Developing novel C-4 analogues of pyrrole-based antitubulin agents: Weak but critical hydrogen bonding in the colchicine site

Article abstract of DOI:10.1039/c2md20320k

The synthesis, biological evaluation and molecular modeling of a series of pyrrole compounds related to 3,5-dibromo-4-(3,4-dimethoxyphenyl)-1H-pyrrole-2- carboxylic acid that evaluates and optimizes C-4 substituents are reported. The key factor for microt

Full text of DOI:10.1039/c2md20320k

MedChemComm
View Article Online
View Journal | View Issue
CONCISE ARTICLE
Developing novel C-4 analogues of pyrrole-based
antitubulin agents: weak but critical hydrogen bonding
in the colchicine site†
Cite this: Med. Chem. Commun., 2013,
4, 417
Chenxiao Da,a Nakul Telang,b Kayleigh Hall,b Emily Kluball,b Peter Barelli,b
Kara Finzel, Xin Jia, John T. Gupton, Susan L. Mooberry and Glen E. Kellogg
b
b
b
c
a
*
Received 24th October 2012
Accepted 10th December 2012
The synthesis, biological evaluation and molecular modeling of a series of pyrrole compounds related to
3,5-dibromo-4-(3,4-dimethoxyphenyl)-1H-pyrrole-2-carboxylic acid that evaluates and optimizes C-4
substituents are reported. The key factor for microtubule depolymerization activity appears to be the
presence of an appropriately positioned acceptor for Cys241b in the otherwise hydrophobic subpocket A.
DOI: 10.1039/c2md20320k
www.rsc.org/medchemcomm
We previously reported the synthesis and modeling of a
series of pyrrole-based antitubulin agents targeting the colchi-
Introduction
cine site5 and predicted, based on a series of C-2 analogues, two
distinct binding modalities distinguishing highly active
analogues from those with weaker activities.6 The most active
analogue, 3,5-dibromo-4-(3,4-dimethoxyphenyl)-1H-pyrrole-2-
carboxylic acid ethyl ester (2, JG-03-14, Fig. 1), demonstrated
potent antiproliferative activity against a wide range of cancer
cell lines, strong microtubule-destabilizing activity and is a poor
substrate of the P-glycoprotein drug efflux pump that transports
taxanes and vinca alkaloids.7 Further studies showed that the
compound disrupts multiple endothelial cell functions, sug-
gesting the potential for vascular-disrupting activities.8 In this
respect, 2 has become a valuable lead candidate and the ve
atoms on its pyrrole scaffold can be easily modied for struc-
tural-activity relationships (SAR), providing a basis for the
future optimization and development.
In this study, we retained the two bromine groups at C-3 and
C-5 and the ethyl ester at the C-2 position and focused on
modications to the 3,4-dimethoxylphenyl ring at the C-4 posi-
tion. Previously, we showed that 20s ethyl ester at C-2 is an ideally
suited substituent for that position and this induces the 3,4-
dimethoxylphenyl moiety of 2 to overlap with ring A of colchicine
in the colchicine site and bind in a subpocket formed mainly by
hydrophobic residues and one polar residue, Cysb241.6 Here, we
Microtubules have been recognized as a target for cancer
treatment for a long period of time, owing to their multifarious
and critical functions in eukaryotic cells, such as maintenance
of cell shape, protein trafficking, signaling and segregation of
chromosomes during mitosis. Traditionally, microtubule-tar-
geting agents are classied as microtubule stabilizing or
destabilizing agents based on their effects on microtubule
polymer mass at high concentrations. A more practical classi-
cation in terms of drug design divides them according to their
binding sites on tubulin, which are the taxane domain, the
vinca domain, the colchicine site and new sites discovered as
more structurally diverse agents are developed.1 Unlike the
taxanes and vinca alkaloids, neither colchicine (1) nor any
colchicine site agents have been successful in cancer chemo-
therapy due to their severe toxicity to normal tissues.2 However,
as an alternative to conventional therapy, the combretastatins,
which also bind in the colchicine site, have been extensively
developed and some are progressing through clinical trials
(including Phase III) as antitumor vascular disrupting agents.1,3
Moreover, a recent study highlighted that colchicine site agents
might be able to circumvent bIII-tubulin overexpression, which
is related to the emerging drug resistance to taxanes and vinca
alkaloids.4
aDepartment of Medicinal Chemistry & Institute for Structural Biology and Drug
Discovery, Virginia Commonwealth University, Richmond, Virginia, USA
23298-0540. E-mail: glen.kellogg@vcu.edu; Tel: +1 804 828-6452
bDepartment of Chemistry, Gottwald Center for the Sciences, University of Richmond,
Richmond, Virginia, USA 23173
cDepartment of Pharmacology, University of Texas Health Science Center at San
Antonio, San Antonio, Texas, USA 78229-3900
† Electronic supplementary information (ESI) available: Experimental details for
the synthesis and characterization of reported compounds, procedures for the
assays and additional information on the computational methods utilized. See
DOI: 10.1039/c2md20320k
Fig. 1 Structures of colchicine and lead compound JG-03-14.
Med. Chem. Commun., 2013, 4, 417–421 | 417
This journal is ª The Royal Society of Chemistry 2013

View Article Online
MedChemComm
Concise Article
explore the electronic, hydrogen bonding and hydrophobic
characteristics of substituents at C-4 to enrich our understanding
of the SAR of these compounds. We report here the integrated
synthesis, microtubule depolymerizing and antiproliferative
effects and modeling results for a focused set of analogues.
Structure–activity relationships
All structural modications for this study were at the C-4 posi-
tion of the pyrrole core. Antiproliferative activities as well as
microtubule depolymerizing activities were measured (Table 1).
Compounds 5a–5d showed very weak or barely any effect on
microtubule polymerization with EC50 values greater than 75
mM. Compound 5a, the unsubstituted ring analogue, showed
negligible antiproliferative activity, while this activity for 5b–5d
(especially 5c with an IC50 of 0.919 mM) likely indicates a
different mechanism of action, although some form of micro-
tubule inhibition may still be responsible. For the rest of the
compounds, 2 and 5e–5q (and 1), the microtubule inhibitory
activity correlates well with the antiproliferative activity
(pEC50 ¼ 1.10 pIC50 ꢀ 1.57, r2 ¼ 0.79, Fig. 2). Interestingly, for
these compounds, pEC50 ꢀ pIC50 ¼ 1.00 ꢁ 0.43 mM, which
indicates that microtubule inhibition is consistently one order
of magnitude weaker than overall inhibition of proliferation.
The SAR was analyzed with respect to the EC50. The active
lead compound 2 (0.490 mM) bore two methoxy groups on the
phenyl ring attached at the C-4 position. Removing either of the
methoxys showed a signicant decrease in activity by 14-fold
(5e, 7.0 mM) and 5-fold (5f, 2.4 mM), respectively; as noted above,
a complete loss of microtubule inhibitory activity was observed
when all ring substitutions were eliminated (5a), suggesting the
signicance of both methoxys with a particular preference for
the meta-methoxy group. When compared to 5e, replacing the
hydrophobic methyl with a more polar triuoromethyl while
Results and discussion
Chemistry
We have previously reported9 the synthesis of 2 and have utilized
a similar strategy (Method A) to prepare analogues 5a–5i (Table
1): the appropriate aryl vinamidinium salt (3a–3i) was condensed
with glycine ethyl ester to yield the pyrrole ethyl esters 4a–4i. For
analogues 5j–5p (Table 1), the ester intermediates 4j–4p were
prepared (Method B) by a Suzuki cross-coupling of 4-bromo-1H-
pyrrole-2-carboxylic acid ethyl ester (6) with the appropriate aryl
boronic acid. The nal dibromination step for all analogues was
accomplished with pyridinium tribromide in DMF. Compound
5q was prepared by bromination of 2 with dibromdimethylhy-
dantioin. See Scheme S1 (ESI†) for synthetic details.
Biological activity
Antiproliferative activities were measured in MDA-MB-435
cancer cells using the sulforhodamine B assay and effects on
cellular microtubules were evaluated in A-10 cells using
immunouorescence as previously described.7 Results are pre-
sented in Table 1.
Table 1 Structures, antiproliferative and microtubule inhibitory activities of pyrrole compounds
Antiproliferation
Microtubule depolymerization
Cmpd
R
IC50 (mM)
EC50 (mM)
pEC50
HINT score
1 (Colchicine)
2 (JG-03-14)
—
0.016 ꢁ 0.002
0.036 ꢁ 0.002
10.3 ꢁ 1.3
0.030
0.490
>75
7.52
6.31
3.82b
3.82b
3.82b
4.03
5.15
5.62
3.82b
5.15
4.75
4.57
4.73
5.00
4.85
3.82b
4.53
4.68
4.85
549
643
170
579
754
815
563
558
124
805
271
649
541
1012
567
577
428
590
781
3,4-Dimethoxylphenyl
Phenyl
4-Methylphenyl
4-Chlorophenyl
4-Bromophenyl
4-Methoxylphenyl
3-Methoxylphenyl
3,4,5-Trimethoxylphenyl
1-Napthyl
5a
5b
5c
5d
5e
5f
5g
5h
5i
5j
5k
5l
5m
5n
5o
5p
5q
2.24 ꢁ 0.2
>75
0.919 ꢁ 0.020
0.312 ꢁ 0.020
0.843 ꢁ 0.090
0.633 ꢁ 0.01
12.9 ꢁ 1.9
>75
ꢂ94a
7.0
2.4
>75
7.0
3.24 ꢁ 0.20
1.98 ꢁ 0.20
1.70 ꢁ 0.10
0.626 ꢁ 0.020
0.806 ꢁ 0.060
0.539 ꢁ 0.040
1.99 ꢁ 0.20
1.80 ꢁ 0.20
4.36 ꢁ 0.3
3-Indolyl
17.8
27.1
18.5
9.9
14.1
>75
29.7
20.9
14.0
4-Triuoromethoxylphenyl
4-Thiomethylphenyl
3,4-Dichlorophenyl
3-Fluoro-4-methoxylphenyl
6-Ethoxyl-2-napthyl
1,3-Benzodioxol-6-yl
1,4-Benzodioxan-6-yl
2-Bromo-4,5-dimethoxylphenyl
2.64 ꢁ 0.30
a
b
40% microtubule loss at 75 mM, EC50 ꢂ 75/(2 ꢃ 0.4) ¼ 94 mM. Assumed EC50 ¼ 150 mM.
418 | Med. Chem. Commun., 2013, 4, 417–421
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
mostly buried in b-tubulin. It is surrounded by helices H7 and
H8, loop T7 and strands S8 and S9 of b-tubulin and loop T5 of
a-tubulin. Comparison of crystal structures of ab-tubulin het-
erodimers complexed with different ligands reveals the exibility
of the colchicine site, especially for loops T7 and T5. To under-
stand the movement of the sidechains and backbones
surrounding the site, we performed docking studies with ve
crystal structures (PDBIDs: 1sa0, 1sa1, 3hkc, 3hkd and 3hke).10,11
Docking poses were generated by GOLD12 and the resulting
complexes were minimized in Sybyl with the Tripos forceeld13
and rescored with HINT.14,15 These results showed that the
compounds tended to bind most favorably to the 3hkc model as
indicated by higher HINT scores. While 1sa0 is complexed with
colchicine and is thus frequently used for docking colchicine site
agents, 3hkc, is co-crystallized with the structurally unrelated
N-{2-[(4-hydroxylphenyl)amino]pyridin-3-yl}-4-methoxybenzene-
sulfonamide (ESI, Fig. S5†). In docking compounds 2 and 5a–5q,
however, the T5 loop of 3hkc appears to adapt and benet from
hydrogen bonding between the backbone carbonyl of Thr179a
and the pyrrole nitrogen, while in colchicine binding, T5 yields
to colchicine's amide chain as seen in 1sa0 (Fig. 3). This binding
mode is the same as we previously reported.6 The ester chain of 2
and 5a–5q partially overlaps with ring C of colchicine, tting into
subpocket C with the carbonyl oxygen forming a hydrogen bond
with the backbone nitrogen of Val181a. The pyrrole core locates
in the center of the site, forming hydrogen bonds with Asn258b
and Thr179a. The phenyl moiety overlaps with ring A of colchi-
cine, inserting into the hydrophobic subpocket A, which is
formed by Tyr202b, Val238b, Thr239b, Leu242b, Leu248b,
Leu252b, Ile378b and Val318b, with Leu248b and Leu255b
clamping the phenyl moiety. One polar residue, Cys241b,
donates to the ligand in the presence of an appropriately posi-
tioned acceptor. The presence of this latter residue in subpocket
A explains the importance of hydrogen bond accepting character
in C-4 substituents observed in the SAR studies.
Fig. 2 Correlation of pEC50 and pIC50. Compounds indicated by closed circles are
not included in correlation. They have high antiproliferative activity but negligible
microtubule inhibition, which indicates an alternative mechanism of action.
retaining the acceptor ether oxygen (5j) or with the weaker
sulfur acceptor (5k) resulted in minor losses in activity by 4-fold
and 2.5-fold, respectively. Furthermore, attempting to recover
activity with hydrophobic groups at the para-position with 5b
(methyl), 5c (chloro) and 5d (bromo) was completely ineffective
with respect to microtubule inhibition, although the anti-
proliferative activity for these analogues increases with
substituent hydrophobicity. Overall, these results suggest that
the hydrogen bonding properties of the C-4 ring substituents
play the more critical role in microtubule inhibition, although
clearly the ether oxygen in –OMe may also serve to place the
hydrophobic methyl in a more ideal position.
Addition of a second chlorine at the meta-position recovered
activity (5l, 9.9 mM). This may be partially explained by the weak
hydrogen bond accepting character of chlorine, but also its
placement in the meta-position is a factor – as was seen in the
comparison between 5f and 5e. Probably because uorine is less
hydrophobic and smaller than chlorine (although a stronger
acceptor), the uorinated compounds, 5m, was no more effec-
tive as a microtubule inhibitor than its des-uoro analogue 5e.
The inhibitory activity observed for large aromatic rings as C-4
substituents (5h and 5i) can be attributed to their hydrophobic-
ities and also the hydrogen bond acceptor character of the
aromatic p-clouds in napthyl and indolyl. To further investigate
this putative hydrogen bonding, the H-bond acceptor was repo-
sitioned with 6-ethoxyl-2-napthyl at C-4 (5n) with negative effect.
Restriction of the rotation of two methyl groups was achieved
by rst forming a methylene bridge between two oxygens (5o),
which led to a 60-fold decrease in activity compared to 2.
Secondly, an ethylene bridge (5p) fared somewhat better with
only a 40-fold activity decrease. Interestingly, addition of a third
methoxy to the phenyl ring at its 5-position, as in 5g, did not lead
to the expected increase, but, instead, a total loss in activity;
however, placing a bromine at the ring's 2-position and removing
the 3-methoxy (5q) produced a >5-fold activity increase over 5g.
Detailed analysis of the binding conformations assists
further interpretation of the SAR (Fig. 4). In the case of 2, the
thiol hydrogen of Cys241b is pointed towards the methoxy at the
meta-position and away from the para-position. This was
observed for all other cases owing to the steric clashes that the
thiol hydrogen could encounter if oriented in the other
Molecular modelling
Fig. 3 Colchicine (green) and binding modes of pyrrole-based C-4 analogs in red
(compound 2 in heavy sticks). The extents of the colchicine site, as illustrated by
Modelling was performed to rationalize the observed SAR. The
colchicine site is located at the interface of a- and b-tubulin and MOLCAD, are shown in grayish white.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 417–421 | 419

View Article Online
MedChemComm
Concise Article
Fig. 5 HINT H-bond component score for ring interactions with Cys241b. Closed
squares represent compounds possessing both H-bond acceptors and appropri-
ately placed hydrophobic groups. These compounds generally possess superior
pEC50s. Open triangles represent compounds with weak or no acceptors. Open
circles represent compounds with steric issues and/or lacking key hydrophobic
interactions.
Fig. 4 Specific hydrogen bonding (yellow) and hydrophobic (green) interactions
in subpocket A. Compounds 2 (solid white), 5h (translucent purple), 5i (translucent
green) and 5m (translucent red) are shown. Notes: (1) the key H-bond interaction is
with Cys241b, which is strongest with the O of methoxy in the ring's meta position;
(2) some analogues, e.g., with F, can weakly H-bond with the NH of Leu252b; (3)
the CH3 of p-methoxy has key hydrophobic interactions with Leu242b; (4) Ile378b
has hydrophobic interactions with m-methoxy or the rings of 5h or 5i.
two-fold; rst, the hydrogen bonding interaction with Cys241b is
the key predictor, absent of steric clashes, for the microtubule
inhibitory activity for this set of analogues; second, other inter-
actions in the pocket, i.e., hydrophobic, are also necessary, but
competitive with this weakly scored hydrogen bonding.
direction. The better hydrogen bonding for a meta-position
substituent explains the activity of 5f compared to 5e and other
similar cases. As for 5h and 5i, the distal (from the pyrrole core)
rings were located directly beneath the thiol hydrogen, thus
acting as acceptors for the weak but critical hydrogen bond, but
pocket steric issues cancelled this advantage. The uorine atom
of 5m is also located at the meta-position, but the docking study
suggested that an 180ꢄ ring ip shied its position in space
such that, although the uorine was anchored by the backbone
NH of Leu252b, it provided no additional bonding to Cys241b
compared to 5e. In the case of 5g, the detrimental effect of the
The importance of both hydrophobic interactions and
hydrogen bonding in subpocket A was seen in the SAR analysis
and modeling studies. The latter dictates whether the C-4
analogues of pyrrole-based antitubulin agents display micro-
tubule depolymerizing activity and the strength of that activity,
while the character of the pocket requires predominantly
hydrophobic moieties. Underestimation of the Cys241b inter-
action was one probable reason that the total HINT score was a
poor predictor of microtubule depolymerizing activity. This
thiol group acts as a hydrogen bond donor and while this type of
hydrogen bonding interaction is generally regarded as weak and
is thusly parameterized by HINT, it is not even considered by
many other scoring functions. For the downstream biological
effect, microtubule depolymerization, the interaction assumed
to be weak surprisingly stands out as a key factor. In fact, its
absence might produce a different mechanism of action even
when other portions of the structure are exactly the same, as
shown particularly by 5d with potent antiproliferative activity
(0.312 mM) but weaker microtubule depolymerization activity
(ꢂ94 mM). Cys241b has been previously identied as an
important target residue for colchicine site agents.16 In a study
of 15 structurally diverse colchicine site inhibitors, the docked
binding modes of all included hydrogen bonding to Cys241b
˚
third methoxy is visually apparent: the tight distance (3.58 A)
between the backbone of Leu252b and the phenyl ring of the
ligand can lead to signicant steric clashes with a large
substituent such as the 5-methoxy.
The total HINT scores of C-4 analogues fail to show a tight
relationship with pEC50 (ESI, Fig. S6†). However, isolating
the HINT score for hydrogen bonding interactions involving
Cys241b for a subset of analogues (2, 5e, 5f, 5j, 5k, 5m and 5q)
that place, as separate entities, appropriately positioned hydro-
phobic groups and a hydrogen bond acceptor in the subpocket
(while not inducing steric clashes),‡ reveals a linear relation with
respect to these compounds' pEC50s (Fig. 5). The implications are
‡ Compounds 5a and 5b do not possess H-bond acceptor substituents to the
phenyl ring. The halogens of 5c, 5d and 5l are considered as weak acceptors by (Cys239b in that study).17 Our combined SAR and modeling
HINT and, while their dominant property is hydrophobicity, these halogen
study conrms the importance of that cysteine. It should be
noted that there is potentially a systematic error in our proce-
dure. As GOLD optimizes ligand placement with a different
atoms are not at ideal hydrophobic contact distance. Similarly, the aromatic
rings of 5h and 5i can simultaneously be weak hydrogen bond acceptors and
strongly hydrophobic, but these analogues are inappropriately shaped to make
forceeld (set of rules) than used by HINT in scoring, subtle
all necessary hydrophobic contacts in the pocket (see Fig. 4). Compounds 5g,
5n–5p have a number of structural or conformational issues as they are docked structural effects, or in this case, the interplay of several of
in subpocket A: there is insufficient room for the trimethoxyphenyl of 5g or the
them, are not well scored post-docking as none of the models
generated by GOLD capture the set of features in a single model
that HINT would score highest. This is likely to be a general
ethoxyl-2-napthyl of 5n; and although the two ether oxygens of 5o and 5p can
participate in hydrogen bonding with Cys241b as in 2, the methylene and
ethylene, respectively, are forced to be close to the thiol sulfur and thus
observation in docking/rescoring studies, irrespective of
utilized scoring functions, when subtle effects are at play.
produce unfavorable hydrophobic–polar interactions. In contrast, the methyls
of 2 can jacknife away from the sulfur.
420 | Med. Chem. Commun., 2013, 4, 417–421
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Concise Article
MedChemComm
3 J. Griggs, J. C. Metcalfe and R. Hesketh, Lancet Oncol., 2001,
2, 82.
Conclusions
We reported the synthesis, biological testing and modeling
studies of C-4 analogues of pyrrole-based antitubulin agents
targeting the colchicine site. For compounds that depolymerized
microtubules, a linear correlation was observed between the
antiproliferative activity and microtubule inhibitory activity,
molecular modeling results explained the SAR very well and they
both revealed that a weak hydrogen bond involved with Cysb241
was the key determiner of microtubule depolymerizing activity,
but the ideal ligand must incorporate (and properly position)
this acceptor within an otherwise hydrophobic framework.
Surprisingly, just the loss of that particular hydrogen bonding
interaction appears to shi the antiproliferative mechanism of
action away from microtubule inhibition.
4 C. Stengel, S. P. Newman, M. P. Leese, B. V. Potter, M. J. Reed
and A. Purohit, Br. J. Cancer, 2010, 102, 316.
5 A. Tripathi, M. Fornabaio, G. E. Kellogg, J. T. Gupton,
D. A. Gewirtz, W. A. Yeudall, N. E. Vega and
S. L. Mooberry, Bioorg. Med. Chem., 2008, 16, 2235.
6 C. Da, N. Telang, P. Barelli, X. Jia, J. T. Gupton, S. L. Mooberry
and G. E. Kellogg, ACS Med. Chem. Lett., 2012, 3, 53.
7 S. L. Mooberry, K. N. Weiderhold, S. Dakshanamurthy,
E. Hamel, E. J. Banner, A. Kharlamova, J. Hempel,
J. T. Gupton and M. L. Brown, Mol. Pharmacol., 2007, 72, 132.
8 N. Dalyot-Herman, F. Delgado-Lopez, D. A. Gewirtz,
J. T. Gupton and E. L. Schwartz, Biochem. Pharmacol., 2009,
78, 1167.
This study has fairly exhaustively probed subpocket A; the
3,4-dimethoxyphenyl substituent at the pyrrole C-4 is – to date –
the most ideal. The development of analogues focusing on other
positions on the pyrrole core is in progress.
9 J. Gupton, B. Burnham, K. Krumpe, K. Du, J. Sikorski,
A. Warren, C. Barnes and I. Hall, Arch. Pharm., 2000, 333, 3.
10 R. B. Ravelli, B. Gigant, P. A. Curmi, I. Jourdain, S. Lachkar,
A. Sobel and M. Knossow, Nature, 2004, 428, 198.
11 A. Dorleans, B. Gigant, R. B. Ravelli, P. Mailliet, V. Mikol and
M. Knossow, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 13775.
12 G. Jones, P. Willett and R. Glen, J. Mol. Biol., 1995, 245, 43.
13 SYBYL 8.1, Tripos Interactional, 1699 South Hanley Rd., St.
Louis, Missouri, 63144, USA.
14 G. E. Kellogg and D. J. Abraham, Eur. J. Med. Chem., 2000, 35,
651.
15 F. Spyrakis, A. Amadasi, M. Fornabaio, D. J. Abraham,
A. Mozzarelli, G. E. Kellogg and P. Cozzini, Eur. J. Med.
Chem., 2007, 42, 921.
Acknowledgements
The excellent technical assistance of Ms Lyda Robb, Ms Cara
Westbrook and Mr Nicholas Dbydal-Hargreaves is gratefully
acknowledged. This work was partially supported by NIH R01
CA135043 (to D.A. Gewirtz, VCU), NIH R15 CA067236 (to JTG)
and the President's Council Research Excellence Award (to SLM).
Notes and references
16 J. Chen, T. Liu, X. Dong and Y. Hu, Mini-Rev. Med. Chem.,
2009, 9, 1174.
17 T. L. Nguyen, C. McGrath, A. R. Hermone, J. C. Burnett,
D. W. Zaharevitz, B. W. Day, P. Wipf, E. Hamel and
R. Gussio, J. Med. Chem., 2005, 48, 6107.
1 C. Dumontet and M. A. Jordan, Nat. Rev. Drug Discovery,
2010, 9, 790.
2 R. A. Stanton, K. M. Gernert, J. H. Nettles and R. Aneja, Med.
Res. Rev., 2011, 31, 443.
This journal is ª The Royal Society of Chemistry 2013
Med. Chem. Commun., 2013, 4, 417–421 | 421