cancer:
(diamonds) PETA vehicle; (circles) 2b 20 mg/kg, 7×/week, 4
weeks; (squares) 2b 30 mg/kg, 7×/week, 4 weeks. All the
compounds were administered ip. (b) Efficacy of 2i in U87
glioblastoma: (diamonds) PETA vehicle; (squares) 2i 25 mg/
kg, 7×/week, 18 days; (circles) 2i 75 mg/kg, 2×/week, 18 days.
All the compounds were administered ip. (c) Efficacy of 2b and
2i in A549 lung cancer: (diamonds) PETA vehicle; (squares)
2b 40 mg/kg, 3×/week, 4 weeks; (circles) 2i 40 mg/kg, 3×/week,
4 weeks. All the compounds were administered ip.
or at high doses twice weekly (Figure 2b). The lack of
schedule dependence along with the pharmacokinetic
properties exhibited by both compounds suggest a fast
diffusion into the tissues and, particularly, accumulation
in solid tumors in mice (data not shown) owing to
enhanced retention of ansamycins to the high-affinity
form of Hsp90.
Con clu sion
During this study, geldanamycin has been derivatized
on the 17-position to produce new analogues with
common pharmacophores such as amides, ureas, car-
bamates, and aryl functional groups. After ex vivo
optimization of their activity using the Her-2 degrada-
tion assay, we were able to refine the data using the
bis-ANS and recombinant Hsp90 competitive binding
assays. More importantly, it was confirmed that tumor
cell lysate Hsp90 has a much higher affinity for this new
series of geldanamycin analogues than recombinant
Hsp90 does, thus explaining the apparent discrepancy
between the binding and the Her-2 degradation assays.
Gen er a l P r oced u r e for th e Syn th esis of 17-(4-Dia lk yl-
a m in om eth ylben zoyl)-17-a m in o-17-d em eth oxygeld a n a -
m ycin Is Exem p lified for 17-(4-Dim eth yla m in om eth yl-
ben zoyl)-17-a m in o-17-d em eth oxygeld a n a m ycin 2a . To a

Geldanamycin Derivatives
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 15 3871
solution of 1d (3.73 g, 5.3 mmol) in CH2Cl2 (50 mL) were added
Me2NH‚HCl salt (1.51 g, 18.5 mmol), Et(i-Pr)2N (2.8 mL), and
NaI (0.79 g, 5.3 mmol). The resulting solution was heated at
reflux for 6 h, whereupon it was cooled to room temperature,
diluted with EtOAc (600 mL), washed with water (100 mL),
dried with Na2SO4, and concentrated under reduced pressure
to give the crude material, which was purified by recrystalli-
zation in EtOH to give 17-(4-dimethylaminomethylbenzoyl)-
17-amino-17-demethoxygeldanamycin as a yellow solid (1.74
g, 46%). Rf ) 0.10 in 80:15:5 CH2Cl2/EtOAc/MeOH. Mp ) 203-
3-(17-geldanamycinyl)urea as a brick solid (5 mg, 22%). Rf )
0.38 in 80:15:5 CH2Cl2/EtOAc/MeOH. HPLC purity: 98.0%.
1H NMR CDCl3 δ 0.92 (d, 3H, J ) 6.0 Hz, C10-Me or C14-
Me), 0.95 (d, 3H, J ) 6.9 Hz, C10-Me or C14-Me), 1.61-1.72
(m, 3H, 2 × H13, H14), 1.78 (s, 3H, C8-Me), 2.03 (s, 3H, C2-
Me), 2.55 (m, 2H, 2 × H15), 2.74-2.78 (m, 1H, H10), 3.31 (s,
3H, OMe), 3.34 (s, 3H, OMe), 3.31-3.34 (m, 1H, H11), 3.43-
3.45 (m, 1H, H12), 3.90 (br s, 2H, NHCH2), 4.31 (d, 1H, J )
9.2 Hz, H6), 5.16 (s, 1H, H7), 5.20-5.29 (m, 3H, CHCH2, NH),
5.79 (d, 1H, J ) 9.4 Hz, H9), 5.84-5.94 (m, 2H, H5, CHCH2),
6.57 (t, 1H, J ) 11.8 Hz, H4), 6.92 (d, 1H, J ) 11.4 Hz, H3),
7.28 (s, 1H, NHCO), 7.41 (s, 1H, H19), 8.79 (s, 1H, N22-H).
MS m/z 731.7 (M+ + Na). HRMS calculated for C32H44N4O9-
Na (M+ + Na): 651.3006. Found 651.2981.
1
206 °C. HPLC purity: 99.9%. H NMR CDCl3 δ 0.93 (d, 3H, J
) 6.5 Hz, C10-Me or C14-Me), 0.95 (d, 3H, J ) 6.5 Hz, C10-
Me or C14-Me), 1.70-1.74 (m, 3H, 2 × H13, H14), 1.79 (s, 3H,
C8-Me), 2.04 (s, 3H, C2-Me), 2.29 (s, 6H, N(Me)2), 2.53-2.59
(m, 1H, H15), 2.63-2.67 (m, 1H, H15), 2.75-2.79 (m, 1H, H10),
3.33 (br s, 7H, 2 × OMe, H11), 3.46 (br s, 1H, H12), 3.54 (s,
2H, NCH2Ar), 4.33 (d, 1H, J ) 9.1 Hz, H6), 5.16 (s, 1H, H7),
5.77 (d, 1H, J ) 9.5 Hz, H9), 5.91 (t, 1H, J ) 10.0 Hz, H5),
6.57 (t, 1H, J ) 11.3 Hz, H4), 6.94 (d, 1H, J ) 11.1 Hz, H3),
7.49 (s, 1H, H19), 7.49 (d, 2H, J ) 8.0 Hz, Ar), 7.87 (d, 2H, J
) 8.2 Hz, Ar), 8.47 (s, 1H, NHCO), 8.77 (s, 1H, N22-H). MS
Gen er al P r ocedu r e for th e Syn th esis of 17-Ar ylgeldan a-
m ycin Is Exem p lified for 17-P h en yl-17-d em eth oxygel-
d a n a m ycin 7a . A solution of 17-OTf-geldanamycin 6 (0.20 g,
0.30 mmol), cesium bromide (128 mg, 0.60 mmol), cesium
fluoride (91 mg, 0.60 mmol), Pd(dba)2 (43 mg, 0.075 mmol),
and phenylboronic acid (73 mg, 0.60 mmol) in dioxane was
heated at 40 °C for 12 h, whereupon it was cooled to room
temperature and concentrated under reduced pressure to give
the crude product. This solid was dissolved in EtOAc and
washed with saturated NaHCO3, dried over Na2SO4, and
concentrated under reduced pressure. The crude material was
purified by flash chromatography to give 17-phenylgeldana-
mycin as a yellow solid (90 mg, 50%). Rf ) 0.49 in 80:15:5 CH2-
Cl2/EtOAc/MeOH. HPLC purity: 98.0%. 1H NMR (CDCl3) δ
0.71 (d, 3H, J ) 6.5 Hz, C10-Me or C14-Me), 0.98 (d, 3H, J )
6.9 Hz, C10-Me or C14-Me), 1.53-1.63 (m, 3H, 2 × H13, H14),
1.79 (s, 3H, C8-Me), 2.04 (s, 3H, C2-Me), 2.35 (dd, 1H, J )
6.5, 13.2 Hz, H15), 2.54-2.60 (m, 1H, H15), 2.73-2.77 (m, 1H,
H10), 3.31 (s, 3H, OMe), 3.34 (s, 3H, OMe), 3.35-3.37 (m, 1H,
H11), 3.47-3.52 (m, 1H, H12), 4.36 (d, 1H, J ) 8.9 Hz, H6),
5.24 (s, 1H, H7), 5.75 (d, 1H, J ) 9.4 Hz, H9), 5.90 (t, 1H, J )
10.0 Hz, H5), 6.58 (t, 1H, J ) 11.5 Hz, H4), 6.96 (d, 1H, J )
11.4 Hz, H3), 7.12-7.14 (m, 2H, Ph), 7.40-7.43 (m, 3H, Ph),
m/z 707.7 (M+ + H). HRMS calculated for C38H51N4O9 (M+
H): 707.3656. Found 707.3631.
+
Gen er a l P r oced u r e for th e Syn th esis of Geld a n a m ycin
Ca r ba m a tes Is Exem p lified for 17-(Eth oxyca r bon yl)-
a m in o-17-d em eth oxygeld a n a m ycin 4a . A solution of 17-
amino-17-demethoxygeldanamycin (0.6 g, 1.10 mmol) in EtOAc
(120 mL) was treated with Na2S2O4 (10%, 10 mL) at room
temperature. After 2 h, the aqueous layer was extracted twice
with EtOAc and the combined organic layers were dried over
Na2SO4 and concentrated under reduced pressure to give
18,21-dihydro-17-amino-17-demethoxygeldanamycin as a yel-
low solid. This solid was dissolved in anhydrous THF (5 mL)
and transferred via cannula to a mixture of ethyl chloroformate
(0.11 mL, 1.15 mmol) and 4 Å molecular sieves (1.2 g) in THF
(3 mL) at 0 °C. After 12 h at room temperature, EtN(i-Pr)2
(0.38 mL), CuSO4 (60 mg, 0.37 mmol), and MeOH (2 mL) were
further added to the reaction mixture. After 2 h, the reaction
mixture was filtered and concentrated under reduced pressure.
The crude material was dissolved in EtOAc (100 mL) and
washed with 2 N HCl (10 mL) and saturated NaHCO3 (10 mL),
dried over Na2SO4, and concentrated under reduced pressure
to give the crude material, which was purified by flash
chromatography to give 17-ethoxycarbonyl-17-amino-17-de-
methoxygeldanamycin as a yellow solid (0.25 g, 40%). Rf ) 0.31
in 80:15:5 CH2Cl2/EtOAc/MeOH. Mp ) 242-245 °C. HPLC
purity: 99.9%. 1H NMR CDCl3 δ 0.93 (d, 3H, J ) 6.0 Hz, C10-
Me or C14-Me), 0.95 (d, 3H, J ) 6.9 Hz, C10-Me or C14-Me),
1.32 (t, 3H, J ) 7.2 Hz, OCH2-Me), 1.62-1.74 (m, 3H, 2 × H13,
H14), 1.78 (s, 3H, C8-Me), 2.02 (s, 3H, C2-Me), 2.58 (m, 2H, 2
× H15), 2.76 (m, 1H, H10), 3.32 (s, 3H, OMe), 3.34 (s, 3H,
OMe), 3.31-3.34 (m, 1H, H11), 3.42-3.44 (m, 1H, H12), 4.23
(q, 2H, J ) 7.0 Hz, OCH2-Me), 4.32 (d, 1H, J ) 9.0 Hz, H6),
5.16 (s, 1H, H7), 5.74 (d, 1H, J ) 9.2 Hz, H9), 5.90 (t, 1H, J )
10.1 Hz, H5), 6.56 (t, 1H, J ) 11.4 Hz, H4), 6.91 (d, 1H, J )
11.6 Hz, H3), 7.44 (s, 1H, H19), 8.73 (s, 1H, N22-H). MS m/z
7.56 (s, 1H, H19), 8.68 (s, 1H, N22-H). MS m/z 629.6 (M+
+
Na). HRMS calculated for C34H42N2O8Na (M+ + Na): 629.2839.
Found 629.2830.
17-(5-(Biot in a m id o)p en t yl)a m in o-17-d em et h oxygel-
d a n a m ycin 8. To 50 mg (0.152 mmol) of 5-(biotinamido)-
pentylamine 4 in 3 mL of 15:1 THF/H2O was added 28 mg
(0.050 mmol) of geldanamycin 2 at room temperature. The
reaction mixture was stirred overnight, quenched with water
(50 mL), and extracted with 2 × 50 mL of EtOAc. The EtOAc
extracts were combined, washed with 2 × 50 mL of H2O and
1 × 50 mL of brine, dried (MgSO4), and purified by silica gel
flash chromatography to give 100 mg (0.114 mmol) of 8 in 75%
yield. Mp ) 143-147 °C. HPLC purity: 99.9%. 1H NMR
(CDCl3) δ 0.99 (d, J ) 7.0 Hz, 3H, C10-Me or C14-Me), 1.04
(d, J ) 7.6 Hz, 3H, C10-Me or C14-Me), 1.46 (m, 4H, 2 × CH2),
1.48 (m, 4H, 2 × CH2), 1.72 (m, 6H, 3 × CH2), 1.79 (m, 1H,
CH), 1.80 (s, 3H, CH3), 2.05 (s, 3H, CH3), 2.26 (m, 4H, 2 ×
CH2), 2.45 (d, 1H, CH2), 2.65 (d, J ) 13.1 Hz, 1H, CH2), 2.77
(m, 2H, CH + CH2), 2.95 (dd, J ) 12.8, 5.0 Hz, 1H, CH2), 3.20
(m, 2H, CH2), 3.34 (s, 3H, OCH3), 3.36 (s, 3H, OCH3), 3.48 (t,
J ) 13.2 Hz, 2H, CH2), 3.58 (m, 1H, CH), 4.30 (m, 2H, CH +
NH), 4.50 (t, J ) 7.6 Hz, 1H, CH), 4.92 (bs, 2H, 2CH), 5.21 (s,
1H, CH), 5.89 (m, 2H, 2CHd), 6.00 (bs, 2H, 2NH), 6.25 (bs,
1H, NH), 6.66 (t, J ) 11.6 Hz, 1H, CHd), 7.00 (d, J ) 11.6
Hz, 1H, CHd), 7.30 (s, 1H, CHd), 9.20 (s, 1H, CONH). MS
m/z 880.2 (M+ + Na). HRMS calculated for C43H64N6O10SNa
(M+ + Na): 879.4302. Found 879.4296.
r Hsp 90 Com p etitive Bin d in g Assa y. An amount of 5 µg
of purified rHsp90 protein (Stressgen, BC, Canada, No. SPP-
770) in phosphated buffered saline (PBS) was coated on 96-
well plates by incubating overnight at 4 °C. Unbound protein
was removed, and the coated wells were washed twice with
200 µL of PBS. DMSO controls (considered as untreated
samples) or test compounds were then added at 100, 30, 10,
3, 1, and 0.3 µM dilutions (in PBS), and samples in the plates
were mixed for 30 s on a plate shaker and then incubated for
60 min at 37 °C. The wells were washed twice with 200 µL of
640.5 (M+ + Na). HRMS calculated for C31H43N3O10Na (M+
Na): 640.2846. Found 640.2830.
+
Gen er a l P r oced u r e for th e Syn th esis of Geld a n a m ycin
Ur ea s Is Exem p lified for 17-(Allyla m in oca r bon yl)a m in o-
17-d em eth oxygeld a n a m ycin 5. A solution of 17-amino-17-
demethoxygeldanamycin (20 mg, 0.036 mmol) in EtOAc (10
mL) was treated with Na2S2O4 (10%, 0.5 mL) at room tem-
perature. After 2 h, the aqueous layer was extracted twice with
EtOAc and the combined organic layers were dried over 18,21-
dihydro-17-amino-17-demethoxygeldanamycin as a brown solid.
This solid was dissolved in anhydrous THF (2 mL) under
nitrogen atmosphere, and to the resulting solution was added
allyl isocyanate (64 µL, 0.072 mmol) at room temperature.
After 1 h, the solvent was removed under reduced pressure
whereby the residue was dissolved in MeOH and the mixture
was stirred overnight in the presence of silica. After filtration,
the solvent was removed under reduced pressure and the
residue was purified by flash chromatography to give 1-allyl-

3872 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 15
Le Brazidec et al.
PBS, and 10 µM biotinylated geldanamycin (biotin-GM) was
added and incubated for 60 min at 37 °C. The wells were
washed again twice with 200 µL of PBS before the addition of
20 µg/mL streptavidin-phycoerythrin (Molecular Probes, Eu-
gene, OR) and incubation for 60 min at 37 °C. The wells were
washed again twice with 200 µL of PBS. Relative fluorescence
units (RFU) was measured using a SpectraMax Gemini XS
spectrofluorometer (Molecular Devices, Sunnyvale, CA) with
an excitation of 485 nm and emission of 580 nm; data were
acquired using SOFTmaxPRO software (Molecular Devices
Corporation, Sunnyvale, CA). The background was defined as
the RFU generated from wells that were not coated with Hsp90
but were treated with the biotin-GM and streptavidin-PE.
Percent inhibition of binding for each sample was calculated
from the background-subtracted values as follows:
[KLH] (Becton Dickinson, No. 340761) control antibody was
added at a dilution of 1:20 or 1:40, respectively (final concen-
tration was 1 µg/mL), and the cells were pipeted up and down
to form a single-cell suspension and incubated for 15 min. The
cells were washed twice with 200 µL of BA buffer, resuspended
in 200 µL of BA buffer, and transferred to FACSCAN tubes
with an additional 250 µL of BA buffer. Samples were analyzed
using an FACSCalibur flow cytometer (Becton Dickinson, San
J ose, CA) equipped with an argon-ion laser that emits 15 mW
of 488 nm light for excitation of the phycoerythrin fluoro-
chrome; 10 000 events were collected per sample. A fluores-
cence histogram was generated, and the mean fluorescence
intensity (MFI) of each sample was determined using Cellquest
software. The background was defined as the MFI generated
from cells incubated with control IgG-PE and was subtracted
from each sample stained with the Her-2/neu antibody. Cells
incubated with DMSO were always done as untreated controls
because the compounds were resuspended in DMSO. The
percent degradation of Her-2 was calculated as follows:
% binding inhibition )
(untreated RFU) - (sample RFU)
× 100
untreated RFU
MFI Her-2 sample
MFI Her-2 untreated cells
IC50 was defined as the concentration of the compound at
which there was 50% inhibition of the biotin-GM binding to
rHsp90.
% Her-2 degradation )
× 100
r Hsp 90 Con for m a tion a l Assa y. This was done by modi-
fication of a previously published method that utilized an
environmentally sensitive fluorophore, 1,1′-bis(4-anilino-5-
naphthalenesulfonic acid (bis-ANS) (Molecular Probes, Eu-
gene, OR), to look at Grp94 conformational change (Wassen-
berg et al., 2000). rHsp90 (Stressgen, BC, Canada, No. SPP-
770) at a final concentration of 2.5 µg in a total volume of 85
µL of buffer A (110 mM KOAc, 20 mM NaCl, 2 mM Mg(OAc)2,
25 mM K-HEPES, pH 7.2, 100 µM CaCl2) was added to 96-
well plates. DMSO control or test compounds were added at
100, 30, 10, 3, 1, and 0.3 µM (10 µL of 10× stocks of each
compound added), and the plates were mixed for 30 s on a
plate shaker before incubation for 60 min at 37°C. Then to
each well, 5 µL of 0.1 mM bis-ANS was added. The plate was
covered with foil and mixed for 30 s on a plate shaker before
incubation for 30 min at 37°C. Relative fluorescence units were
measured using a SpectraMax Gemini XS spectrofluorometer
(Molecular Devices Corporation, Sunnyvale, CA) at an excita-
tion wavelength for bis-ANS at 393 nm, and emission wave-
length was 484 nm. The data were acquired using the
SOFTmaxPRO software (Molecular Devices Corporation, Sunny-
vale, CA). The background was defined as the RFU generated
from wells that did not get rHsp90 but to which the bis-ANS
was added. The percent inhibition of conformational change,
i.e., percent inhibition of bis-ANS activity for each sample, was
calculated from the background-subtracted values as follows:
EC50 was defined as the concentration of the compound at
which there was 50% degradation of the Her-2/neu protein.
Cell Lysa te Bin d in g Assa y. MCF7 cell lysates were
prepared in lysis buffer (20 mM Hepes, pH 7.3, 1 mM EDTA,
5 mM MgCl2, 100 mM KCl) in a Dounce homogenizer and then
incubated with or without test compound for 30 min at 4 °C,
followed by incubation with biotin-GM linked to BioMagTM
streptavidin magnetic beads (Qiagen) for 1 h at 4 °C. Tubes
were placed on a magnetic rack, and the unbound supernatant
was removed. The magnetic beads were washed three times
in lysis buffer and boiled for 5 min at 95 °C in an SDS-PAGE
sample buffer. Samples were analyzed on SDS protein gels,
and Western blots were done for rHsp90. Bands in the Western
blots were quantitated using the Bio-rad Fluor-S MultiImager,
and the percent inhibition of binding of rHsp90 to the biotin-
GM was calculated. The IC50 reported is the concentration of
test compound needed to cause half-maximal inhibition of
binding.
MTS Assa y. Mea su r em en t of Cytotoxicity of Geld a n a -
m ycin Der iva tives. Cells were seeded in 96-well plates at
2000 cells/well and allowed to adhere overnight in Dulbecco’s
modified Eagle’s medium supplemented with 10% fetal bovine
serum. The final culture volume was 100 µL. Viable cell
number was determined by using the Celltiter 96 AQueous
nonradioactive cell proliferation assay (Promega, Madison WI).
The MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxy-
phenyl)-2-(4-sulfophenyl)-2H- tetrazolium/PMS (phenazine
methosulfate) solution was mixed at a ratio of 20:1, and 20
µL was added per well to 100 µL of culture medium. MTS is a
tetrazolium dye that is converted to a formazan product by
dehydrogenase enzymes of metabolically active cells,31 which
is measured at 490 nm absorbance after 2-4 h using a
multiwell plate spectrophotometer. Background was deter-
mined by measuring the Abs 490 nm of cell culture medium
and MTS-PMS in the absence of cells. The background value
was subtracted from all values. The percent viable cells was
calculated as follows:
% inhibition of bis-ANS activity )
(untreated RFU) - (sample RFU)
× 100
untreated RFU
IC50 was defined as the concentration of the compound at
which there was 50% inhibition of bis-ANS activity.
Her -2 Degr a d a tion Assa y. MCF7 breast carcinoma cells
(ATCC) were grown in Dulbecco’s modified Eagle’s medium
(DMEM) containing 10% fetal bovine serum (FBS) and 10 mM
HEPES and plated in 24-well plates (50% confluent). Twenty-
four hours later (cells are 65-70% confluent), test compounds
were added at 1000 nM and six half-log dilutions down to 3
nM and incubated overnight for 16 h. The wells were washed
with 1 mL of phosphate buffered saline (PBS), and 200 µL of
trypsin was added to each well. After trypsinization was
complete, 50 µL of FBS was added to each well. Then 200 µL
of cells was transferred to 96-well plates The cells were
pipetted up and down to obtain a single-cell suspension. The
plates were centrifuged at 2500 rpm for 1 min using a Sorvall
Legend RT tabletop centrifuge (Kendro Laboratory Products,
Asheville, NC). The cells were then washed once in PBS
containing 0.2% BSA and 0.2% sodium azide (BA buffer). PE
conjugated anti-Her-2/neu antibody (Becton Dickinson, No.
340552) or PE conjugated anti keyhole limpet hemacyanin
Abs 490 sample
Abs 490 nm untreated cells
% viable cells )
× 100
IC50 was defined as the concentration that gave rise to 50%
viable cell number.
In t r a ven ou s a n d Or a l P h a r m a cok in et ic St u d y in
Mice. Female Balb/C mice were administered 2i (dispersion
formulation) as a 1 min intravenous infusion (n ) 3) or by
oral gavage (n ) 3) at doses of 20 mg/kg (iv) or 100 mg/kg
(oral). Blood samples were collected at intervals, and the total
2i concentration in serum was determined using a reversed-
phase HPLC-UV method with a 50 ng/mL limit of quantita-
tion. Pharmacokinetic parameters were estimated by fitting

Geldanamycin Derivatives
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 15 3873
(12) Schneider, C.; Sepp-Lorenzino, L.; Nimmesgern, E.; Ouerfelli,
O.; Danishefsky, S.; Rosen, N.; Hartl, F. U. Pharmacologic
Shifting of a Balance between Protein Refolding and Degradation
Mediated by Hsp90. Proc. Natl. Acad. Sci. U.S.A. 1996, 93,
14536-14541.
(13) Solit, D. B.; Zheng, F. F.; Drobnjak, M.; Mu¨nster, P. N.; Higgins,
B.; Verbel, D.; Heller, G.; Tong, W.; Cordon-Cardo, C.; Agus, D.
B.; Scher, H. I.; Rosen, N. 17-Allylamino-17-demethoxygeldana-
mycin Induces the Degradation of Androgen Receptor and Her-
2/neu and Inhibits the Growth of Prostate Cancer Xenografts.
Clin. Cancer Res. 2002, 8, 986-993.
(14) Solit, D. B.; Basso, A. D.; Olshen, A. B.; Scher, H. I.; Rosen, N.
Inhibition of Heat Shock Protein 90 Function Down-Regulates
Akt Kinase and Sensitizes Tumors to Taxol. Cancer Res. 2003,
63, 2139-2144.
(15) Basso, A. D.; Solit, D. B.; Chiosis, G.; Giri, B.; Tsichlis, P.; Rosen,
N. Akt Forms an Intracellular Complex with Heat Shock Protein
90 (Hsp90) and Cdc37 and Is Destabilized by Inhibitors of Hsp90
Function. J . Biol. Chem. 2002, 277, 39858-39866.
(16) Minami, Y.; Kiyoi, H.; Yamamoto, Y.; Yamamoto, K.; Ueda, R.;
Saito, H.; Naoe, T. Selective Apoptosis of Tandemly Duplicated
FLT3-Transformed Leukemia Cells by Hsp90 Inhibitors. Leu-
kemia 2002, 16, 1535-1540.
(17) Whitesell, L.; Mimnaugh, E. G.; De Costa, B.; Myers, C. E.;
Neckers, L. M. Inhibition of Heat Shock Protein Hsp90-pp60v-
src Heteroprotein Complex Formation by Benzoquinone Ansa-
mycins: Essential Role for Stress Protein in Oncogenic Trans-
formation. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 8324-8328.
(18) Roe, S. M.; Podromou, C.; O’Brien, R.; Ladbury, J . E.; Piper, P.
W.; Pearl, L. H. Structural Basis for Inhibition of the Hsp90
Molecular Chaperone by the Antitumor Antibiotics Radicicol and
Geldanamycin. J . Med. Chem. 1999, 42, 260-266.
(19) J ez, J . M.; Chen, J . C.-H.; Rastelli, G.; Stroud, R. M.; Santi, D.
V. Crystal Structure and Molecular Modeling of 17-DMAG in
Complex with Human Hsp90. Chem. Biol. 2003, 10, 361-368.
(20) (a) Egorin, M. J .; Lagattuta, T. F.; Hamburger, D. R.; Covey, J .
M.; White, K. D.; Musser, S. M.; Eiseman, J . L. Cancer
Chemother. Pharmacol. 2002, 49, 7-19. (b) Egorin, M. J .; Rosen,
D. M.; Wolff, J . H.; Callery, P. S.; Musser, S. M.; Eiseman, J . M.
Metabolism of 17-(Allylamino)-17-demethoxygeldanamycin by
Murine and Human Hepatic Preparations. Cancer Res. 1998,
58, 2385.
the mean 2i concentration versus time data to a two-compart-
ment model (WinNonlin).
HP LC Con dition s. An Agilent 1100 HPLC system equipped
with a photodiode array detector was used. A gradient method
was used with mobile phase A consisting of water with 1%
acetic acid and 0.5% triethylamine and mobile phase B
consisting of acetonitrile with 1% acetic acid and 0.5% tri-
ethylamine. The content of mobile phase B was increased from
5% to 100% over 10 min, and the column was reequilibrated
to the starting condition over 5 min (total run time ) 15
min).The flow rate was 1.0 mL/min, and a Zorbax 300SB-C18
column was used (3.5 µM; 4.6 mm × 150 mm). Quantitation
was based on UV absorbance at 254 nM compared to a
standard.
Xen ogr a ft Stu d ies in Mice. The human tumor cell lines
A549 (non-small-cell lung carcinoma), H69 (small-cell lung
cancer), and U87 (glioblastoma multiforme) were obtained
from the American Type Tissue Culture Collection (ATCC) and
maintained in DMEM supplemented with nonessential amino
acids, sodium pyruvate, and HEPES. Female BALB/c nu/nu
mice were obtained from Harlan and maintained in microiso-
lator boxes. Recipient mice (five per group) were injected with
5 × 106 tumor cells on each flank, and tumors were allowed
to become established (starting volumes of 60-160 mm3) before
onset of treatment. Hsp90 inhibitors were formulated in PETA
(PEG 400/ethanol/Tween 80 mixed 6:3:1 in 0.05 M acetate
buffer, pH 5). Animals received either PETA alone or 2b or 2i
in PETA intraperitoneally (ip) at the doses and schedules
indicated in the figure legends. Tumors were measured in
three dimensions with calipers, and tumor volume was esti-
mated using the algorithm volume ()length × width × height
π
× /6).
Su p p or tin g In for m a tion Ava ila ble: Rf, HPLC purity, 1H
NMR, and HRMS data are provided for the compounds
described in the manuscript. This material is available free of
charge via the Internet at http://pubs.acs.org.
(21) Pan, S.-S.; Gonzalez, H. Mitomycin Antibiotic Reductive Poten-
tial and Related Pharmacological Activities. Mol. Pharmacol.
1990, 37, 966-970.
(22) Sasaki, K.; Inoue, Y. Geldanamycin Derivatives and Antitumor
Drug. U.S. Patent 4,261,989, April 14, 1981.
Refer en ces
(1) Gadjos, C.; J ordan, V. Selective Estrogen Receptor Modulators
as a New Therapeutic Drug Group: Concept to Reality in a
Decade. Clin. Breast Cancer 2002, 2, 272-281.
(23) Neel, D. A.; J irousek, M. R.; McDonald, J . H., III. Synthesis of
(2) Carswell, C. I.; Figgitt, D. P. Bicalutamide: In Early-Stage
Prostate Cancer. Drugs 2002, 62, 2471-2481.
Bisindolylmaleimides Using
a Palladium Catalyzed Cross-
Coupling Reaction. Bioorg. Med. Chem. Lett. 1998, 8, 47-50.
(24) Wassenberg, J . J .; Reed, R. C.; Nicchitta, C. V. Ligand Interac-
tions in the Adenosine Nucleotide-Binding Domain of the Hsp90
Chaperone, GRP94. J . Biol. Chem. 2000, 275, 22806-22814.
(25) The standard error for this assay is (30%.
(26) Bundgaard, H. The Double Prodrug Concept and Its Applica-
tions. Adv. Drug Delivery Rev. 1989, 3, 39-65.
(27) Kamal, A.; Thao, L.; Sensintaffar, J .; Zhang, L.; Boehm, M. F.;
Fritz, L. C.; Burrows, F. J . Nature 2003, 425, 407-410.
(28) J ez, J . M.; Chen, J . C.; Rastelli, G.; Stroud, R. M.; Santi, D. V.
Crystal Structure and Molecular Modeling of 17-DMAG in
Complex with Human Hsp90. Chem. Biol. 2003, 10, 361-368.
(29) (a) Stebbins, C. E.; Russo, A. A.; Schneider, C.; Rosen, N.; Hartl,
F. U.; Pavletich, N. P. Crystal Structure of an Hsp90-Geldana-
mycin Complex: Targeting of a Protein Chaperone by an
Antitumor Agent. Cell 1997, 89, 239-250. (b) Roe, S. M.;
Podromou, C.; O’Brien, R.; Ladbury, J . E.; Piper, P. W.; Pearl,
L. H. Structural Basis for Inhibition of the Hsp90 Molecular
Chaperone by the Antitumor Antibiotics Radicicol and Geldana-
mycin. J . Med. Chem. 1999, 42, 260-266.
(30) Egorin, M. J .; Zuhowski, E. G.; Rosen, M. D.; Sentz, D. L.; Covey,
J . M.; Eiseman, J . L. Plasma Pharmacokinetics of Tissue
Distribution of 17-(Allylamino)-17-demethoxygeldanamycin (NSC
330507) in SD2F1 Mice. Cancer Chemother Pharmacol. 2001,
47, 291-302.
(31) Corey, A.; Owen, T.; Barltrop, J .; Cory, J . Use of an aqueous
soluble tetrazolium/formazan assay for cell growth assays in
culture. Cancer Commun. 1991, 3, 207-212.
(3) Zhang, H.; Richter, M.; Greene, M. I. Therapeutic Monoclonal
Antibodies for the ErbB Family of Receptor Tyrosine Kinases.
Cancer Biol. Ther. 2003, 2 (4), S122-S126.
(4) Cohen, M. H.; Williams, G. A.; Sridhara, R.; Chen, G.; Pazdur,
R. FDA Drug Approval Summary: Gefitinib (ZD1839) (Iressa)
Tablets. Oncologist 2003, 8, 303-306.
(5) Capdeville, R.; Buchdunger, E.; Zimmermann, J .; Matter, A.
Glivec (STI571, imatinib), a Rationally Developed Targeted
Anticancer Drug. Nat. Rev. Drug Discovery 2002, 7, 493-502.
(6) Pearl, L. H.; Prodromou, C. Curr. Opin. Struct. Biol. 2000, 10,
46.
(7) Maloney, A.; Workman, P. Hsp90 as a New Therapeutic Target
for Cancer Therapy: The Story Unfolds. Expert Opin. Biol. Ther.
2002, 2, 3-24.
(8) Blank, M.; Mandel, M.; Keisari, Y.; Muruelo, D.; Lavie, G.
Enhanced Ubiquinylation of Heat Shock Protein 90 as a Poten-
tial Mechanism for Mitotic Cell Death in Cancer Cells Induced
with Hypericin. Cancer Res. 2003, 63, 8241-8247.
(9) Gorre, M. E.; Ellwood-Yen, K.; Chiosis, G.; Rosen, N.; Sawyers,
C. L. BCR-ABL point mutants isolated from patients with
imatinib mesylate-resistant chronic myeloid leukemia remain
sensitive to inhibitors of the BCR-ABL chaperone heat shock
protein 90. Blood 2002, 100, 3041-3044.
(10) Solit, D. B.; Scher, H. I.; Rosen, N. Hsp90 as a Therapeutic
Target in Prostate Cancer. Semin. Oncol. 2003, 30, 709-716.
(11) Beliakoff, J .; Bagatell, R.; Paine-Murrieta, G.; Taylor, C. W.;
Lykkesfeldt, A. E.; Whitesell, L. Hormone-Refractory Breast
Cancer Remains Sensitive to the Antitumor Activity of Heat
Shock Protein 90 Inhibitors. Clin. Cancer Res. 2003, 9, 4961-
4971.
J M0306125