Antagonism of the Stat3-Stat3 Protein Dimer with Salicylic Acid Based Small Molecules

Article abstract of DOI:10.1002/cmdc.201100194

More than 50 new inhibitors of the oncogenic Stat3 protein were identified through a structure-activity relationship (SAR) study based on the previously identified inhibitor S3I-201 (IC₅₀=86μM, K_i>300μM). A key structural feature of

Full text of DOI:10.1002/cmdc.201100194

DOI: 10.1002/cmdc.201100194
Antagonism of the Stat3–Stat3 Protein Dimer with Salicylic
Acid Based Small Molecules
Steven Fletcher,[a] Brent D. G. Page,[a] Xialoei Zhang,[b] Peibin Yue,[b] Zhi Hua Li,[c]
Sumaiya Sharmeen,[d] Jagdeep Singh,[a] Wei Zhao,[b] Aaron D. Schimmer,[d] Suzanne Trudel,[c]
James Turkson,*[b] and Patrick T. Gunning*[a]
More than 50 new inhibitors of the oncogenic Stat3 protein
were identified through a structure–activity relationship (SAR)
study based on the previously identified inhibitor S3I-201
(IC50 =86 mm, Ki >300 mm). A key structural feature of these in-
hibitors is a salicylic acid moiety, which, by acting as a phos-
photyrosine mimetic, is believed to facilitate binding to the
Stat3 SH2 domain. Several of the analogues exhibit higher po-
tency than the lead compound in inhibiting Stat3 DNA binding
activity, with an in vitro IC50 range of 18.7–51.9 mm, and disrup-
tion of Stat3–pTyr peptide interactions with Ki values in the
15.5–41 mm range. One agent in particular exhibited potent in-
hibition of Stat3 phosphorylation in both breast and multiple
myeloma tumor cells, suppressed the expression of Stat3
target genes, and induced antitumor effects in tumor cells har-
boring activated Stat3 protein.
Introduction
The signal transducer and activator of transcription 3 (Stat3)
protein mediates the relay of extracellular cytokine or growth
factor stimulation to the nucleus, where it initiates the expres-
sion of gene profiles that promote cell proliferation, differentia-
tion, and cell survival.[1] In normal cells, Stat3 transcriptional ac-
tivity is transient and responsive to physiological cues. Howev-
er, numerous human cancer cell lines, including breast,[2] pros-
tate,[3] ovarian,[4] brain,[5] and lung[6] have been found to harbor
persistently activated Stat3 protein. Aberrant Stat3 activity is
widely acknowledged to be a master regulator of the cancer
phenotype and to play a critical role in malignant transforma-
tion and tumorigenesis.[1] Moreover, dysregulated Stat3 tran-
scriptional function has been implicated in the induction of
tumor immune tolerance.[7] Overactivation of Stat3 promotes
tumorigenesis by the up-regulation of cell survival proteins,
cell cycle regulators and induction of angiogenesis.[8] Inhibition
of Stat3 signaling correlates with suppression of cell transfor-
mation, motility, growth, and the induction of apoptosis in ma-
lignant cells.[9] Cell lines that lack aberrant Stat3 activation are
more tolerant to Stat3 inhibitors, possibly identifying an irre-
versible dependence on persistent Stat3 activation for survival
in vulnerable cell lines.[9] Of clinical and therapeutic signifi-
cance, earlier studies from our research group and others have
shown that in vivo administration of inhibitors of Stat3–Stat3
dimerization induce tumor regression in xenograft models.[10,11]
In summary, Stat3 protein is considered an exciting and high-
value target for cancer therapeutics.
mic domain, creating docking sites for the recruitment of mon-
omeric unphosphorylated Stat3 protein via its SH2 domain.
Stat3 is phosphorylated on a key tyrosine residue, Tyr705,
which leads to receptor dissociation and the formation of acti-
vated Stat3–Stat3 dimers through reciprocal SH2–pTyr705 in-
teractions. After translocation to the nucleus, dimeric Stat3
complexes bind to DNA response elements and promote gene
transcription.[13,14]
Inhibition of constitutive Stat3–Stat3 complexes by disrup-
tion of binding interfaces offers significant value as a molecular
targeted therapy for cancer treatment.[10] Disruption of Stat3
complexes has been achieved through SH2 domain binders
[a] Dr. S. Fletcher,+ B. D. G. Page,+ J. Singh, Prof. P. T. Gunning
Department of Chemistry, University of Toronto Mississauga
Mississauga, ON, L5L 1C6 (Canada)
Fax: (+1)905-828-5425
E-mail: patrick.gunning@utoronto.ca
[b] X. Zhang, P. Yue, W. Zhao, Prof. J. Turkson
Department of Molecular Biology and Microbiology
Burnett College of Biomedical Sciences
University of Central Florida, Orlando, FL, 32826 (USA)
Fax: (+1)407-384-2062
E-mail: jturkson@mail.ucf.edu
[c] Z. H. Li, Prof. S. Trudel
Division of Medical Oncology and Hematology
University Health Network
Princess Margaret Hospital, McLaughlin Centre of Molecular Medicine
620 University Ave, Toronto, ON, M5G 2C1 (Canada)
[d] S. Sharmeen, Prof. A. D. Schimmer
Ontario Cancer Institute/Princess Margaret Hospital
610 University Avenue, Toronto, ON, M5G 2M9 (Canada)
[+] These authors contributed equally to this work.
The canonical view of Stat3 signaling describes latent Stat3
protein (monomeric[1] or dimeric[12]) residing predominantly in
the cytoplasm. Ligand binding to the extracellular domain of
transmembrane receptors induces intracellular activation of ty-
rosine kinases such as Janus kinases (JAKs). Receptors are
phosphorylated on critical tyrosine residues of their cytoplas-
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201100194: full synthetic protocols, char-
acterization of intermediate compounds, whole-cell tumor data, fluores-
cence polarization binding data, and EMSA results.
ChemMedChem 2011, 6, 1459 – 1470
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1459

J. Turkson, P. T. Gunning, et al.
MED
that compete with phosphorylated Stat3 monomers for the
phosphotyrosine (pTyr) binding module. Numerous research
groups, including our own, have shown that disruption of
Stat3 transcriptional activity through dimer disruption leads to
suppression of Stat3 gene expression profiles and induction of
apoptosis. Stat3 dimers have been effectively disrupted by
peptides,[15] peptidomimetics,[16] small molecules,[17] and metal
complexes (Figure 1).[18] Peptidic inhibitors have been derived
Ser613, and two hydrophobic domains; the first comprises
Ile634 and the hydrocarbon portions of the side chains of
Lys591 and Arg595, and the second comprises Trp623, Val637,
Ile659, and Phe716. The structural core of S3I-201 is glycolic
acid, the carboxylic acid of which has been condensed with 4-
aminosalicylic acid to furnish the amide bond, and the hydroxy
group of which has been tosylated. Because S3I-201 carries
only two appendages off the main scaffold, GOLD[19] docking
unsurprisingly demonstrated that this small molecule
can simultaneously occupy only two of these three
subpockets (Figure 2). The salicylic acid moiety of
S3I-201 is a known pTyr mimetic,[20] and low-energy
GOLD docking studies consistently placed it in the
pTyr binding site. The potential for hydrogen bonds
and salt bridges here suggests that this component
is responsible for a considerable portion of the bind-
ing energy with the Stat3 SH2 domain. GOLD dock-
ing studies suggested that the O-toluenesulfonyl
(tosyl) group binds in the Arg595/Ile634 subpocket,
leaving the Trp623/Phe716 subpocket unoccupied;
the secondary amide NH of S3I-201 offers an excel-
lent opportunity to gain access to this third subpock-
et. Thus, we were confident that a rational, synthetic
program, facilitated by the inherent modular design
of S3I-201, would allow the optimization of contacts
between small molecule and the Stat3 SH2 domain
to furnish more potent analogues of S3I-201.
The tosylate moiety in S3I-201 is an excellent leav-
ing group, allowing nucleophilic attack at the carbon
atom to which it is attached. Moreover, in this case,
S3I-201 is especially prone to nucleophilic attack, due
to the interaction of the s* orbital (the LUMO) of the
CꢀOTs bond with the p* orbital of the adjacent car-
bonyl group. Whilst we have no conclusive proof that S3I-201
functions as an irreversible inhibitor, there are several nucleo-
Figure 1. Structures of Stat3 inhibitors 1–8.
from the cognate binding sequence of Stat3 (pYLKTK) and
from the Stat3-binding gp130 receptor (GpYLPQTV).[10] Inspired
by these proof-of-principle peptidic probes, our group-
s[15a,16a,d,17d–g] and many others have synthesized optimized,
more drug-like second-generation peptidomimetic inhibi-
tors.[16b–c,e–g] Most notably, these include ISS610 (4-CN-Ph-pTyr-
Leu (1))[16a] derived from pYLKTK, pCin-Leu-Pro-Glu-NHBn (2)[16b]
derived from GpYLPQTV, and, most recently, the cell-permeable
macrocyclic compound CJ-1383 (3).[16c] In addition to peptido-
mimetics, small-molecule inhibitors such as Stattic (4),[17a]
LLL12 (5),[17b] STA-21 (6),[17c] and S3I-M2001 (7)[17d,e] have been
identified through a combination of in silico and in vitro
screening of chemical libraries as well as de novo rational
design.
By conducting an in silico structure-based virtual screen of
the National Cancer Institute (NCI) chemical libraries, our re-
search groups recently identified the potent Stat3 inhibitor
S3I-201 (Figure 1, compound 8: IC50 =86 mm as determined by
an electrophoretic mobility shift assay (EMSA)).[17f] We identi-
fied that S3I-201 offers several opportunities for structural di-
versification, and embarked on a medicinal chemistry program
to identify more potent analogues of S3I-201. Broadly speak-
ing, the Stat3 SH2 domain is composed of three subpockets: a
hydrophilic domain bounded by Lys591, Arg609, Ser611, and
Figure 2. Low-energy GOLD[19] docking conformation of S3I-201 (8); hydro-
phobic residues are indicated in light grey, hydrophilic residues are dark
grey.
1460
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1459 – 1470

Stat3–Stat3 Dimer Antagonists
philic residues on the Stat3 SH2 domain surface, including
Cys418 and Cys712, that may form a covalent bond to S3I-
201. Such an event might compromise an SAR study because
the majority of the inhibitory activity would be derived from
the irreversible conjugation to the protein surface, which
would be common to all S3I-201 analogues. Moreover, it is
probable that irreversible inhibitors would exhibit poorer pro-
tein selectivity profiles than their reversible inhibitor counter-
parts. Therefore, we decided to replace the scaffold oxygen
atom with a nitrogen atom to convert the labile tosylate into a
non-labile tosylamide. The resulting secondary sulfonamide
possesses a polar NH group, but despite this, GOLD docking
studies consistently placed the tosylamide in the same hydro-
phobic subpocket (Arg595/Ile634) as the parent tosylate in
S3I-201 (compound 9 or SF-1-082,[16g] Figure 3A). Nevertheless,
N(CH3) unit of the tosylamide. The salicylic acid moiety is a
known phosphotyrosine mimetic, and because its modification
would add considerably to the synthetic effort required for this
research, we chose to keep this component constant. The re-
maining three sites would be subjected to SAR studies. Herein
we elaborate on our previous communication[16g] by expanding
on the SAR work of our initial lead compound S3I-201 and by
providing additional biological characterizations in vitro and in
whole cells.
Results and Discussion
If the considerable inhibitory activity of S3I-201 is due to its
ability to covalently modify the Stat3 target, then conversion
of the labile O-tosyl group to the non-labile N(CH3)-tosyl group
would be expected to cause a significant decrease in the inhib-
ition of Stat3. To investigate this, we first prepared a focused
set of non-labile analogues of S3I-201 (shown in Table 1), the
Table 1. EMSA inhibition data for disruption of the Stat3–Stat3:DNA com-
plex in vitro by a focused set of S3I-201 (8) analogues.
Compd
R1
X
IC50 [mm]
8 (S3l-201)
H
H
H
H
O
NH
NCH3
NBoc
86ꢁ33
>300
>300
9
10
11
>300
12
13
14
15
O
>300
NH
>300
Figure 3. Low-energy GOLD[19] docking conformations of A) compound 9
and B) compound 14; hydrophobic residues are indicated in light grey, hy-
drophilic residues are dark grey.
NCH3
NBoc
292ꢁ35
>300
to further encourage occupancy of this hydrophobic subpock-
et, we elected to convert the NH group of the tosylamide to
the more hydrophobic NCH3 group. In addition, as alluded to
previously, a key aspect of this work was to functionalize the
secondary amide NH, as it was anticipated that doing so
would allow access to the third, as-yet-unexplored hydropho-
bic subpocket (Trp623, Val637, Ile659, and Phe716). Indeed,
several low-energy GOLD docked poses of the N-benzyl deriva-
tive 14 (previously reported as SF-1-062)[16g] revealed that, as
well as the salicylic acid and the N(CH3)-tosyl components
binding the same subpockets as the corresponding compo-
nents in the parent S3I-201, the N-benzyl group is projected
into the third subpocket (Trp623/Phe716) as predicted (Fig-
ure 3B).
syntheses of which are described in full in the Supporting In-
formation. Unfortunately, replacement of the scaffold oxygen
atom with NH, NCH3, or NBoc led to a decrease in activity in all
cases, from an IC50 value of 86 mm (by EMSA) to >300 mm for
all non-labile analogues, suggesting that S3I-201 might indeed
operate, at least in part, as an irreversible inhibitor. On the
other hand, benzylation of the amide NH of S3I-201 also led to
a loss in inhibitory activity (compound 12, SF-1-120:[16g] IC50 >
300 mm) despite the alkylating potential of this analogue re-
maining intact. Nevertheless, within this series (R1 =benzyl), we
observed a slight recovery in activity if the scaffold oxygen
atom is replaced with the NCH3 unit (14: IC50 =292 mm), and
thus we elected to constrain the X heteroatom/group as NCH3
for most of this research project. Because 14 demonstrated
some activity against Stat3, and we believed the R1 =benzyl
The N(CH3)-tosylamide analogue of S3I-201 offers four po-
tential optimization sites: 1) the salicylic acid component,
2) the secondary amide NH, 3) the tosyl moiety, and 4) the
ChemMedChem 2011, 6, 1459 – 1470
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1461

J. Turkson, P. T. Gunning, et al.
MED
group of that inhibitor makes fa-
vorable interactions with the
Table 2. EMSA inhibition data for disruption of the Stat3–Stat3:DNA complex in vitro by a series of R1-function-
alized analogues of compound 10.
Trp623/Phe716
hydrophobic
subpocket, as predicted by
GOLD docking experiments (Fig-
ure 3B), we decided to investi-
gate the effects of modifying the
benzyl group, in particular at the
para and meta positions, where
deeper access to the subpocket
might be realized.
Compd
R1
H
IC50 [mm]
>300
Compd
R1
IC50 [mm]
>300
10
27e
14
295ꢁ35
27 f
194ꢁ47
115 ꢁ60
Probing the Trp623, Val637,
Ile659, and Phe716 hydro-
phobic subpocket: SAR of the
R1 group
27a
290
27g
The series of S3I-201 analogues
listed in Table 2, where X=NCH3,
were furnished by following the
synthetic steps outlines in
Scheme 1. After the one-pot and
stepwise benzylations of the car-
boxylic acid and hydroxy func-
tionalities of 4-aminosalicylic
acid (16), which proceeded in
moderate yield (54%), several
hydrophobic aldehydes (RCHO)
were reductively aminated with
27b
>300
27h
27i
35ꢁ9
27c
27d
260ꢁ47
298ꢁ11
280ꢁ15
highly reactive peptide coupling agent dichlorotriphenylphos-
phorane (PPh3Cl2), which is believed to generate the corre-
sponding acid chloride of 23 in situ. Finally, a global debenzy-
lation of compounds 24, 25, and 26a–i with hydrogen gas
over 10% palladium on carbon yielded the series of S3I-201
analogues 10, 14, and 27a–i. Importantly, the aryl nitrile moiet-
ies were essentially untouched in the debenzylation reactions,
with the reducing conditions proving chemoselective for re-
moval of the benzyl protecting groups. These phenomena are
likely due to a combination of rapid reaction times (both aryl
the resultant aniline 17 to afford the series of secondary ani-
lines 18 and 19a–i in very good to excellent yields. Meanwhile,
sulfonylation of glycine methyl ester (20) with para-toluenesul-
fonyl chloride (p-TsCl) furnished secondary sulfonamide 21,
which was subsequently N-methylated with methyl iodide, and
then saponified with lithium hydroxide to generate carboxylic
acid 23 (75% yield over three steps). Condensation of the pri-
mary aniline 17 and the secondary anilines 18 and 19a–i with
acid 23 to deliver the secondary amide 24 and the tertiary
amides 25 and 26a–i, respectively, was achieved with the
Scheme 1. a) 1. BnBr, KOtBu, DMF, 08C!RT, 5 h; 2. BnBr, KOtBu, DMF, 08C!RT, 16 h, 54%; b) 1. RCHO, AcOH, 4 ꢁ MS, MeOH, 458C, 3 h; 2. NaCNBH3, RT, 12 h,
75–96%; c) p-TsCl, DIPEA, CH3CN, 08C!RT, 1 h, 93%; d) MeI, Cs2CO3, DMF, RT, 16 h, 85%; e) LiOH·H2O, THF/MeOH/H2O (3:1:1), RT, 1 h, 95%; f) PPh3Cl2, CHCl3,
608C, 12 h, 89–95%; g) H2, 10% Pd/C, MeOH/THF (1:1), RT, 1–16 h, 85–100%; or for 26a and 26b: h) LiOH·H2O, THF/H2O (3:1), RT, 24 h, 76–86%; i) TFA/toluene
(1:2), RT, 16 h, 85–93%.
1462
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1459 – 1470

Stat3–Stat3 Dimer Antagonists
nitrile-containing intermediates 26c and 26d were doubly de-
benzylated in ~1 h), which limited the exposure of the nitrile
functional group to the reducing conditions, and the fortuitous
limited solubilities of the compounds in neat methanol, requir-
ing the use of THF as co-solvent, which is known to suppress
hydrogenation of nitriles.[21] Conversely, because the reduction
of aryl–bromide bonds with H2 and Pd/C catalyst is known to
be a relatively facile reaction,[22] hydrogenolysis of the benzyl
protecting groups in intermediates 26a and 26b was not at-
tempted. Instead, we employed a high-yielding, non-reducing,
two-step protocol. First, the benzyl ester was selectively hydro-
lyzed with lithium hydroxide (the tertiary amide was slowly hy-
drolyzed under these conditions), and then the benzyl ether
was cleaved under acidic conditions with trifluoroacetic acid
(TFA). Removal of the benzyl ether under these conditions is
believed to be facilitated by chelation of a proton between the
carboxylic acid and ether functionalities.[23]
Replacement of the R1 benzyl group in 14 with 4-cyanoben-
zyl (27c, SF-1-073)[16g] led to an improvement in activity (IC50 =
260 mm for 27c; cf. IC50 =292 mm for 14). This enhancement in
Stat3 inhibition may be due to improved hydrophobic interac-
tions with the larger and more electron-poor aromatic system,
and/or from a hydrogen bond between the nitrile group and
the SH2 domain. More interesting is the observation that Stat3
inhibition improved with increasing size of the hydrophobic R1
group. Specifically, 4-(tert-butyl)benzylated agent 27 f (SF-1-
068)[16g] showed marked improvement in activity over both 14
(R1 =benzyl) and 27c (R1 =4-cyanobenzyl), whilst replacement
of the tert-butyl group with a phenyl ring to give the large bi-
phenyl-based inhibitor 27g (SF-1-070, R1 =4-phenylbenzyl) led
to a further approximate twofold increase in potency (27g,
IC50 =115 mm; cf. IC50 =194 mm for 27 f). Furthermore, the inclu-
sion of the especially hydrophobic 4-cyclohexylbenzyl group at
the R1 position furnished an inhibitor that exhibited Stat3 in-
hibitory activity with more than double the potency of our
lead agent: IC50 =35 mm for 27h (SF-1-066);[16g] cf. IC50 =86 mm
for S3I-201 (8).
Scheme 2. a) R3 =Boc: Boc2O, cat. DMAP, CH2Cl2, RT, 1 h, 95%; R3 =aryl: R3F
or R3Cl, DIPEA, DMSO, 1208C, 16 h, 76–96%; b) H2, 10% Pd/C, MeOH/THF
(1:1), RT, 1–16 h, 85–100%.
a hydrophobic subpocket, we appreciated that conjugation of
groups to the piperidine nitrogen that would considerably de-
crease its basicity would be required. Thus, the transformations
conducted on the piperidine nitrogen (Scheme 2) included re-
tert-butoxycarbonylation and arylation with 4-fluorobenzoni-
trile or 2-chloropyrimidine to afford, after benzyl deprotections,
inhibitors 27jb, 27jc, and 27jd, respectively. Unfortunately, as
shown in Table 3, none of the inhibitors were active; all exhib-
ited EMSA IC50 values >300 mm.
Table 3. EMSA inhibition data for disruption of the Stat3–Stat3:DNA com-
plex in vitro by a series of R1 =N-(4-piperidinyl)methyl-based analogues
of compound 10.
Compd
R3
H
IC50 [mm]
>300
27ja
27jb
27jc
27jd
>300
>300
>300
N-Substituted piperidinylmethyl derivatives and N-substitut-
ed 4-(piperidinyl)benzyl derivatives
Because greater Stat3 inhibitory activity was furnished by sub-
stitution at the para position of the R1 benzyl group in 14, we
were keen to functionalize this position further still. However,
owing to a simpler synthetic demand, it was decided to deter-
mine whether substitution at the 4-position of the cyclohexyl
group (a good match for benzyl) would also enhance inhibitor
activity. Replacement of the cyclohexylmethyl moiety in 27e
with 4-piperidinylmethyl would allow facile elaboration of the
inhibitor through functionalization of the piperidine nitrogen
to probe deeper into the proposed subpocket. To this end,
compound 26j (Scheme 2) was accessed by following the
steps in Scheme 1, where the RCHO aldehyde was N-Boc-piper-
idinylformaldehye (the Boc group was inadvertently removed
during the peptide coupling step with PPh3Cl2; full details for
the synthesis of 26j are given in the Supporting Information).
Because the piperidinylmethyl group was proposed to bind in
Next, we tackled functionalization of the 4-position of the
cyclohexyl component of inhibitor 27h in a similar manner.
This time, preparation of the requisite aldehyde 4-[N-trifluoroa-
cetyl(piperidin-4-yl)]benzaldehyde (32) was slightly more com-
plicated, and its synthesis is illustrated in Scheme 3. Briefly,
protection of the piperidine nitrogen of 4-phenylpiperidine
(29) was accomplished as its acid-stable trifluoroacetamide 30.
Subsequently, regioselective para-chlorocarbonylation of 30
was effected under Friedel–Crafts conditions,[24] and then the
crude acid chloride 31 was reduced to the target aldehyde 32
in a modification of the Rosenmund reaction. Employing 32 as
ChemMedChem 2011, 6, 1459 – 1470
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1463

J. Turkson, P. T. Gunning, et al.
MED
Scheme 3. a) (CF3CO)2O, DIPEA, CH2Cl2, 08C!RT, 3 h, 93%; b) (COCl)2, AlCl3,
CH2Cl2, 08C, 1 h; c) H2, 10% Pd/C, DIPEA, EtOAc, RT, 2 h, 63% (two steps).
the RCHO aldehyde, the corresponding compound 26k was
then furnished by following the appropriate steps in
Scheme 1. Next, as shown in Scheme 4, the trifluoroacetyl
group of 26k was cleaved in excellent yield by brief treatment
with lithium hydroxide to reveal the piperidine nitrogen atom
Scheme 5. a) TFA/toluene (1:1), RT, 4 h, 95%; b) NH4Cl, DIPEA, HBTU, DMF, RT,
16 h, 99%; c) H2, 10% Pd/C, MeOH/THF (1:1), RT, 1–16 h, 85–100%.
by us previously, to deliver monobenzyl-protected
compound 35. Facile condensation of the carboxylic
acid of 35 with ammonium chloride, employing O-
(benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexa-
fluorophosphate (HBTU) as the coupling agent, gen-
erated carboxamide 36 in excellent yield. Deprotec-
tion of the benzyl esters of 35 and 36 under the
usual hydrogenolytic conditions furnished the corre-
sponding inhibitors 27kh and 27ki. As the N-(piperi-
din-4-yl)benzyl moiety was predicted to bind in a hy-
drophobic subpocket, we anticipated that the polar
acid and carboxamide-containing inhibitors might
demonstrate poor activity against Stat3. In fact, as
Table 4 illustrates, among the entire series 27ka–ki,
only 4-cyanophenyl-based 27kd and 4-cyanobenze-
nesulfonyl-based 27 kg exhibited Stat3 inhibitory ac-
Scheme 4. a) LiOH·H2O, THF/H2O (3:1), RT, 10 min, 98%; b) R3 =Boc: Boc2O, cat. DMAP,
CH2Cl2, RT, 1 h, 95%; R3 =aryl: R3F or R3Cl, DIPEA, DMSO, 1208C, 16 h, 80–99%; R3 =p-
CNC6H4SO2: p-CNC6H4SO2Cl, DIPEA, RT, 16 h, 99%; R3 =p-CNC6H4CO2: p-CNC6H4CO2H,
HBTU, DIPEA, DMF, RT, 16 h, 89%; c) H2, 10% Pd/C, MeOH/THF (1:1), RT, 1–16 h, 85–
100%.
Table 4. EMSA inhibition data for disruption of the Stat3–Stat3:DNA complex in vitro by a series of R1 =N-(4-pi-
peridinyl)benzyl-based analogues of compound 10.
in 33. Subsequent functionaliza-
tion of this nitrogen was accom-
plished with a variety of re-
agents to furnish, after the stan-
dard benzyl deprotections, the
series of compounds 27ka–kg
depicted in Table 4. As in the
case of the N-piperidinylmethyl
series of inhibitors 27ja–jd, we
elected to substitute the piperi-
dine nitrogen atom in 33 with
functionalities that would de-
crease its basicity through with-
drawal of its lone pair of elec-
trons into aryl systems, and acyl
and sulfonyl groups. Inhibitors
27kh and 27ki were prepared
as shown in Scheme 5. Specifi-
cally, deprotection of the tert-
butyl ester of 34h with TFA also
led to the concomitant removal
of the benzyl ether, as reported
Compd
R3
IC50 [mm]
>300
Compd
R3
IC50 [mm]
>300
27ka
27kf
27kb
27kc
27kd
27ke
H
>300
>300
45ꢁ12
>300
27kg
27kah
27ki
50ꢁ15
>300
>300
1464
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1459 – 1470

Stat3–Stat3 Dimer Antagonists
into the Trp623/Phe716 hydrophobic subpocket was to
modify the biphenyl unit of 27g (IC50 =115 mm). The aryl bro-
mide moiety in 26a provides an excellent handle for facile
substitution reactions via Suzuki chemistry, facilitating access
to the desired biphenyl analogues of 27g. To this end, and as
described in Scheme 6, 26a was treated with a variety of aryl
boronic acids in the presence of catalytic Pd(PPh3)4 to furnish,
after the standard benzyl group deprotections, the series of
meta- and para-substituted biphenyl-based inhibitors 27la–lh
shown in Table 5. Likewise, the corresponding 4-(4-bromophe-
nyl)benzyl derivative 26m furnished the terphenyl-based inhib-
itors 27na–nh (Table 5).
As shown in Table 5, none of the biphenyl-based inhibitors
27la–27lh offered any improvement in Stat3 inhibitory activity
over the parent biphenyl inhibitor 27g (IC50 =115 mm). Howev-
er, excluding the carboxylic acid-substituted compounds 27nb
and 27nf, the terphenyl-based inhibitors 27na–nh proved
more potent that the parent inhibitor 27g, with the most
active compound 27nh disrupting the Stat3–Stat3:DNA terna-
ry complex with an IC50 value of 43 mm. The improved activity
Scheme 6. a) R4B(OH)2, Pd(PPh3)4, K2CO3,
DMF, 1008C, 24 h, 16–73%; b) LiOH·H2O,
THF/H2O (3:1), RT, 24 h, 76–99%; c) TFA/
toluene (1:2), RT, 16 h, 85%; d) H2, 10%
Pd/C, MeOH/THF (1:1), RT, 1–16 h, 85–
100%.
Table 5. EMSA inhibition data for disruption of the Stat3–Stat3:DNA complex in vitro by a series of biphenyl-
and terphenyl-based derivatives of inhibitor 27a.
tivity (<300 mm), with IC50 values
of 45 and 50 mm, respectively.
Both 27kd and 27kg share a 4-
benzonitrile moiety, so it may
seem curious that the related in-
hibitor 4-cyanobenzamide 27kf
demonstrated no inhibition of
Stat3. This result could be due
to the nature of the EMSA,
which is conducted on nuclear
extracts that contain various
other members of the STAT pro-
tein family. It may be the case
that 27kf is a potent inhibitor of
a different STAT isoform, decreas-
ing the concentration of free
compound available to inhibit
Stat3, leading to an apparent
IC50 lower than the actual value.
Compd
R4
IC50 [mm]
Compd
R4
IC50 [mm]
63ꢁ2
27g
115 ꢁ760
27na
27la
27lb
27lc
191ꢁ8
27nb
27nc
27nd
>300
82ꢁ1
90
250ꢁ23
>300
27gld
141ꢁ10
27ne
97ꢁ17
Biphenyl and terphenyl deriva-
tives
27le
27lf
27lg
>300
27nf
27ng
27nh
>300
After elaborating our inhibitors
through functionalization of the
4-position of the cyclohexyl moi-
eties in 27e (IC50 >300 mm) and
27h (IC50 =35 mm), the next logi-
cal approach to probe deeper
274ꢁ10
74ꢁ9
43ꢁ1.3
>300
ChemMedChem 2011, 6, 1459 – 1470
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1465

J. Turkson, P. T. Gunning, et al.
MED
of the terphenyl-based inhibitors over their biphenyl-based
counterparts is likely due, at least in part, to enhanced van der
Waals contacts between the larger terphenyl moieties and the
protein surface, possibly in the proposed Trp623/Phe716 sub-
pocket. The observation that the 4-carboxamide terphenyl
27nh was the most potent of the series is probably due to a
hydrogen bond between the carboxamide functional group
and the protein surface.
tion of the resultant primary alcohol to give 45 was nontrivial
and required the use of 20 equiv p-TsCl in order to suppress
symmetrical ether formation through the reaction of the start-
ing alcohol with the product tosylate. Debenzylation of 45 was
closely monitored, and fresh Pd catalyst (10 mol%) was added
every 2 h to minimize reaction time and the likelihood of loss
of the O-tosyl group through nucleophilic attack by the metha-
nol co-solvent. S3I-201 analogue 46 (SF-1-121)[16g] was thus fur-
nished in very good yield (85%).
The EMSA data for compounds 41 and 43 in Table 6 indicate
that changing the X=NCH3 group in compound 27h to NH or
NBoc, respectively, had a detrimental effect on Stat3 inhibitory
activity, increasing the IC50 value from 35 to ~100 mm. Howev-
Probing the Arg595/Ile634 hydrophobic subpocket: SAR of
the sulfonamide X group
To complete our research program, we modified the X=NCH3
component (X=O in S3I-201) whilst invoking optimized R1 and
R2 groups to help identify even more potent Stat3 inhibitors.
Once more, R1 was constrained as the 4-cyclohexylbenzyl
group. Thus, a focused variation of the NCH3 group in 27h
was executed. The substitutes chosen were the more hydro-
phobic NBoc group, the more polar NH group, and oxygen, af-
fording, in the latter case, a potentially irreversible inhibitor.
The syntheses of these target molecules are depicted in
Scheme 7. Briefly, secondary aniline 19h was coupled to TsN-
(Boc)CH2CO2H (39) using PPh3Cl2, which, due to the generation
of HCl in situ, led to the inadvertent loss of the Boc group to
furnish 40. Standard hydrogenolytic debenzylation of 40 gave
41 (SF-1-083),[16g] or, alternatively, the NH of 40 was re-tert-bu-
toxycarbonylated and then debenzylated as usual to deliver 43
(SF-1-087).[16g] To synthesize the labile O-tosyl analogue of 41,
compound 19h was first coupled to 2-acetoxyacetyl chloride
to produce 44. Simple hydrolysis of the acetate group in the
presence of the aryl benzyl ester proceeded smoothly. Tosyla-
Table 6. EMSA inhibition data for disruption of the Stat3–Stat3:DNA com-
plex in vitro by a series of X-substituted para-toluenesulfonyl analogues
of inhibitor 27h.
Compd
X
IC50 [mm]
27h
41
43
NCH3
NH
NBoc
O
35ꢁ9
95ꢁ9
115ꢁ35
43ꢁ13
46
er, more interestingly, the O-tosyl analogue 46 was equipotent
with the parent inhibitor 27h, within experimental error. Com-
pound 46, carrying the labile O-tosyl group, has the capacity
to function as an irreversible inhibitor, whilst 27h, with the
non-labile N(CH3)-tosyl moiety, possesses no such potential.
These data thus suggest that the inhibitory activity of 46 likely
arises chiefly from noncovalent interactions with the Stat3 SH2
domain. Interestingly, however, the similar activities of 46 and
27h are in stark contrast to the very different activities of the
analogous R1 =H derivatives S3I-201 (8) and 10, respectively,
for which replacement of the X=O atom with NCH3 abolished
Stat3 inhibitory activity (>300 mm). Taken together, these re-
sults suggest that the R1 =4-(cyclohexyl)benzyl moiety in 27h
and 46 contributes significantly to the inhibition of Stat3. Fur-
thermore, it is evident that the nature of the X group in the
S3I-201 scaffold plays a considerable role in the subsequent
Stat3 inhibitory activity, and for this reason our current re-
search efforts are focused toward a more extensive SAR study
of this group. This work shall be reported in due course.
We selected several of our analogues of S3I-201 (8) for more
thorough biophysical characterization by evaluating their in-
hibition of the Stat3 protein in isolation by using an in vitro
fluorescence polarization (FP) assay.[25] The principle of this
assay works on the decrease in FP that occurs upon displace-
ment of the 5-carboxyfluorescein (F*)-labeled Stat3 SH2
Scheme 7. a) TsN(Boc)CH2CO2H (46), PPh3Cl2, CHCl3, 608C, 12 h, 48%;
b) AcOCH2COCl, DIPEA, CH2Cl2, RT, 4 h, 64%; c) (Boc)2O, cat. DMAP, THF, 12 h,
81%; d) H2, 10% Pd/C, MeOH/THF (1:1), RT, 1–16 h, 85–94%; e) LiOH·H2O,
THF/MeOH/H2O (3:1:1); 89% f) p-TsCl, DIPEA, CH2Cl2, RT, 3 h, 85%.
1466
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1459 – 1470

Stat3–Stat3 Dimer Antagonists
tivities against Stat1, which ex-
hibits 78% sequence identity to
Stat3.[27] The results of our find-
ings are disclosed in Table 8 and
in the Supporting Information.
Table 7. EMSA and FP inhibition data of selected inhibitors for disruption of the Stat3–Stat3:DNA and Stat3–
gp130 sequence complexes in vitro.
Compound 27h exhibited
a
greater than threefold selectivity
for Stat3 (Stat3: Ki =15 mm,
Stat1: Ki ^50 mm). In contrast to
27h, the 4-cyanobenzenesulfon-
Compd
R1
H
X
R2
EMSA IC50 [mm]
86ꢁ33
FP Ki [mm]
>50
8 (S3I-201)
O
yl-based
compound
27kg
showed only limited isoform
specificity (Stat3: Ki =21 mm,
Stat1: Ki =28 mm), which, given
the structural similarities of
these two compounds, suggests
27h
NCH3
NCH3
35ꢁ9
15ꢁ4.9
27kd
45ꢁ12
18.7ꢁ1.4
that
the
4-cyclohexylbenzyl
group at the R1 position is also a
source of Stat3 isoform specifici-
ty.
27kd
27kd
NCH3
50ꢁ15
21.5ꢁ2.2
21.6ꢁ0.9
Whole-cell cytotoxicity
Agents 27h, 27kd, 27 kg, 27nh,
and 46 all show significantly im-
proved in vitro inhibitory activity
against Stat3 DNA binding activi-
ty (as determined by EMSA),
with IC50 values of 18.7–51.9 mm
(Tables 5–8) and Stat3–pTyr pep-
tide interaction in the FP assay,
with Ki values of 13–26.5 mm
(Tables 7 and 8 and Supporting
Information). For select active
compounds, whole-cell activities
were investigated by screening
inhibitors at a concentration of
NCH3
74.3ꢁ9.3
27kd
NCH3
43ꢁ1.3
16.7ꢁ1.7
domain inhibitor F*-GpYLPQTV from the Stat3 protein by the
small molecule of interest. Generally, the FP Ki data for the se-
lected inhibitors (Table 7) corroborate those data observed in
the EMSA, with potent activity in one assay reflected by potent
activity in the other. These data support our hypothesis that
the S3I-201 analogues disrupt the ternary Stat3–Stat3:DNA
complex, as quantified in the EMSA, through direct inhibition
of the Stat3 protein. It is reasonable to expect that disruption
of the Stat3–Stat3 dimer (full-length protein) bound to DNA in
the EMSA may be more difficult than disrupting the Stat3–
phosphopeptide interaction in the FP assay, and this is proba-
bly the reason why, in many cases, the FP-determined Ki values
were lower than the corresponding EMSA-determined IC50
values.
up to 100 mm across a range of human tumor cell lines,
namely breast cancer (MDA-MB-468),[28] prostate cancer
(DU145),[29] acute myeloid leukemia (OCI-AML-2),[30] and human
multiple myeloma (JJN-3), all of which harbor constitutively
active Stat3 (data not shown). The inhibitory activities (IC50
values) for select compounds are listed in Table 9.
Indeed, there was good correlation between the whole-cell
effects and the inhibition of Stat3 in nuclear extracts. Treat-
ment of cells with compounds 10, 14, 27a–e, 27i, 27ja–jd,
27ld, 27lh, and 27nd had no effect on cell growth at inhibitor
Table 8. Comparative Stat isoform selectivity as assessed by a Stat3 and
Stat1 FP assay.
Compd
Ki [mm]
Stat3
Stat1
STAT isoform selectivities
27h
27kd
27kg
27nh
15ꢁ5
8.4ꢁ1
21ꢁ2
8.4ꢁ2
>50
>50
Using a similar Stat1 SH2 domain FP-based binding assay,[26]
we also investigated the isoform selectivity of some of our
most potent Stat3 inhibitors by evaluating their inhibitory ac-
28ꢁ2
9.5ꢁ2
ChemMedChem 2011, 6, 1459 – 1470
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1467

J. Turkson, P. T. Gunning, et al.
MED
active or worse in the prostate
cancer DU145 cell line. Com-
pound 27h shows strong cellu-
lar effects (Table 10), consistent
with the inhibition of Stat3 activ-
ity in cells (Figure 4).
Table 9. IC50 values for selected R1-substituted analogues in whole-cell viability studies.
As detailed in Table 10, ana-
logues of S3I-201 (8) carrying
the optimized R1 =4-cyclohexyl-
benzyl and R2 =p-tolyl groups all
exhibited sub-100 mm activities
in the three Stat3-dependent
tumor cell lines. Compounds
27h (X=NCH3) and 41 (X=H)
were roughly equipotent. The
most potent compound of this
series in the Stat3-dependent
cell lines was compound 43, the
X=NBoc analogue of 27h, in-
Compd
R1
IC50 [mm]
MDA-MB-468[a]
17ꢁ4
DU145[b]
OCI-AML-2[c]
36ꢁ13
27h
37ꢁ12
>100
27kd
>100
>100
27kg
95ꢁ9
>100
>100
27ng
27nh
33ꢁ16
>100
94ꢁ7
>100
58ꢁ14
>100
hibiting
MDA-MB-468
cell
[a] MDA-MB-468: breast cancer. [b] DU145: prostate cancer. [c] OCI-AML-2: acute myeloid leukemia.
growth with an IC50 value of
10.5 mm. The improved whole-
cell activities of 43 relative to
27h is possibly due to the great-
concentrations <100 mm, reflecting their poor IC50 values in
the EMSA (data not shown), whereas the active Stat3 inhibitors
in the in vitro EMSA inhibited the growth of cells dependent
on constitutively active Stat3. Accordingly, the whole-cell activ-
ities observed with 27h, 27 kg, and 46 mirror their inhibitory
activities in the Stat3 DNA binding activity/EMSA (Table 2) and
the Stat3–pTyr peptide interactions in the FP assay (Tables 7
and 8). Agent 27h inhibited the growth of MDA-MB-468 breast
cancer cells with an IC50 value of 17 mm, of DU145 with IC50 =
37.2 mm, and of OCI-AML-2 with IC50 =35.9 mm, consistent with
its EMSA results or the Ki value for the inhibition of Stat3–pTyr
peptide interactions (Tables 7 and 8). Notably, although 27h in-
hibited Stat3 activity in EMSA with an IC50 value of 35 mm, it in-
hibited Stat3–pTyr peptide interactions with an IC50 value of
15 mm (Table 8), suggesting that in cells, it might be more ef-
fective to disrupt Stat3 binding to pTyr peptide motifs of re-
ceptors, as has been previously reported for dimerization dis-
ruptors by our group.[15,17g] Consistent with the findings from
the EMSA (Tables 3 and 4), the N-(4-piperidinyl)methyl- (27ja–
27d) and N-(4-piperidinyl)benzyl-based (27ka–kf) inhibitors
proved ineffective toward the growth of Stat3-dependent
tumor cell lines.
er hydrophobicity of the NBoc group over the NCH3 group,
which might facilitate more efficient cellular entry. Encourag-
ingly, when assessed by FP for Stat3 binding potency, com-
pound 43 was shown to have Ki =15ꢁ0.2 mm (Supporting In-
formation). The O-tosyl derivative 46, a potentially irreversible
inhibitor, demonstrated activities in the Stat3-dependent cell
lines that were around half that exhibited by the parent 27h.
These findings may, in part, be a consequence of the possible
hydrolysis of 46 to primary alcohol 47, which was synthesized
and found to show no inhibition of Stat3 in vitro (IC50 >
300 mm), nor appreciable whole-cell activity (MDA-MB-468:
IC50 >100 mm).
Table 10. IC50 values for a series of para-toluenesulfonyl X analogues of
inhibitor 27h in whole-cell viability studies.
Generally, a good correlation was observed between the
EMSA data for the biphenyl- and terphenyl-based compounds
(Table 5) and the whole-cell data for the Stat3-dependent cell
lines, particularly with the MDA-MB-468 cell line. It is probable
that the polar carboxamide functional groups in 27nd and
27nh hindered cellular entry of these compounds, hence their
poorer whole-cell activities than might have otherwise been
anticipated based on their activities in the EMSA. Generally
speaking, these series of compounds were equipotent at inhib-
iting the growth of breast cancer MDA-MB-468 cells and acute
myeloid leukemia OCI-AML-2 cells, and were around half as
Compd
X
IC50 [mm]
MDA-MB-468[a]
OCI-AML-2[b]
DU145[c]
27h
41
43
NCH3
NH
NBoc
O
17ꢁ4
21ꢁ6
10ꢁ9
43ꢁ6
37ꢁ12
38ꢁ14
24ꢁ16
52ꢁ13
36ꢁ13
33ꢁ8
18ꢁ9
62ꢁ12
46
[a] MDA-MB-468: breast cancer. [b] OCI-AML-2: acute myeloid leukemia.
[c] DU145: prostate cancer.
1468
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1459 – 1470

Stat3–Stat3 Dimer Antagonists
Inhibition of intracellular aberrant Stat3 phosphorylation
and the induction of known Stat3-regulated genes
only two of these subdomains. Importantly, with a labile O-
tosyl group a to a carbonyl group, S3I-201 has the capacity to
operate, at least in part, as an irreversible inhibitor, which
would be anticipated to lead to poor protein selectivity pro-
files.
Consistent with the effects on viability, 27h strongly inhibited
constitutively active Stat3 in tumor cells, including the human
breast cancer (MDA-MB-468) and multiple myeloma (JJN-3)
lines, as measured by Western blot analysis (Figure 4A), con-
Experimental Section
Anhydrous solvents MeOH, DMSO, CH2Cl2, THF, and DMF were pur-
chased from Sigma–Aldrich and were used directly from Sure-Seal
bottles. Molecular sieves were activated by heating at 3008C under
vacuum overnight. All reactions were performed under an atmos-
phere of dry N2 in oven-dried glassware and were monitored for
completeness by thin-layer chromatography (TLC) using silica gel
(visualized by UV light, or developed by treatment with KMnO4
1
stain or phosphomolybdic acid stain). H and 13C NMR spectra were
recorded on Bruker 400 MHz and Varian 500 MHz spectrometers
in either CDCl3, CD3OD, or [D6]DMSO. Chemical shifts (d) are report-
ed in ppm after calibration to residual isotopic solvent. Coupling
constants (J) are reported in Hz. Before biological testing, inhibitor
purity was evaluated by reversed-phase HPLC (RP-HPLC). Analysis
by RP-HPLC was performed by using a Microsorb-MV 300A C18
250ꢂ4.6 mm column run at 1 mLminꢀ1, and using gradient mix-
tures of A) H2O with 0.1m CH3COONH4 and B) MeOH. Ligand purity
was confirmed by using linear gradients from 75% A and 25% B
to 100% B after an initial 2 min period of 100% A. The linear gradi-
ent consisted of a changing solvent composition of either I) 4.7%
per minute and UV detection at l 254 nm, or II) 1.4% per minute
and detection at l 214 nm, each ending with 100% B for 5 min.
For reporting HPLC data, percentage purity is given in parentheses
after the retention time for each condition. All biologically evaluat-
ed compounds are of >95% chemical purity as measured by
HPLC. Full characterization for all final compounds and intermedi-
ate compounds are provided in the Supporting Information.
Figure 4. Western blot analysis showing A) inhibition of Stat3 phosphoryla-
tion (pYStat3) and B) repression of Stat3-regulated gene products, Bcl-xL
and Survivin, in human breast (MDA-MB-468) and multiple myeloma (JJN-3)
cells as a function of treatment with 27h (100 mm, 24 h). Glyceraldehyde 3-
phosphate dehydrogenase (GAPDH) was used as a lane loading control.
firming that select agents inhibit aberrant Stat3 activation in
tumor cells. Furthermore, treatment with 27h inhibited the ex-
pression of Bcl-xL and Survivin, the genes for which are known
to be regulated by Stat3 (Figure 4B). These findings suggest
that the modulation of aberrant Stat3 in MDA-MB-468 and JJN-
3 cells leads to suppression of Stat3-mediated gene regulation.
These events contribute to the loss of viability observed fol-
lowing the treatment of malignant cells that harbor aberrant
Stat3 activity by the newly identified small-molecule inhibitors.
Acknowledgements
We thank Jeffrey L. Wrana and Alessandro Datti (Mt. Sinai Hospi-
tal) for access to the SMART high-throughput screening facility.
This work was supported by the Leukemia and Lymphoma Soci-
ety of Canada (P.T.G.), NSERC (P.T.G.), the University of Toronto
(P.T.G.), and by National Cancer Institute grants CA106439 (J.T.)
and CA128865 (J.T.). A.S. is a Leukemia and Lymphoma Society
Scholar in Clinical Research.
Conclusions
In summary, we have conducted an extensive SAR study cen-
tered on the previously identified Stat3 inhibitor S3I-201 (8) to
derive analogues with improved Stat3 inhibitory activity. These
studies have led to the identification of several diverse classes
of agents equipped with an additional appendage that pro-
motes interaction with the hitherto unexplored pocket on the
Stat3 protein surface, thereby intensifying the binding to Stat3
and enhancing the Stat3 inhibitory activity. Specifically, com-
pounds 27h, 27nh, 27kd, 27 kg, and 46 all show significantly
improved in vitro inhibitory activity against Stat3, with IC50
values of 18.7–51.9 mm. Moreover, at these concentrations,
select compounds inhibit constitutively active Stat3 and Stat3
tyrosine phosphorylation in malignant cells and promote anti-
tumor cell effects consistent with the inhibition of aberrant
Stat3 activity. The improved inhibitory activity against Stat3 ac-
tivation is derived in part from the successful occupation of
the third subdomain of the Stat3 SH2 domain, as supported by
computational modeling; S3I-201 can simultaneously occupy
Keywords: antitumor agents
·
molecular recognition
·
molecular therapeutics · protein–protein interactions · stat3
[1] a) R. Buettner, L. B. Mora, R. Jove, Clin. Cancer Res. 2002, 8, 945–954;
b) J. E. Darnell, Jr., Nat. Med. 2005, 11, 595–596; c) T. Bowman, R. Garcia,
J. Turkson, R. Jove, Oncogene 2000, 19, 2474–2488.
[2] M. Berishaj, S. P. Gao, K. Leslie, H. Al-Ahmadie, W. L. Gerald, W. Born-
mann, J. F. Bromberg, Breast Cancer Res. 2007, 9, R32.
[3] J. Abdulghani, L. Gu, A. Dagvadorj, J. Lutz, B. Leiby, G. Bonuccelli, M. P.
Lisanti, T. Zellweger, K. Alanen, T. Mirtti, T. Visakorpi, L. Bubendorf, M. T.
Nevalainen, Am. J. Pathol. 2008, 172, 1717–1728.
[4] W. M. Burke, X. Jin, H.-J. Lin, M. Huang, R. Liu, K. R. Reynolds, J. Lin, On-
cogene 2001, 20, 7925–7934.
[5] M. Carro, W. K. Lim, M. J. Alvarez, R. J. Bollo, X. Zhao, E. Y. Snyder, E. P.
Sulman, S. L. Anne, F. Doetsch, H. Colman, A. Lasorella, K. Aldape, A. Cal-
ifano, A. Iavarone, Nature 2010, 463, 318–325.
ChemMedChem 2011, 6, 1459 – 1470
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1469

J. Turkson, P. T. Gunning, et al.
MED
[6] H. Akca, M. Tani, T. Hishida, S. Matsumoto, Yokota, J. Lung Cancer 2006,
54, 25–33.
[7] a) G. Inghirami, R. Chiarle, W. J. Simmons, R. Piva, K. Schlessinger, D. E.
Levy, Cell Cycle 2005, 4, 1131–1133; b) K. Schlessinger, D. E. Levy, Cancer
Res. 2005, 65, 5282–5834.
[8] J. Bromberg, J. E. Darnell, Jr., Oncogene 2000, 19, 2468–2473.
[9] J. F. Bromberg, M. H. Wrzeszczynska, G. Devagan, Y. Zhao, R. G. Pestell,
C. Albanese, J. E. Darnell, Jr., Cell 1999, 98, 295–303.
[10] a) S. Fletcher, J. Turkson, P. T. Gunning, ChemMedChem 2008, 3, 1159–
1168; b) S. Fletcher, J. Drewry, V. Shahani, B. D. G. Page, P. T. Gunning,
Biochem. Cell Biol. 2009, 87, 825–833; c) B. D. G. Page, D. Ball, P. T. Gun-
ning, Expert Opin. Ther. Pat. 2011, 21, 1–19; d) S. Haftchenary, M. Avadi-
sian, P. T. Gunning, Anti-Cancer Drugs 2011, 22, 115–127.
[11] K. A. Z. Siddiquee, P. T. Gunning, M. Glenn, W. P. Katt, S. Zhang, C.
Schroeck, S. M. Sebti, R. Jove, A. D. Hamilton, J. Turkson, ACS Chem. Biol.
2007, 12, 787–798.
[12] a) M. Schrçder, K. Kroeger, H. D. Volk, K. A. Eidne, G. Grꢃtz, J. Leukoc.
Biol. 2004, 75, 792–797; b) P. B. Sehgal, Semin. Cell Dev. Biol. 2008, 19,
329–340.
[13] V. Paillard, A. Kaptein, M. Saunders, J. Biol. Chem. 1996, 271, 5961–5964.
[14] T. Faruqi, D. Gomez, X. Bustelo, D. Bar-Sagi, N. Reich, Proc. Natl. Acad.
Sci. USA 2001, 98, 9014–9019.
[17] a) J. Schust, B. Sperl, A. Hollis, T. U. Mayer, T. Berg, Chem. Biol. 2006, 13,
1235–1242; b) L. Lin, B. Hutzen, P.-K. Li, S. Ball, M. Zuo, S. DeAngelis, E.
Foust, M. Sobo, L. Friedman, D. Bhasin, L. Cen, C. Li, J. Lin, Neoplasia
2010, 12, 39–50; c) H. Song, R. Wang, S. Wang, J. Lin, Proc. Natl. Acad.
Sci. USA 2005, 102, 4700–4705; d) P. T. Gunning, M. Glenn, K. A. Z. Siddi-
quee, W. P. Katt, E. Masson, S. M. Sebti, J. Turkson, A. D. Hamilton, Chem-
BioChem 2008, 9, 2800–2803; e) K. A. Z. Siddiquee, P. T. Gunning, M.
Glenn, W. P. Glenn, S. M. Sebti, R. Jove, A. D. Hamilton, J. Turkson, ACS
Chem. Biol. 2007, 12, 787–796; f) K. Siddiquee, S. Zhang, W. C. Guida,
M. A. Blaskovich, B. Greedy, H. R. Lawrence, M. L. R. Yip, R. Jove, M. M.
Laughlin, N. J. Lawrence, S. M. Sebti, J. Turkson, Proc. Natl. Acad. Sci.
USA 2007, 104, 7391–7396; g) S. Fletcher, J. Singh, X. Zhang, P. Yue,
Page, B. D. G. , S. Sharmeen, V. M. Shahani, W. Zhao, A. D. Schimmer, J.
Turkson, P. T. Gunning, ChemBioChem 2009, 10, 1959–1964; h) V. M.
Shahani, P. Yue, S. Haftchenary, W. Zhao, J. Lukkarila, X. Zhang, D. Ball,
C. Nona, P. T. Gunning, J. Turkson, ACS Med. Chem. Lett. 2011, 2, 79–81.
[18] a) J. Turkson, S. Zhang, J. Palmer, H. Kay, J. Stanko, L. B. Mora, S. Sebti,
H. Yu, R. Jove, Mol. Cancer Ther. 2004, 3, 1533–1542; b) J. Turkson, S.
Zhang, L. B. Mor, S. Sebti, R. Jove, J. Biol. Chem. 2005, 280, 32979–
32988.
[19] G. Jones, P. Willett, R. C. Glen, A. C. Leach, R. Taylor, J. Mol. Biol. 1997,
267, 727–748.
[15] a) J. Turkson, D. Ryan, J. S. Kim, Y. Zhang, Z. Chen, E. Haura, A. Laudano,
S. M. Sebti, A. D. Hamilton, R. Jove, J. Biol. Chem. 2001, 276, 45443–
45455; b) Z. Ren, L. A. Cabell, T. S. Schaefer, J. S. McMurray, Bioorg. Med.
Chem. Lett. 2003, 13, 633–636.
[20] S. Zhang, Z. Y. Zhang, Drug Discovery Today 2007, 12, 373–381.
[21] T. Maegawa, Y. Fujita, A. Sakurai, A. Akashi, M. Sato, K. Oono, H. Sajiki,
Chem. Pharm. Bull. 2007, 55, 837–839.
[22] P. N. Pandey, M. L. Purkayastha, Synthesis 1982, 876–878.
[23] S. Fletcher, P. T. Gunning, Tetrahedron Lett. 2008, 49, 4817–4819.
[24] C. Valerio, F. Moulines, J. Ruiz, J.-C. Blais, D. Astruc, J. Org. Chem. 2000,
65, 1996–2002.
[25] J. Schust, T. Berg, Anal. Biochem. 2004, 330, 114–118.
[26] P. Wu, M. Brasseur, U. Schindler, Anal. Biochem. 1997, 249, 29–36.
[27] W. Zhao, S. Jaganathan, J. Turkson, J. Biol. Chem. 2010, 285, 35855–
35865.
[28] F. Alimirah, J. Chen, Z. Basrawala, H. Xin, D. Choubey, FEBS Lett. 2006,
580, 2294–2300.
[16] a) J. Turkson, J. S. Kim, S. Zhang, J. Yaun, M. Haung, Glenn, M. P. E.
Haura, S. M. Sebti, J. Turkson, A. D. Hamilton, Mol. Cancer Ther. 2004, 3,
261–269; b) P. K. Mandal, P. A. Heard, Z. Ren, X. Chen, J. S. McMurray,
Bioorg. Med. Chem. Lett. 2007, 17, 654–656; c) J. Chen, L. Bai, D. Ber-
nard, Z. Nikolovska-Coleska, C. Gomez, J. Zhang, H. Yi, S. Wang, ACS
Med. Chem. Lett. 2010, 1, 85–89; d) P. T. Gunning, W. P. Katt, M. P. Glenn,
K. A. Z. Siddiquee, J. S. Kim, R. Jove, S. M. Sebti, J. Turkson, A. D. Hamil-
ton, Bioorg. Med. Chem. Lett. 2007, 17, 1875–1878; e) D. R. Coleman IV,
K. Kaluarachchi, Z. Ren, X. Chen, J. S. McMurray, Int. J. Pept. Res. Ther.
2008, 14, 1–9; f) J. Dourlat, B. Valentin, W.-C. Liu, C. Garbay, Bioorg. Med.
Chem. Lett. 2007, 17, 3943–3946; g) J. Chen, Z. Nikolovska-Coleska, C.-Y.
Yang, C. Gomez, W. Gao, K. Krajewski, S. Jiang, P. Roller, S. Wang, Bioorg.
Med. Chem. Lett. 2007, 17, 3939–3942; h) V. M. Shahani, P. Yue, S.
Fletcher, S. Sharmeen, M. A. Sukhai, D. P. Luu, X. Zhang, H. Sun, W.
Zhao, A. D. Schimmer, J. Turkson, P. T. Gunning, Bioorg. Med. Chem.
2011, 19, 1823–1838.
[29] C. Wang, J. E. Curtis, M. D. Minden, E. A. McCulloch, Leukemia 1989, 3,
264–269.
[30] D. A. N. Carlesso, Frank, J. D. Griffin, J. Exp. Med. 1996, 183, 811–820.
Received: April 16, 2011
Published online on May 25, 2011
1470
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1459 – 1470